| Title: | Data Sets from Ramsey and Schafer's "Statistical Sleuth (3rd Ed)" |
|---|---|
| Description: | Data sets from Ramsey, F.L. and Schafer, D.W. (2013), "The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed)", Cengage Learning. |
| Authors: | Original by F.L. Ramsey and D.W. Schafer; modifications by Daniel W. Schafer, Jeannie Sifneos and Berwin A. Turlach; vignettes contributed by Nicholas Horton, Linda Loi, Kate Aloisio and Ruobing Zhang, with corrections by Randall Pruim |
| Maintainer: | Berwin A Turlach <[email protected]> |
| License: | GPL (>= 2) |
| Version: | 1.0-6 |
| Built: | 2024-10-24 05:31:51 UTC |
| Source: | https://github.com/r-forge/sleuth2 |
Data sets from Ramsey and Schafer's "Statistical Sleuth (3rd ed)"
This package contains a variety of datasets. For a complete
list, use library(help="Sleuth3") or
Sleuth3Manual().
Original by F.L. Ramsey and D.W. Schafer
Modifications by Daniel W Schafer, Jeannie Sifneos and Berwin A Turlach
Maintainer: Berwin A Turlach [email protected]
Data from an experiment concerning the effects of intrinsic and extrinsic motivation on creativity. Subjects with considerable experience in creative writing were randomly assigned to on of two treatment groups.
case0101case0101
A data frame with 47 observations on the following 2 variables.
creativity score
factor denoting the treatment group, with
levels "Extrinsic" and "Intrinsic"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Amabile, T. (1985). Motivation and Creativity: Effects of Motivational Orientation on Creative Writers, Journal of Personality and Social Psychology 48(2): 393–399.
attach(case0101) str(case0101) boxplot(Score ~ Treatment) # Basic boxplots for each level of Treatment boxplot(Score ~ Treatment, # Boxplots with labels ylab= "Average Creativity Score From 11 Judges (on a 40-point scale)", names=c("23 'Extrinsic' Group Students","24 'Intrinsic' Group Students"), main= "Haiku Creativity Scores for 47 Creative Writing Students") detach(case0101)attach(case0101) str(case0101) boxplot(Score ~ Treatment) # Basic boxplots for each level of Treatment boxplot(Score ~ Treatment, # Boxplots with labels ylab= "Average Creativity Score From 11 Judges (on a 40-point scale)", names=c("23 'Extrinsic' Group Students","24 'Intrinsic' Group Students"), main= "Haiku Creativity Scores for 47 Creative Writing Students") detach(case0101)
The data are the beginning salaries for all 32 male and all 61 female skilled, entry–level clerical employees hired by a bank between 1969 and 1977.
case0102case0102
A data frame with 93 observations on the following 2 variables.
starting salaries (in US)
sex of the clerical employee, with levels "Female"
and "Male"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Roberts, H.V. (1979). Harris Trust and Savings Bank: An Analysis of Employee Compensation, Report 7946, Center for Mathematical Studies in Business and Economics, University of Chicago Graduate School of Business.
attach(case0102) str(case0102) boxplot(Salary ~ Sex, ylab= "Starting Salary (U.S. Dollars)", names=c("61 Females","32 Males"), main= "Harris Bank Entry Level Clerical Workers, 1969-1971") hist(Salary[Sex=="Female"]) dev.new() hist(Salary[Sex=="Male"]) t.test(Salary ~ Sex, var.equal=TRUE) # Equal var. version; 2-sided by default t.test(Salary ~ Sex, var.equal=TRUE, alternative = "less") # 1-sided; that group 1 (females) mean is less detach(case0102)attach(case0102) str(case0102) boxplot(Salary ~ Sex, ylab= "Starting Salary (U.S. Dollars)", names=c("61 Females","32 Males"), main= "Harris Bank Entry Level Clerical Workers, 1969-1971") hist(Salary[Sex=="Female"]) dev.new() hist(Salary[Sex=="Male"]) t.test(Salary ~ Sex, var.equal=TRUE) # Equal var. version; 2-sided by default t.test(Salary ~ Sex, var.equal=TRUE, alternative = "less") # 1-sided; that group 1 (females) mean is less detach(case0102)
In the 1980s, biologists Peter and Rosemary Grant caught and measured all the birds from more than 20 generations of finches on the Galapagos island of Daphne Major. In one of those years, 1977, a severe drought caused vegetation to wither, and the only remaining food source was a large, tough seed, which the finches ordinarily ignored. Were the birds with larger and stronger beaks for opening these tough seeds more likely to survive that year, and did they tend to pass this characteristic to their offspring? The data are beak depths (height of the beak at its base) of 89 finches caught the year before the drought (1976) and 89 finches captured the year after the drought (1978).
case0201case0201
A data frame with 178 observations on the following 2 variables.
Year the finch was caught, 1976 or 1978
Beak depth of the finch (mm)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Grant, P. (1986). Ecology and Evolution of Darwin's Finches, Princeton University Press, Princeton, N.J.
attach(case0201) str(case0201) mean(Depth[Year==1978]) - mean(Depth[Year==1976]) yearFactor <- factor(Year) # Convert the numerical variable Year into a factor # with 2 levels. 1976 is "group 1" (it comes first alphanumerically) t.test(Depth ~ yearFactor, var.equal=TRUE) # 2-sample t-test; 2-sided by default t.test(Depth ~ yearFactor, var.equal=TRUE, alternative = "less") # 1-sided; alternative: group 1 mean is less boxplot(Depth ~ Year, ylab= "Beak Depth (mm)", names=c("89 Finches in 1976","89 Finches in 1978"), main= "Beak Depths of Darwin Finches in 1976 and 1978") ## BOXPLOTS FOR PRESENTATION boxplot(Depth ~ Year, ylab="Beak Depth (mm)", names=c("89 Finches in 1976","89 Finches in 1978"), main="Beak Depths of Darwin Finches in 1976 and 1978", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) detach(case0201)attach(case0201) str(case0201) mean(Depth[Year==1978]) - mean(Depth[Year==1976]) yearFactor <- factor(Year) # Convert the numerical variable Year into a factor # with 2 levels. 1976 is "group 1" (it comes first alphanumerically) t.test(Depth ~ yearFactor, var.equal=TRUE) # 2-sample t-test; 2-sided by default t.test(Depth ~ yearFactor, var.equal=TRUE, alternative = "less") # 1-sided; alternative: group 1 mean is less boxplot(Depth ~ Year, ylab= "Beak Depth (mm)", names=c("89 Finches in 1976","89 Finches in 1978"), main= "Beak Depths of Darwin Finches in 1976 and 1978") ## BOXPLOTS FOR PRESENTATION boxplot(Depth ~ Year, ylab="Beak Depth (mm)", names=c("89 Finches in 1976","89 Finches in 1978"), main="Beak Depths of Darwin Finches in 1976 and 1978", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) detach(case0201)
Are any physiological indicators associated with schizophrenia? In a
1990 article, researchers reported the results of a study that
controlled for genetic and socioeconomic differences by examining 15
pairs of monozygotic twins, where one of the twins was schizophrenic
and the other was not. The researchers used magnetic resonance
imaging to measure the volumes (in cm) of several regions and
subregions of the twins' brains.
case0202case0202
A data frame with 15 observations on the following 2 variables.
volume of left hippocampus of unaffected
twin (in cm)
volume of left hippocampus of affected twin
(in cm)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Suddath, R.L., Christison, G.W., Torrey, E.F., Casanova, M.F. and Weinberger, D.R. (1990). Anatomical Abnormalities in the Brains of Monozygotic Twins Discordant for Schizophrenia, New England Journal of Medicine 322(12): 789–794.
attach(case0202) str(case0202) diff <- Unaffected-Affected summary(diff) t.test(diff) # Paired t-test is a one-sample t-test on differences t.test(Unaffected,Affected,pair=TRUE) # Alternative coding for the same test boxplot(diff, ylab="Difference in Hippocampus Volume (cubic cm)", xlab="15 Sets of Twins, One Affected with Schizophrenia", main="Hippocampus Difference: Unaffected Twin Minus Affected Twin") abline(h=0,lty=2) # Draw a dashed (lty=2) horizontal line at 0 ## BOXPLOT FOR PRESENTATION: boxplot(diff, ylab="Difference in Hippocampus Volume (cubic cm)", xlab="15 Sets of Twins, One Affected with Schizophrenia", main="Hippocampus Difference: Unaffected Minus Affected Twin", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) abline(h=0,lty=2) detach(case0202)attach(case0202) str(case0202) diff <- Unaffected-Affected summary(diff) t.test(diff) # Paired t-test is a one-sample t-test on differences t.test(Unaffected,Affected,pair=TRUE) # Alternative coding for the same test boxplot(diff, ylab="Difference in Hippocampus Volume (cubic cm)", xlab="15 Sets of Twins, One Affected with Schizophrenia", main="Hippocampus Difference: Unaffected Twin Minus Affected Twin") abline(h=0,lty=2) # Draw a dashed (lty=2) horizontal line at 0 ## BOXPLOT FOR PRESENTATION: boxplot(diff, ylab="Difference in Hippocampus Volume (cubic cm)", xlab="15 Sets of Twins, One Affected with Schizophrenia", main="Hippocampus Difference: Unaffected Minus Affected Twin", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) abline(h=0,lty=2) detach(case0202)
Does dropping silver iodide onto clouds increase the amount of rainfall they produce? In a randomized experiment, researchers measured the volume of rainfall in a target area (in acre-feet) on 26 suitable days in which the clouds were seeded and on 26 suitble days in which the clouds were not seeded.
case0301case0301
A data frame with 52 observations on the following 2 variables.
the volume of rainfall in the target area (in acre-feet)
a factor with levels "Unseeded" and
"Seeded" indicating whether the clouds were unseeded or seeded.
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Simpson, J., Olsen, A., and Eden, J. (1975). A Bayesian Analysis of a Multiplicative Treatment Effect in Weather Modification. Technometrics 17: 161–166.
attach(case0301) str(case0301) #Seeded is level 1 of Treatment (it's first alphabetically) boxplot(Rainfall ~ Treatment) boxplot(log(Rainfall) ~ Treatment) # Boxplots of natural logs of Rainfall t.test(log(Rainfall) ~ Treatment, var.equal=TRUE, alternative="greater") # 1-sided t-test; alternative: level 1 mean is greater myTest <- t.test(log(Rainfall) ~ Treatment, var.equal=TRUE, alternative="two.sided") # 2-sided alternative to get confidence interval exp(myTest$est[1] - myTest$est[2]) # Back-transform estimate on log scale exp(myTest$conf) # Back transform endpoints of confidence interval boxplot(log(Rainfall) ~ Treatment, ylab="Log of Rainfall Volume in Target Area (Acre Feet)", names=c("On 26 Seeded Days", "On 26 Unseeded Days"), main="Distributions of Rainfalls from Cloud Seeding Experiment") ## POLISHED BOXPLOTS FOR PRESENTATION: opar <- par(no.readonly=TRUE) # Store device graphics parameters par(mar=c(4,4,4,4)) # Change margins to allow more space on right boxplot(log(Rainfall) ~ Treatment, ylab="Log Rainfall (Acre-Feet)", names=c("on 26 seeded days","on 26 unseeded days"), main="Boxplots of Rainfall on Log Scale", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5 ) myTicks <- c(1,5,10,100,500,1000,2000,3000) # some tick marks for original scale axis(4, at=log(myTicks), label=myTicks) # Add original-scale axis on right mtext("Rainfall (Acre Feet)", side=4, line=2.5) # Add right-side axis label par(opar) # Restore previous graphics parameter settings detach(case0301)attach(case0301) str(case0301) #Seeded is level 1 of Treatment (it's first alphabetically) boxplot(Rainfall ~ Treatment) boxplot(log(Rainfall) ~ Treatment) # Boxplots of natural logs of Rainfall t.test(log(Rainfall) ~ Treatment, var.equal=TRUE, alternative="greater") # 1-sided t-test; alternative: level 1 mean is greater myTest <- t.test(log(Rainfall) ~ Treatment, var.equal=TRUE, alternative="two.sided") # 2-sided alternative to get confidence interval exp(myTest$est[1] - myTest$est[2]) # Back-transform estimate on log scale exp(myTest$conf) # Back transform endpoints of confidence interval boxplot(log(Rainfall) ~ Treatment, ylab="Log of Rainfall Volume in Target Area (Acre Feet)", names=c("On 26 Seeded Days", "On 26 Unseeded Days"), main="Distributions of Rainfalls from Cloud Seeding Experiment") ## POLISHED BOXPLOTS FOR PRESENTATION: opar <- par(no.readonly=TRUE) # Store device graphics parameters par(mar=c(4,4,4,4)) # Change margins to allow more space on right boxplot(log(Rainfall) ~ Treatment, ylab="Log Rainfall (Acre-Feet)", names=c("on 26 seeded days","on 26 unseeded days"), main="Boxplots of Rainfall on Log Scale", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5 ) myTicks <- c(1,5,10,100,500,1000,2000,3000) # some tick marks for original scale axis(4, at=log(myTicks), label=myTicks) # Add original-scale axis on right mtext("Rainfall (Acre Feet)", side=4, line=2.5) # Add right-side axis label par(opar) # Restore previous graphics parameter settings detach(case0301)
In 1987, researchers measured the TCDD concentration in blood samples from 646 U.S. veterans of the Vietnam War and from 97 U.S. veterans who did not serve in Vietnam. TCDD is a carcinogenic dioxin in the herbicide called Agent Orange, which was used to clear jungle hiding areas by the U.S. military in the Vietnam War between 1962 and 1970.
data(case0302)data(case0302)
A data frame with 743 observations on the following 2 variables.
the concentration of TCDD, in parts per trillion
factor variable with two levels,
"Vietnam" and "Other", to indicate the type of veteran
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Centers for Disease Control Veterans Health Studies: Serum 2,3,7,8-Tetraclorodibenzo-p-dioxin Levels in U.S. Army Vietnam-era Veterans. Journal of the American Medical Association 260: 1249–1254.
attach(case0302) str(case0302) # Note: Level 1 of Veteran is "Other" (first alphabeticall) boxplot(Dioxin ~ Veteran) t.test(Dioxin ~ Veteran, var.equal=TRUE, alternative="less") # 1-sided t-test; alternative: group 1 mean is less t.test(Dioxin ~ Veteran, alternative="less", var.equal=TRUE, subset=(Dioxin < 40)) # t-test on subset for which Dioxin < 40 t.test(Dioxin ~ Veteran, alternative="less", var.equal=TRUE, subset=(Dioxin < 20)) t.test(Dioxin ~ Veteran, var.equal=TRUE) # 2-sided--to get confidence interval ## HISTOGRAMS FOR PRESENTATION opar <- par(no.readonly=TRUE) # Store device graphics parameter settings par(mfrow=c(2,1), mar=c(3,3,1,1)) # 2 by 1 layout of plots; change margins myBreaks <- (0:46) - .5 # Make breaks for histogram bins hist(Dioxin[Veteran=="Other"], breaks=myBreaks, xlim=range(Dioxin), col="green", xlab="", ylab="", main="") text(10,25, "Dioxin in 97 'Other' Veterans; Estimated mean = 4.19 ppt (95% CI: 3.72 to 4.65 ppt)", pos=4, cex=.75) # CI from 1-sample t-test & subset=(Veteran="Other") hist(Dioxin[Veteran=="Vietnam"],breaks=myBreaks,xlim=range(Dioxin), col="green", xlab="", ylab="", main="") text(10,160, "Dioxin in 646 Vietnam Veterans; Estimated mean = 4.26 ppt (95% CI: 4.06 to 4.64 ppt)", pos=4, cex=.75) text(13,145,"[Estimated Difference in Means: 0.07 ppt (95% CI: -0.63 to 0.48 ppt)]", pos=4, cex=.75) par(opar) # Restore previous graphics parameter settings detach(case0302)attach(case0302) str(case0302) # Note: Level 1 of Veteran is "Other" (first alphabeticall) boxplot(Dioxin ~ Veteran) t.test(Dioxin ~ Veteran, var.equal=TRUE, alternative="less") # 1-sided t-test; alternative: group 1 mean is less t.test(Dioxin ~ Veteran, alternative="less", var.equal=TRUE, subset=(Dioxin < 40)) # t-test on subset for which Dioxin < 40 t.test(Dioxin ~ Veteran, alternative="less", var.equal=TRUE, subset=(Dioxin < 20)) t.test(Dioxin ~ Veteran, var.equal=TRUE) # 2-sided--to get confidence interval ## HISTOGRAMS FOR PRESENTATION opar <- par(no.readonly=TRUE) # Store device graphics parameter settings par(mfrow=c(2,1), mar=c(3,3,1,1)) # 2 by 1 layout of plots; change margins myBreaks <- (0:46) - .5 # Make breaks for histogram bins hist(Dioxin[Veteran=="Other"], breaks=myBreaks, xlim=range(Dioxin), col="green", xlab="", ylab="", main="") text(10,25, "Dioxin in 97 'Other' Veterans; Estimated mean = 4.19 ppt (95% CI: 3.72 to 4.65 ppt)", pos=4, cex=.75) # CI from 1-sample t-test & subset=(Veteran="Other") hist(Dioxin[Veteran=="Vietnam"],breaks=myBreaks,xlim=range(Dioxin), col="green", xlab="", ylab="", main="") text(10,160, "Dioxin in 646 Vietnam Veterans; Estimated mean = 4.26 ppt (95% CI: 4.06 to 4.64 ppt)", pos=4, cex=.75) text(13,145,"[Estimated Difference in Means: 0.07 ppt (95% CI: -0.63 to 0.48 ppt)]", pos=4, cex=.75) par(opar) # Restore previous graphics parameter settings detach(case0302)
The number of space shuttle O-ring incidents for 4 space shuttle launches when the air temperatures was below 65 degrees F and for 20 space shuttle launches when the air temperature was above 65 degrees F.
case0401case0401
A data frame with 24 observations on the following 2 variables.
the number of O-ring incidents
factor variable with two levels—"Cool" and
"Warm"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Feynman, R.P. (1988). What do You Care What Other People Think? W. W. Norton.
str(case0401) attach(case0401) mCool <- mean(Incidents[Launch=="Cool"]) mWarm <- mean(Incidents[Launch=="Warm"]) mDiff <- mCool - mWarm c(mCool,mWarm,mDiff) # Show the values of these variables ## PERMUTATION TEST , VIA REPEATED RANDOM RE-GROUPING (ADVANCED) numRep <- 50 # Number of random groupings. CHANGE TO LARGER NUMBER; eg 50,000. rDiff <- rep(0,numRep) # Initialize this variable to contain numRep 0s. for (rep in 1:numRep) { # Repeat the following commands numRep times: randomGroup <- rep("rWarm",24) # Set randomGroup to have 24 values "rWarm" randomGroup[sample(1:24,4)] <- "rCool" # Replace 4 at random with "rCool" mW <- mean(Incidents[randomGroup=="rWarm"]) # average of random "rWarm" group mC <- mean(Incidents[randomGroup=="rCool"]) # average of random "rCool" group rDiff[rep] <- mC-mW # Store difference in averages in 'rep' cell of rDiff } # End of loop hist(rDiff, # Histogram of difference in averages from numRep random groupings main="Approximate Permutation Distribution", xlab="Possible Values of Difference in Averages", ylab="Frequency of Occurrence") abline(v=mDiff) # Draw a vertical line at the actually observed difference pValue <- sum(rDiff >= 1.3)/numRep # 1-sided p-value pValue text(mDiff,75000, paste(" -->",round(pValue,4)), adj=-0.1) detach(case0401)str(case0401) attach(case0401) mCool <- mean(Incidents[Launch=="Cool"]) mWarm <- mean(Incidents[Launch=="Warm"]) mDiff <- mCool - mWarm c(mCool,mWarm,mDiff) # Show the values of these variables ## PERMUTATION TEST , VIA REPEATED RANDOM RE-GROUPING (ADVANCED) numRep <- 50 # Number of random groupings. CHANGE TO LARGER NUMBER; eg 50,000. rDiff <- rep(0,numRep) # Initialize this variable to contain numRep 0s. for (rep in 1:numRep) { # Repeat the following commands numRep times: randomGroup <- rep("rWarm",24) # Set randomGroup to have 24 values "rWarm" randomGroup[sample(1:24,4)] <- "rCool" # Replace 4 at random with "rCool" mW <- mean(Incidents[randomGroup=="rWarm"]) # average of random "rWarm" group mC <- mean(Incidents[randomGroup=="rCool"]) # average of random "rCool" group rDiff[rep] <- mC-mW # Store difference in averages in 'rep' cell of rDiff } # End of loop hist(rDiff, # Histogram of difference in averages from numRep random groupings main="Approximate Permutation Distribution", xlab="Possible Values of Difference in Averages", ylab="Frequency of Occurrence") abline(v=mDiff) # Draw a vertical line at the actually observed difference pValue <- sum(rDiff >= 1.3)/numRep # 1-sided p-value pValue text(mDiff,75000, paste(" -->",round(pValue,4)), adj=-0.1) detach(case0401)
Educational researchers randomly assigned 28 ninth-year students in Australia to receive coordinate geometry training in one of two ways: a conventional way and a modified way. After the training, the students were asked to solve a coordinate geometry problem. The time to complete the problem was recorded, but five students in the “conventional” group did not complete the solution in the five minute alloted time.
case0402case0402
A data frame with 28 observations on the following 3 variables.
the time (in seconds) that the student worked on the problem
factor variable with two
levels—"Modified" and "Conventional"
1 if the individual did not complete the problem in 5 minutes, 0 if they did
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Sweller, J., Chandler, P., Tierney, P. and Cooper, M. (1990). Cognitive Load as a Factor in the Structuring of Technical Material, Journal of Experimental Psychology General 119(2): 176–192.
str(case0402) # level 1 of Treatment is "Conventional" (1st alphabetically) attach(case0402) boxplot(Time ~ Treatment) median(Time[Treatment=="Conventional"])-median(Time[Treatment=="Modified"]) wilcox.test(Time ~ Treatment, exact=FALSE, correct=TRUE, alternative="greater") # Rank-sum test; alternative: group 1 is greater wilcox.test(Time ~ Treatment, exact=FALSE, correct=TRUE, alternative="two.sided", conf.int=TRUE) # Use 2-sided to get confidence int. ## DOT PLOTS FOR PRESENTATION xTreatment <- ifelse(Treatment=="Conventional",1,2) # Make numerical values myPointCode <- ifelse(Censored==0,21,24) plot(Time ~ jitter(xTreatment,.2), # Jitter the 1's and 2's for visibility ylab="Completion Time (Sec.)", xlab="Training Method (jittered)", main="Test Completion Times from Cognitive Load Experiment", axes=FALSE, pch=myPointCode, bg="green", cex=2, xlim=c(.5,2.5) ) axis(2) # Draw y-axis as usual axis(1, tick=FALSE, at=c(1,2), # Draw x-axis without ticks labels=c("Conventional (n=14 Students)","Modified (n=14 Students)") ) legend(1.5,300, legend=c("Did not Complete in 300 sec","Completed in 300 sec."), pch=c(24,21), pt.cex=2, pt.bg="green") detach(case0402)str(case0402) # level 1 of Treatment is "Conventional" (1st alphabetically) attach(case0402) boxplot(Time ~ Treatment) median(Time[Treatment=="Conventional"])-median(Time[Treatment=="Modified"]) wilcox.test(Time ~ Treatment, exact=FALSE, correct=TRUE, alternative="greater") # Rank-sum test; alternative: group 1 is greater wilcox.test(Time ~ Treatment, exact=FALSE, correct=TRUE, alternative="two.sided", conf.int=TRUE) # Use 2-sided to get confidence int. ## DOT PLOTS FOR PRESENTATION xTreatment <- ifelse(Treatment=="Conventional",1,2) # Make numerical values myPointCode <- ifelse(Censored==0,21,24) plot(Time ~ jitter(xTreatment,.2), # Jitter the 1's and 2's for visibility ylab="Completion Time (Sec.)", xlab="Training Method (jittered)", main="Test Completion Times from Cognitive Load Experiment", axes=FALSE, pch=myPointCode, bg="green", cex=2, xlim=c(.5,2.5) ) axis(2) # Draw y-axis as usual axis(1, tick=FALSE, at=c(1,2), # Draw x-axis without ticks labels=c("Conventional (n=14 Students)","Modified (n=14 Students)") ) legend(1.5,300, legend=c("Did not Complete in 300 sec","Completed in 300 sec."), pch=c(24,21), pt.cex=2, pt.bg="green") detach(case0402)
Female mice were randomly assigned to six treatment groups to investigate whether restricting dietary intake increases life expectancy. Diet treatments were:
"NP"—mice ate unlimited amount of nonpurified, standard
diet
"N/N85"—mice fed normally before and after weaning.
After weaning, ration was controlled at 85 kcal/wk
"N/R50"—normal diet before weaning and reduced calorie
diet (50 kcal/wk) after weaning
"R/R50"—reduced calorie diet of 50 kcal/wk both before
and after weaning
"N/R50 lopro"—normal diet before weaning, restricted
diet (50 kcal/wk) after weaning and dietary protein content
decreased with advancing age
"N/R40"—normal diet before weaning and reduced diet (40
Kcal/wk) after weaning.
case0501case0501
A data frame with 349 observations on the following 2 variables.
the lifetime of the mice (in months)
factor variable with six levels—"NP",
"N/N85", "lopro", "N/R50", "R/R50" and
"N/R40"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Weindruch, R., Walford, R.L., Fligiel, S. and Guthrie D. (1986). The Retardation of Aging in Mice by Dietary Restriction: Longevity, Cancer, Immunity and Lifetime Energy Intake, Journal of Nutrition 116(4):641–54.
str(case0501) attach(case0501) # Re-order levels for better boxplot organization: myDiet <- factor(Diet, levels=c("NP","N/N85","N/R50","R/R50","lopro","N/R40") ) myNames <- c("NP(49)","N/N85(57)","N/R50(71)","R/R50(56)","lopro(56)", "N/R40(60)") # Make these for boxplot labeling. boxplot(Lifetime ~ myDiet, ylab= "Lifetime (months)", names=myNames, xlab="Treatment (and sample size)") myAov1 <- aov(Lifetime ~ Diet) # One-way analysis of variance plot(myAov1, which=1) # Plot residuals versus estimated means. summary(myAov1) pairwise.t.test(Lifetime,Diet, pool.SD=TRUE, p.adj="none") # All t-tests ## p-VALUES AND CONFIDENCE INTERVALS FOR SPECIFIED COMPARISONS OF MEANS if(require(multcomp)){ diet <- factor(Diet,labels=c("NN85", "NR40", "NR50", "NP", "RR50", "lopro")) myAov2 <- aov(Lifetime ~ diet - 1) myComparisons <- glht(myAov2, linfct=c("dietNR50 - dietNN85 = 0", "dietRR50 - dietNR50 = 0", "dietNR40 - dietNR50 = 0", "dietlopro - dietNR50 = 0", "dietNN85 - dietNP = 0") ) summary(myComparisons,test=adjusted("none")) # No multiple comparison adjust. confint(myComparisons, calpha = univariate_calpha()) # No adjustment } ## EXAMPLE 5: BOXPLOTS FOR PRESENTATION boxplot(Lifetime ~ myDiet, ylab= "Lifetime (months)", names=myNames, main= "Lifetimes of Mice on 6 Diet Regimens", xlab="Diet (and sample size)", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) detach(case0501)str(case0501) attach(case0501) # Re-order levels for better boxplot organization: myDiet <- factor(Diet, levels=c("NP","N/N85","N/R50","R/R50","lopro","N/R40") ) myNames <- c("NP(49)","N/N85(57)","N/R50(71)","R/R50(56)","lopro(56)", "N/R40(60)") # Make these for boxplot labeling. boxplot(Lifetime ~ myDiet, ylab= "Lifetime (months)", names=myNames, xlab="Treatment (and sample size)") myAov1 <- aov(Lifetime ~ Diet) # One-way analysis of variance plot(myAov1, which=1) # Plot residuals versus estimated means. summary(myAov1) pairwise.t.test(Lifetime,Diet, pool.SD=TRUE, p.adj="none") # All t-tests ## p-VALUES AND CONFIDENCE INTERVALS FOR SPECIFIED COMPARISONS OF MEANS if(require(multcomp)){ diet <- factor(Diet,labels=c("NN85", "NR40", "NR50", "NP", "RR50", "lopro")) myAov2 <- aov(Lifetime ~ diet - 1) myComparisons <- glht(myAov2, linfct=c("dietNR50 - dietNN85 = 0", "dietRR50 - dietNR50 = 0", "dietNR40 - dietNR50 = 0", "dietlopro - dietNR50 = 0", "dietNN85 - dietNP = 0") ) summary(myComparisons,test=adjusted("none")) # No multiple comparison adjust. confint(myComparisons, calpha = univariate_calpha()) # No adjustment } ## EXAMPLE 5: BOXPLOTS FOR PRESENTATION boxplot(Lifetime ~ myDiet, ylab= "Lifetime (months)", names=myNames, main= "Lifetimes of Mice on 6 Diet Regimens", xlab="Diet (and sample size)", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) detach(case0501)
In 1968, Dr. Benjamin Spock was tried in Boston on charges of conspiring to violate the Selective Service Act by encouraging young men to resist being drafted into military service for Vietnam. The defence in the case challenged the method of jury selection claiming that women were underrepresented. Boston juries are selected in three stages. First 300 names are selected at random from the City Directory, then a venire of 30 or more jurors is selected from the initial list of 300 and finally, an actual jury is selected from the venire in a nonrandom process allowing each side to exclude certain jurors. There was one woman on the venire and no women on the final list. The defence argued that the judge in the trial had a history of venires in which women were systematically underrepresented and compared the judge's recent venires with the venires of six other Boston area district judges.
case0502case0502
A data frame with 46 observations on the following 2 variables.
is the percent of women on the venire's of the Spock trial judge and 6 other Boston area judges
is a factor with levels "Spock's", "A",
"B", "C", "D", "E" and "F"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Zeisel, H. and Kalven, H. Jr. (1972). Parking Tickets and Missing Women: Statistics and the Law in Tanur, J.M. et al. (eds.) Statistics: A Guide to the Unknown, Holden-Day.
str(case0502) attach(case0502) # Make new factor level names (with sample sizes) for boxplots myNames <- c("A (5)", "B (6)", "C (9)", "D (2)", "E (6)", "F (9)", "Spock's (9)") boxplot(Percent ~ Judge, ylab = "Percent of Women on Judges' Venires", names = myNames, xlab = "Judge (and number of venires)", main = "Percent Women on Venires of 7 Massachusetts Judges") myAov1 <- aov(Percent ~ Judge) plot(myAov1, which=1) # Residual plot summary(myAov1) # Initial screening. Any evidence of judge differences? (yes) ## ANALYSIS 1. TWO-SAMPLE t-TEST (ASSUMING NON-SPOCK JUDGES HAVE A COMMON MEAN) SpockOrOther <- factor(ifelse(Judge=="Spock's","Spock","Other")) aovFull <- aov(Percent ~ Judge) aovReduced <- aov(Percent ~ SpockOrOther) anova(aovReduced,aovFull) #Any evidence that 7 mean fits better than the 2 mean? t.test(Percent ~ SpockOrOther, var.equal=TRUE) # Evidence that 2 means differ? ## ANALYSIS 2. COMPARE SPOCK MEAN TO AVERAGE OF OTHER MEANS myAov3 <- aov(Percent ~ Judge - 1) myContrast <- rbind(c(1/6, 1/6, 1/6, 1/6, 1/6, 1/6, - 1)) if(require(multcomp)){ # use multcomp library myComparison <- glht(myAov3, linfct=myContrast) summary(myComparison, test=adjusted("none")) confint(myComparison) } ## BOXPLOTS FOR PRESENTATION boxplot(Percent ~ Judge, ylab= "Percent of Women on Judges' Venires", names=myNames, xlab="Judge (and number of venires)", main= "Percent Women on Venires of 7 Massachusetts Judges", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) detach(case0502)str(case0502) attach(case0502) # Make new factor level names (with sample sizes) for boxplots myNames <- c("A (5)", "B (6)", "C (9)", "D (2)", "E (6)", "F (9)", "Spock's (9)") boxplot(Percent ~ Judge, ylab = "Percent of Women on Judges' Venires", names = myNames, xlab = "Judge (and number of venires)", main = "Percent Women on Venires of 7 Massachusetts Judges") myAov1 <- aov(Percent ~ Judge) plot(myAov1, which=1) # Residual plot summary(myAov1) # Initial screening. Any evidence of judge differences? (yes) ## ANALYSIS 1. TWO-SAMPLE t-TEST (ASSUMING NON-SPOCK JUDGES HAVE A COMMON MEAN) SpockOrOther <- factor(ifelse(Judge=="Spock's","Spock","Other")) aovFull <- aov(Percent ~ Judge) aovReduced <- aov(Percent ~ SpockOrOther) anova(aovReduced,aovFull) #Any evidence that 7 mean fits better than the 2 mean? t.test(Percent ~ SpockOrOther, var.equal=TRUE) # Evidence that 2 means differ? ## ANALYSIS 2. COMPARE SPOCK MEAN TO AVERAGE OF OTHER MEANS myAov3 <- aov(Percent ~ Judge - 1) myContrast <- rbind(c(1/6, 1/6, 1/6, 1/6, 1/6, 1/6, - 1)) if(require(multcomp)){ # use multcomp library myComparison <- glht(myAov3, linfct=myContrast) summary(myComparison, test=adjusted("none")) confint(myComparison) } ## BOXPLOTS FOR PRESENTATION boxplot(Percent ~ Judge, ylab= "Percent of Women on Judges' Venires", names=myNames, xlab="Judge (and number of venires)", main= "Percent Women on Venires of 7 Massachusetts Judges", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) detach(case0502)
Study explores how physical handicaps affect people's perception of employment qualifications. Researchers prepared 5 videotaped job interviews using actors with a script designed to reflect an interview with an applicant of average qualifications. The 5 tapes differed only in that the applicant appeared with a different handicap in each one. Seventy undergraduate students were randomly assigned to view the tapes and rate the qualification of the applicant on a 0-10 point scale.
case0601case0601
A data frame with 70 observations on the following 2 variables.
is the score each student gave to the applicant
is a factor variable with 5
levels—"None", "Amputee", "Crutches",
"Hearing" and "Wheelchair"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Cesare, S.J., Tannenbaum, R.J. and Dalessio, A. (1990). Interviewers' Decisions Related to Applicant Handicap Type and Rater Empathy, Human Performance 3(3): 157–171.
str(case0601) attach(case0601) ## EXPLORATION myHandicap <- factor(Handicap, levels=c("None","Amputee","Crutches","Hearing","Wheelchair")) boxplot(Score ~ myHandicap, ylab= "Qualification Score Assigned by Student to Interviewee", xlab= "Treatment Group--Handicap Portrayed (14 Students in each Group)", main= "Handicap Discrimination Experiment on 70 Undergraduate Students") myAov <- aov(Score ~ myHandicap) plot(myAov, which=1) # Plot residuals versus estimated means summary(myAov) ## COMPARE MEAN QUALIFICATION SCORE OF EVERY HANDICAP GROUP TO "NONE" if(require(multcomp)){ # Use the multcomp library myDunnett <- glht(myAov, linfct = mcp(myHandicap = "Dunnett")) summary(myDunnett) confint(myDunnett,level=.95) opar <- par(no.readonly=TRUE) # Save current graphics parameter settings par(mar=c(4.1,8.1,4.1,1.1)) # Change margins plot(myDunnett, xlab="Difference in Mean Qualification Score (and Dunnet-adjusted CIs)") par(opar) # Restore original graphics parameter settings } ## COMPARE EVERY MEAN TO EVERY OTHER MEAN if(require(multcomp)){ # Use the multcomp library myTukey <- glht(myAov, linfct = mcp(myHandicap = "Tukey")) summary(myTukey) } ## TEST THE CONTRAST OF DISPLAY 6.4 myAov2 <- aov(Score ~ myHandicap - 1) myContrast <- rbind(c(0, -1/2, 1/2, -1/2, 1/2)) if(require(multcomp)){ # Use the multcomp library myComparison <- glht(myAov2, linfct=myContrast) summary(myComparison, test=adjusted("none")) confint(myComparison) } # BOXPLOTS FOR PRESENTATION boxplot(Score ~ myHandicap, ylab= "Qualification Score Assigned by Student to Video Job Applicant", xlab="Handicap Portrayed by Job Applicant in Video (14 Students in each Group)", main= "Handicap Discrimination Experiment on 70 Undergraduate Students", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) detach(case0601)str(case0601) attach(case0601) ## EXPLORATION myHandicap <- factor(Handicap, levels=c("None","Amputee","Crutches","Hearing","Wheelchair")) boxplot(Score ~ myHandicap, ylab= "Qualification Score Assigned by Student to Interviewee", xlab= "Treatment Group--Handicap Portrayed (14 Students in each Group)", main= "Handicap Discrimination Experiment on 70 Undergraduate Students") myAov <- aov(Score ~ myHandicap) plot(myAov, which=1) # Plot residuals versus estimated means summary(myAov) ## COMPARE MEAN QUALIFICATION SCORE OF EVERY HANDICAP GROUP TO "NONE" if(require(multcomp)){ # Use the multcomp library myDunnett <- glht(myAov, linfct = mcp(myHandicap = "Dunnett")) summary(myDunnett) confint(myDunnett,level=.95) opar <- par(no.readonly=TRUE) # Save current graphics parameter settings par(mar=c(4.1,8.1,4.1,1.1)) # Change margins plot(myDunnett, xlab="Difference in Mean Qualification Score (and Dunnet-adjusted CIs)") par(opar) # Restore original graphics parameter settings } ## COMPARE EVERY MEAN TO EVERY OTHER MEAN if(require(multcomp)){ # Use the multcomp library myTukey <- glht(myAov, linfct = mcp(myHandicap = "Tukey")) summary(myTukey) } ## TEST THE CONTRAST OF DISPLAY 6.4 myAov2 <- aov(Score ~ myHandicap - 1) myContrast <- rbind(c(0, -1/2, 1/2, -1/2, 1/2)) if(require(multcomp)){ # Use the multcomp library myComparison <- glht(myAov2, linfct=myContrast) summary(myComparison, test=adjusted("none")) confint(myComparison) } # BOXPLOTS FOR PRESENTATION boxplot(Score ~ myHandicap, ylab= "Qualification Score Assigned by Student to Video Job Applicant", xlab="Handicap Portrayed by Job Applicant in Video (14 Students in each Group)", main= "Handicap Discrimination Experiment on 70 Undergraduate Students", col="green", boxlwd=2, medlwd=2, whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5) detach(case0601)
Do female Platyfish prefer male Platyfish with yellow swordtails? A.L. Basolo proposed and tested a selection model in which females have a pre-existing bias for a male trait even before the males possess it. Six pairs of males were surgically given artificial, plastic swordtails—one pair received a bright yellow sword, the other a transparent sword. Females were given the opportunity to engage in courtship activity with either of the males. Of the total time spent by each female engaged in courtship during a 20 minute observation period, the percentages of time spent with the yellow-sword male were recorded.
case0602case0602
A data frame with 84 observations on the following 3 variables.
The percentage of courtship time spent by 84 females with the yellow-sword males
Factor variable with 6 levels—"Pair1",
"Pair2", "Pair3", "Pair4", "Pair5" and
"Pair6"
Body size of the males
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Basolo, A.L. (1990). Female Preference Predates the Evolution of the Sword in Swordtail Fish, Science 250: 808–810.
str(case0602) attach(case0602) ## EXPLORATION plot(Percentage ~ Length, xlab="Length of the Two Males", ylab="Percentage of Time Female Spent with Yellow-Sword Male", main="Percentage of Time Spent with Yellow Rather than Transparent Sword Male") abline(h=50) # Draw a horizontal line at 50% (i.e. the "no preference" line) myAov <- aov(Percentage ~ Pair) plot(myAov, which=1) # Resdiual plot summary(myAov) # Explore possibility of linear effect, as in Display 6.5 myAov2 <- aov(Percentage ~ Pair - 1) # Show the estimated means. myContrast <- rbind(c(5, -3, 1, 3, -9, 3)) if(require(multcomp)){ # Use the multcomp library myComparison <- glht(myAov2, linfct=myContrast) summary(myComparison, test=adjusted("none")) } # Simpler exploration of linear effect, via regression (Ch. 7) myLm <- lm(Percentage ~ Length) summary(myLm) # ONE-SAMPLE t-TEST THAT MEAN PERCENTAGE = 50%, IGNORING MALE PAIR EFFECT t.test(Percentage, mu=50, alternative="greater") # Get 1-sided p-value t.test(Percentage, alternative="two.sided") # Get C.I. ## SCATTERPLOT FOR PRESENTATION plot(Percentage ~ Length, xlab="Length of the Two Males (mm)", ylab="Percentage of Time Female Spent with Yellow-Sword Male", main="Female Preference for Yellow Rather than Transparent Sword Male", pch=21, lwd=2, bg="green", cex=1.5 ) abline(h=50,lty=2,col="blue",lwd=2) text(29.5,52,"50% (no preference)", col="blue") detach(case0602)str(case0602) attach(case0602) ## EXPLORATION plot(Percentage ~ Length, xlab="Length of the Two Males", ylab="Percentage of Time Female Spent with Yellow-Sword Male", main="Percentage of Time Spent with Yellow Rather than Transparent Sword Male") abline(h=50) # Draw a horizontal line at 50% (i.e. the "no preference" line) myAov <- aov(Percentage ~ Pair) plot(myAov, which=1) # Resdiual plot summary(myAov) # Explore possibility of linear effect, as in Display 6.5 myAov2 <- aov(Percentage ~ Pair - 1) # Show the estimated means. myContrast <- rbind(c(5, -3, 1, 3, -9, 3)) if(require(multcomp)){ # Use the multcomp library myComparison <- glht(myAov2, linfct=myContrast) summary(myComparison, test=adjusted("none")) } # Simpler exploration of linear effect, via regression (Ch. 7) myLm <- lm(Percentage ~ Length) summary(myLm) # ONE-SAMPLE t-TEST THAT MEAN PERCENTAGE = 50%, IGNORING MALE PAIR EFFECT t.test(Percentage, mu=50, alternative="greater") # Get 1-sided p-value t.test(Percentage, alternative="two.sided") # Get C.I. ## SCATTERPLOT FOR PRESENTATION plot(Percentage ~ Length, xlab="Length of the Two Males (mm)", ylab="Percentage of Time Female Spent with Yellow-Sword Male", main="Female Preference for Yellow Rather than Transparent Sword Male", pch=21, lwd=2, bg="green", cex=1.5 ) abline(h=50,lty=2,col="blue",lwd=2) text(29.5,52,"50% (no preference)", col="blue") detach(case0602)
Hubble's initial data on 24 nebulae outside the Milky Way.
case0701case0701
A data frame with 24 observations on the following 2 variables.
recession velocity (in kilometres per second)
distance from earth (in magaparsec)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Hubble, E. (1929). A Relation Between Distance and Radial Velocity Among Extragalactic Nebulae, Proceedings of the National Academy of Science 15: 168–173.
str(case0701) attach(case0701) ## EXPLORATION plot(Distance ~ Velocity) myLm <- lm(Distance ~ Velocity) abline(myLm) myResiduals <- myLm$res myFits <- myLm$fit plot(myResiduals ~ myFits) # Plot residuals versus estimated means. abline(h=0) # Draw a horizontal line at 0. # OR, use this shortcut... plot(myLm, which=1) # Residual plot (red curve is a scatterplot smooother) ## INFERENCE summary(myLm) confint(myLm,level=.95) myLm2 <- lm(Distance ~ Velocity - 1) # Drop the intercept. summary(myLm2) confint(myLm2) ## DISPLAY FOR PRESENTATION plot(Distance ~ Velocity, xlab="Recession Velocity (km/sec)", ylab="Distance from Earth (megaparsecs)", main="Measured Distance Versus Velocity for 24 Extra-Galactic Nebulae", pch=21, lwd=2, bg="green", cex=1.5 ) abline(myLm, lty=2, col="blue", lwd=2) abline(myLm2, lty=3, col="red", lwd=2) legend(-250,2.05, c("unrestricted regression line","regression through the origin"), lty=c(2,3), lwd=c(2,2), col=c("blue","red")) detach(case0701)str(case0701) attach(case0701) ## EXPLORATION plot(Distance ~ Velocity) myLm <- lm(Distance ~ Velocity) abline(myLm) myResiduals <- myLm$res myFits <- myLm$fit plot(myResiduals ~ myFits) # Plot residuals versus estimated means. abline(h=0) # Draw a horizontal line at 0. # OR, use this shortcut... plot(myLm, which=1) # Residual plot (red curve is a scatterplot smooother) ## INFERENCE summary(myLm) confint(myLm,level=.95) myLm2 <- lm(Distance ~ Velocity - 1) # Drop the intercept. summary(myLm2) confint(myLm2) ## DISPLAY FOR PRESENTATION plot(Distance ~ Velocity, xlab="Recession Velocity (km/sec)", ylab="Distance from Earth (megaparsecs)", main="Measured Distance Versus Velocity for 24 Extra-Galactic Nebulae", pch=21, lwd=2, bg="green", cex=1.5 ) abline(myLm, lty=2, col="blue", lwd=2) abline(myLm2, lty=3, col="red", lwd=2) legend(-250,2.05, c("unrestricted regression line","regression through the origin"), lty=c(2,3), lwd=c(2,2), col=c("blue","red")) detach(case0701)
A certain kind of meat processing may begin once the pH in postmortem muscle of a steer carcass has decreased sufficiently. To estimate the timepoint at which pH has dropped sufficiently, 10 steer carcasses were assigned to be measured for pH at one of five times after slaughter.
case0702case0702
A data frame with 10 observations on the following 2 variables.
time after slaughter (hours)
pH level in postmortem muscle
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Schwenke, J.R. and Milliken, G.A. (1991). On the Calibration Problem Extended to Nonlinear Models, Biometrics 47(2): 563–574.
str(case0702) attach(case0702) # EXPLORATION plot(pH ~ Time) myLm <- lm(pH ~ Time) abline(myLm, col="blue", lwd=2) lines(lowess(Time,pH), col="red", lty=2, lwd=2) # Add scatterplot smoother plot(myLm, which=1) # Residual plot logTime <- log(Time) plot(pH ~ logTime) myLm2 <- lm(pH ~ logTime) abline(myLm2) plot(myLm2, which=1) ## PREDICTION BAND ABOUT REGRESSION LINE xToPredict <- seq(1,8,length=100) # sequence from 1 to 8 of length 100 logXToPredict <- log(xToPredict) newData <- data.frame(logTime = logXToPredict) myPredict <- predict(myLm2,newData, interval="prediction", level=.90) plot(pH ~ logTime) abline(myLm2) lines(myPredict[,3]~ logXToPredict, lty=2) lines(myPredict[,2] ~ logXToPredict, lty=2) # Find smallest time at which the upper endpoint of a 90% prediction # interval is less than or equal to 6: minTime <- min(xToPredict[myPredict[,3] <= 6.0]) minTime abline(v=log(minTime),col="red") # DISPLAY FOR PRESENTATION plot(pH ~ Time, xlab="Time After Slaughter (Hours); log scale", ylab="pH in Muscle", main="pH and Time after Slaughter for 10 Steers", log="x", pch=21, lwd=2, bg="green", cex=2 ) lines(xToPredict,myPredict[,1], col="blue", lwd=2) lines(xToPredict, myPredict[,3], lty=2, col="blue", lwd=2) lines(xToPredict, myPredict[,2], lty=2, col="blue", lwd=2) legend(3,7, c("Estimated Regression Line","90% Prediction Band"), lty=c(1,2), col="blue", lwd=c(2,2)) abline(h=6, lty=3, col="purple", lwd=2) text(1.5,6.05,"Desired pH", col="purple") lines(c(minTime,minTime),c(5,6.15), col="purple", lwd=2) text(minTime,6.2,"4.9 hours",col="purple",cex=1.25) detach(case0702)str(case0702) attach(case0702) # EXPLORATION plot(pH ~ Time) myLm <- lm(pH ~ Time) abline(myLm, col="blue", lwd=2) lines(lowess(Time,pH), col="red", lty=2, lwd=2) # Add scatterplot smoother plot(myLm, which=1) # Residual plot logTime <- log(Time) plot(pH ~ logTime) myLm2 <- lm(pH ~ logTime) abline(myLm2) plot(myLm2, which=1) ## PREDICTION BAND ABOUT REGRESSION LINE xToPredict <- seq(1,8,length=100) # sequence from 1 to 8 of length 100 logXToPredict <- log(xToPredict) newData <- data.frame(logTime = logXToPredict) myPredict <- predict(myLm2,newData, interval="prediction", level=.90) plot(pH ~ logTime) abline(myLm2) lines(myPredict[,3]~ logXToPredict, lty=2) lines(myPredict[,2] ~ logXToPredict, lty=2) # Find smallest time at which the upper endpoint of a 90% prediction # interval is less than or equal to 6: minTime <- min(xToPredict[myPredict[,3] <= 6.0]) minTime abline(v=log(minTime),col="red") # DISPLAY FOR PRESENTATION plot(pH ~ Time, xlab="Time After Slaughter (Hours); log scale", ylab="pH in Muscle", main="pH and Time after Slaughter for 10 Steers", log="x", pch=21, lwd=2, bg="green", cex=2 ) lines(xToPredict,myPredict[,1], col="blue", lwd=2) lines(xToPredict, myPredict[,3], lty=2, col="blue", lwd=2) lines(xToPredict, myPredict[,2], lty=2, col="blue", lwd=2) legend(3,7, c("Estimated Regression Line","90% Prediction Band"), lty=c(1,2), col="blue", lwd=c(2,2)) abline(h=6, lty=3, col="purple", lwd=2) text(1.5,6.05,"Desired pH", col="purple") lines(c(minTime,minTime),c(5,6.15), col="purple", lwd=2) text(minTime,6.2,"4.9 hours",col="purple",cex=1.25) detach(case0702)
The data are the numbers of reptile and amphibian species and the island areas for seven islands in the West Indies.
case0801case0801
A data frame with 7 observations on the following 2 variables.
area of island (in square miles)
number of reptile and amphibian species on island
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Wilson, E.O., 1992, The Diversity of Life, W. W. Norton, N.Y.
str(case0801) attach(case0801) ## EXPLORATION logSpecies <- log(Species) logArea <- log(Area) plot(logSpecies ~ logArea, xlab="Log of Island Area", ylab="Log of Number of Species", main="Number of Reptile and Amphibian Species on 7 Islands") myLm <- lm(logSpecies ~ logArea) abline(myLm) ## INFERENCE AND INTERPRETATION summary(myLm) slope <- myLm$coef[2] slopeConf <- confint(myLm,2) 100*(2^(slope)-1) # Back-transform estimated slope 100*(2^(slopeConf)-1) # Back-transform confidence interval # Interpretation: Associated with each doubling of island area is a 19% increase # in the median number of bird species (95% CI: 16% to 21% increase). ## DISPLAY FOR PRESENTATION plot(Species ~ Area, xlab="Island Area (Square Miles); Log Scale", ylab="Number of Species; Log Scale", main="Number of Reptile and Amphibian Species on 7 Islands", log="xy", pch=21, lwd=2, bg="green",cex=2 ) dummyArea <- c(min(Area),max(Area)) beta <- myLm$coef meanLogSpecies <- beta[1] + beta[2]*log(dummyArea) medianSpecies <- exp(meanLogSpecies) lines(medianSpecies ~ dummyArea,lwd=2,col="blue") island <- c(" Cuba"," Hispaniola"," Jamaica", " Puerto Rico", " Montserrat"," Saba"," Redonda") for (i in 1:7) { offset <- ifelse(Area[i] < 10000, -.2, 1.5) text(Area[i],Species[i],island[i],col="dark green",adj=offset,cex=.75) } detach(case0801)str(case0801) attach(case0801) ## EXPLORATION logSpecies <- log(Species) logArea <- log(Area) plot(logSpecies ~ logArea, xlab="Log of Island Area", ylab="Log of Number of Species", main="Number of Reptile and Amphibian Species on 7 Islands") myLm <- lm(logSpecies ~ logArea) abline(myLm) ## INFERENCE AND INTERPRETATION summary(myLm) slope <- myLm$coef[2] slopeConf <- confint(myLm,2) 100*(2^(slope)-1) # Back-transform estimated slope 100*(2^(slopeConf)-1) # Back-transform confidence interval # Interpretation: Associated with each doubling of island area is a 19% increase # in the median number of bird species (95% CI: 16% to 21% increase). ## DISPLAY FOR PRESENTATION plot(Species ~ Area, xlab="Island Area (Square Miles); Log Scale", ylab="Number of Species; Log Scale", main="Number of Reptile and Amphibian Species on 7 Islands", log="xy", pch=21, lwd=2, bg="green",cex=2 ) dummyArea <- c(min(Area),max(Area)) beta <- myLm$coef meanLogSpecies <- beta[1] + beta[2]*log(dummyArea) medianSpecies <- exp(meanLogSpecies) lines(medianSpecies ~ dummyArea,lwd=2,col="blue") island <- c(" Cuba"," Hispaniola"," Jamaica", " Puerto Rico", " Montserrat"," Saba"," Redonda") for (i in 1:7) { offset <- ifelse(Area[i] < 10000, -.2, 1.5) text(Area[i],Species[i],island[i],col="dark green",adj=offset,cex=.75) } detach(case0801)
In an industrial laboratory, under uniform conditions, batches of electrical insulating fluid were subjected to constant voltages until the insulating property of the fluids broke down. Seven different voltage levels were studied and the measured reponses were the times until breakdown.
case0802case0802
A data frame with 76 observations on the following 3 variables.
times until breakdown (in minutes)
voltage applied (in kV)
factor variable (group number)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Nelson, W.B., 1970, G.E. Co. Technical Report 71-C-011, Schenectady, N.Y.
str(case0802) attach(case0802) ## EXPLORATION plot(Time ~ Voltage) myLm <- lm(Time ~ Voltage) plot(myLm, which=1) # Residual plot logTime <- log(Time) plot(logTime ~ Voltage) myLm <- lm(logTime ~ Voltage) abline(myLm) plot(myLm,which=1) # Residual plot myOneWay <- lm(logTime ~ factor(Voltage)) anova(myLm, myOneWay) # Lack of fit test for simple regression (seems okay) ## INFERENCE AND INTERPREATION beta <- myLm$coef 100*(1 - exp(beta[2])) # Back-transform estimated slope 100*(1 - exp(confint(myLm,"Voltage"))) # Interpretation: Associated with each 1 kV increase in voltage is a 39.8% # decrease in median breakdown time (95% CI: 32.5% decrease to 46.3% decrease). ## DISPLAY FOR PRESENTATION options(scipen=50) # Do this to avoid scientific notation on y-axis plot(Time ~ Voltage, log="y", xlab="Voltage (kV)", ylab="Breakdown Time (min.); Log Scale", main="Breakdown Time of Insulating Fluid as a Function of Voltage Applied", pch=21, lwd=2, bg="green", cex=1.75 ) dummyVoltage <- c(min(Voltage),max(Voltage)) meanLogTime <- beta[1] + beta[2]*dummyVoltage medianTime <- exp(meanLogTime) lines(medianTime ~ dummyVoltage, lwd=2, col="blue") detach(case0802)str(case0802) attach(case0802) ## EXPLORATION plot(Time ~ Voltage) myLm <- lm(Time ~ Voltage) plot(myLm, which=1) # Residual plot logTime <- log(Time) plot(logTime ~ Voltage) myLm <- lm(logTime ~ Voltage) abline(myLm) plot(myLm,which=1) # Residual plot myOneWay <- lm(logTime ~ factor(Voltage)) anova(myLm, myOneWay) # Lack of fit test for simple regression (seems okay) ## INFERENCE AND INTERPREATION beta <- myLm$coef 100*(1 - exp(beta[2])) # Back-transform estimated slope 100*(1 - exp(confint(myLm,"Voltage"))) # Interpretation: Associated with each 1 kV increase in voltage is a 39.8% # decrease in median breakdown time (95% CI: 32.5% decrease to 46.3% decrease). ## DISPLAY FOR PRESENTATION options(scipen=50) # Do this to avoid scientific notation on y-axis plot(Time ~ Voltage, log="y", xlab="Voltage (kV)", ylab="Breakdown Time (min.); Log Scale", main="Breakdown Time of Insulating Fluid as a Function of Voltage Applied", pch=21, lwd=2, bg="green", cex=1.75 ) dummyVoltage <- c(min(Voltage),max(Voltage)) meanLogTime <- beta[1] + beta[2]*dummyVoltage medianTime <- exp(meanLogTime) lines(medianTime ~ dummyVoltage, lwd=2, col="blue") detach(case0802)
Meadowfoam is a small plant found growing in moist meadows of the US Pacific Northwest. Researchers reported the results from one study in a series designed to find out how to elevate meadowfoam production to a profitable crop. In a controlled growth chamber, they focused on the effects of two light–related factors: light intensity and the timeing of the onset of the ligth treatment.
case0901case0901
A data frame with 24 observations on the following 3 variables.
average number of flowers per meadowfoam plant
time light intensity regiments started; 1=Late, 2=Early
light intensity (in mol/m/sec)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(case0901) attach(case0901) ## EXPLORATION plot(Flowers ~ Intensity, pch=ifelse(Time ==1, 19, 21)) myLm <- lm(Flowers ~ Intensity + factor(Time) + Intensity:factor(Time)) plot(myLm, which=1) summary(myLm) # Note p-value for interaction term # INFERENCE myLm2 <- lm(Flowers ~ Intensity + factor(Time)) summary(myLm2) confint(myLm2) # DISPLAY FOR PRESENTATION plot(Flowers ~ jitter(Intensity,.3), xlab=expression("Light Intensity ("*mu*"mol/"*m^2*"/sec)"), # Include symbols ylab="Average Number of Flowers per Plant", main="Effect of Light Intensity and Timing on Meadowfoam Flowering", pch=ifelse(Time ==1, 21, 22), bg=ifelse(Time==1, "orange","green"), cex=1.7, lwd=2) beta <- myLm2$coef abline(beta[1],beta[2],lwd=2, lty=2) abline(beta[1]+beta[3],beta[2],lwd=2,lty=3) legend(700,79,c("Early Start","Late Start"), pch=c(22,21),lwd=2,pt.bg=c("green","orange"),pt.cex=1.7,lty=c(3,2)) detach(case0901)str(case0901) attach(case0901) ## EXPLORATION plot(Flowers ~ Intensity, pch=ifelse(Time ==1, 19, 21)) myLm <- lm(Flowers ~ Intensity + factor(Time) + Intensity:factor(Time)) plot(myLm, which=1) summary(myLm) # Note p-value for interaction term # INFERENCE myLm2 <- lm(Flowers ~ Intensity + factor(Time)) summary(myLm2) confint(myLm2) # DISPLAY FOR PRESENTATION plot(Flowers ~ jitter(Intensity,.3), xlab=expression("Light Intensity ("*mu*"mol/"*m^2*"/sec)"), # Include symbols ylab="Average Number of Flowers per Plant", main="Effect of Light Intensity and Timing on Meadowfoam Flowering", pch=ifelse(Time ==1, 21, 22), bg=ifelse(Time==1, "orange","green"), cex=1.7, lwd=2) beta <- myLm2$coef abline(beta[1],beta[2],lwd=2, lty=2) abline(beta[1]+beta[3],beta[2],lwd=2,lty=3) legend(700,79,c("Early Start","Late Start"), pch=c(22,21),lwd=2,pt.bg=c("green","orange"),pt.cex=1.7,lty=c(3,2)) detach(case0901)
The data are the average values of brain weight, body weight, gestation lengths (length of pregnancy) and litter size for 96 species of mammals.
case0902case0902
A data frame with 96 observations on the following 5 variables.
species
average brain weight (in grams)
average body weight (in kilograms)
gestation period (in days)
average litter size
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(case0902) attach(case0902) ## EXPLORATION myMatrix <- cbind(Brain, Body, Litter, Gestation) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix, # Matrix of scatterplots smooth=FALSE, # Omit scatterplot smoother on plots diagonal="histogram") # Draw histograms on diagonals } myLm <- lm(Brain ~ Body + Litter + Gestation) plot(myLm, which=1) logBrain <- log(Brain) logBody <- log(Body) logGestation <- log(Gestation) myMatrix2 <- cbind(logBrain,logBody,Litter, logGestation) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix2, smooth=FALSE, diagonal="histogram") } myLm2 <- lm(logBrain ~ logBody + Litter + logGestation) plot(myLm2,which=1) # Residual plot. if(require(car)){ # Use the car library crPlots(myLm2) # Partial residual plots (Sleuth Ch.11) } plot(logBrain ~ logBody) identify(logBrain ~ logBody,labels=Species) # Identify points on scatterplot # Place the cursor over a point of interest, then left-click. # Continue with other points if desired. When finished, pres Esc. ## INFERENCE summary(myLm2) confint(myLm2) # DISPLAYS FOR PRESENTATION myLm3 <- lm(logBrain ~ logBody + logGestation) beta <- myLm3$coef logBrainAdjusted <- logBrain - beta[2]*logBody y <- exp(logBrainAdjusted) ymod <- 100*y/median(y) plot(ymod ~ Gestation, log="xy", xlab="Average Gestation Length (Days); Log Scale", ylab="Brain Weight Adjusted for Body Weight, as a Percentage of the Median", main="Brain Weight Adjusted for Body Weight, Versus Gestation Length, for 96 Mammal Species", pch=21,bg="green",cex=1.3) identify(ymod ~ Gestation,labels=Species, cex=.7) # Identify points, as desired # Press Esc to complete identify. abline(h=100,lty=2) # Draw horizontal line at 100% myLm4 <- lm(logBrain ~ logBody + Litter) beta <- myLm4$coef logBrainAdjusted <- logBrain - beta[2]*logBody y2 <- exp(logBrainAdjusted) y2mod <- 100*y2/median(y2) plot(y2mod ~ Litter, log="y", xlab="Average Litter Size", ylab="Brain Weight Adjusted for Body Weight, as a Percentage of the Median", main="Brain Weight Adjusted for Body Weight, Versus Litter Size, for 96 Mammal Species", pch=21,bg="green",cex=1.3) identify(y2mod ~ Litter,labels=Species, cex=.7) abline(h=100,lty=2) detach(case0902)str(case0902) attach(case0902) ## EXPLORATION myMatrix <- cbind(Brain, Body, Litter, Gestation) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix, # Matrix of scatterplots smooth=FALSE, # Omit scatterplot smoother on plots diagonal="histogram") # Draw histograms on diagonals } myLm <- lm(Brain ~ Body + Litter + Gestation) plot(myLm, which=1) logBrain <- log(Brain) logBody <- log(Body) logGestation <- log(Gestation) myMatrix2 <- cbind(logBrain,logBody,Litter, logGestation) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix2, smooth=FALSE, diagonal="histogram") } myLm2 <- lm(logBrain ~ logBody + Litter + logGestation) plot(myLm2,which=1) # Residual plot. if(require(car)){ # Use the car library crPlots(myLm2) # Partial residual plots (Sleuth Ch.11) } plot(logBrain ~ logBody) identify(logBrain ~ logBody,labels=Species) # Identify points on scatterplot # Place the cursor over a point of interest, then left-click. # Continue with other points if desired. When finished, pres Esc. ## INFERENCE summary(myLm2) confint(myLm2) # DISPLAYS FOR PRESENTATION myLm3 <- lm(logBrain ~ logBody + logGestation) beta <- myLm3$coef logBrainAdjusted <- logBrain - beta[2]*logBody y <- exp(logBrainAdjusted) ymod <- 100*y/median(y) plot(ymod ~ Gestation, log="xy", xlab="Average Gestation Length (Days); Log Scale", ylab="Brain Weight Adjusted for Body Weight, as a Percentage of the Median", main="Brain Weight Adjusted for Body Weight, Versus Gestation Length, for 96 Mammal Species", pch=21,bg="green",cex=1.3) identify(ymod ~ Gestation,labels=Species, cex=.7) # Identify points, as desired # Press Esc to complete identify. abline(h=100,lty=2) # Draw horizontal line at 100% myLm4 <- lm(logBrain ~ logBody + Litter) beta <- myLm4$coef logBrainAdjusted <- logBrain - beta[2]*logBody y2 <- exp(logBrainAdjusted) y2mod <- 100*y2/median(y2) plot(y2mod ~ Litter, log="y", xlab="Average Litter Size", ylab="Brain Weight Adjusted for Body Weight, as a Percentage of the Median", main="Brain Weight Adjusted for Body Weight, Versus Litter Size, for 96 Mammal Species", pch=21,bg="green",cex=1.3) identify(y2mod ~ Litter,labels=Species, cex=.7) abline(h=100,lty=2) detach(case0902)
In 1609 Galileo proved mathematically that the trajectory of a body falling with a horizontal velocity component is a parabola. His search for an experimental setting in which horizontal motion was not affected appreciably (to study inertia) let him to construct a certain apparatus. The data comes from one of his experiments.
case1001case1001
A data frame with 7 observations on the following 2 variables.
horizontal distances (in punti)
initial height (in punti)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(case1001) attach(case1001) ## EXPLORATION plot(Distance ~ Height) myLm <- lm(Distance ~ Height) plot(myLm, which=1) height2 <- Height^2 myLm2 <- lm(Distance ~ Height + height2) plot(myLm2, which=1) summary(myLm2) # Note p-value for quadratic term (it's small) height3 <- Height^3 myLm3 <- update(myLm2, ~ . + height3) plot(myLm3,which=1) summary(myLm3) # Note p-value for cubic term (it's small) height4 <- Height^4 myLm4 <- update(myLm3, ~ . + height4) summary(myLm4) # Note p-value for quartic term (it's not small) ## DISPLAY FOR PRESENTATION plot(Distance ~ Height, xlab="Initial Height (Punti)", ylab="Horizontal Distance Traveled (Punti)", main="Galileo's Falling Body Experiment", pch=21, bg="green", lwd=2, cex=2) dummyHeight <- seq(min(Height),max(Height),length=100) betaQ <- myLm2$coef quadraticCurve <- betaQ[1] + betaQ[2]*dummyHeight + betaQ[3]*dummyHeight^2 lines(quadraticCurve ~ dummyHeight,col="blue",lwd=3) betaC <- myLm3$coef # coefficients of cubic model cubicCurve <- betaC[1] + betaC[2]*dummyHeight + betaC[3]*dummyHeight^2 + betaC[4]*dummyHeight^3 lines(cubicCurve ~ dummyHeight,lty=3,col="red",lwd=3) legend(590,290,legend=c(expression("Quadratic Fit "*R^2*" = 99.0%"), expression("Cubic Fit "*R^2*" = 99.9%")), lty=c(1,3),col=c("blue","red"), lwd=c(3,3)) detach(case1001)str(case1001) attach(case1001) ## EXPLORATION plot(Distance ~ Height) myLm <- lm(Distance ~ Height) plot(myLm, which=1) height2 <- Height^2 myLm2 <- lm(Distance ~ Height + height2) plot(myLm2, which=1) summary(myLm2) # Note p-value for quadratic term (it's small) height3 <- Height^3 myLm3 <- update(myLm2, ~ . + height3) plot(myLm3,which=1) summary(myLm3) # Note p-value for cubic term (it's small) height4 <- Height^4 myLm4 <- update(myLm3, ~ . + height4) summary(myLm4) # Note p-value for quartic term (it's not small) ## DISPLAY FOR PRESENTATION plot(Distance ~ Height, xlab="Initial Height (Punti)", ylab="Horizontal Distance Traveled (Punti)", main="Galileo's Falling Body Experiment", pch=21, bg="green", lwd=2, cex=2) dummyHeight <- seq(min(Height),max(Height),length=100) betaQ <- myLm2$coef quadraticCurve <- betaQ[1] + betaQ[2]*dummyHeight + betaQ[3]*dummyHeight^2 lines(quadraticCurve ~ dummyHeight,col="blue",lwd=3) betaC <- myLm3$coef # coefficients of cubic model cubicCurve <- betaC[1] + betaC[2]*dummyHeight + betaC[3]*dummyHeight^2 + betaC[4]*dummyHeight^3 lines(cubicCurve ~ dummyHeight,lty=3,col="red",lwd=3) legend(590,290,legend=c(expression("Quadratic Fit "*R^2*" = 99.0%"), expression("Cubic Fit "*R^2*" = 99.9%")), lty=c(1,3),col=c("blue","red"), lwd=c(3,3)) detach(case1001)
The data are on in–flight energy expenditure and body mass from 20 energy studies on three types of flying vertebrates: echolocating bats, non–echolocating bats and non–echolocating birds.
case1002case1002
A data frame with 20 observations on the following 3 variables.
mass (in grams)
a factor with 3 levels indicating the type of flying vertebrate: non-echolocating bats, non-echolocating birds, echolocating bats
in–flight energy expenditure (in W)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Speakman, J.R. and Racey, P.A. (1991). No cost of Echolocation for Bats in Flight, Nature 350: 421–423.
str(case1002) attach(case1002) ## EXPLORATION plot(Energy~Mass, case1002, log="xy", xlab = "Body Mass (g) (log scale)", ylab = "Energy Expenditure (W) (log scale)", pch = ifelse(Type=="echolocating bats", 19, ifelse(Type=="non-echolocating birds", 21, 24))) legend(7, 50, pch=c(24, 21, 19), c("Non-echolocating bats", "Non-echolocating birds","Echolocating bats")) logEnergy <- log(Energy) logMass <- log(Mass) myLm2 <- lm(logEnergy ~ logMass + Type + logMass:Type) plot(myLm2, which=1) myLm3 <- update(myLm2, ~ . - logMass:Type) anova(myLm3, myLm2) # Test for interaction with extra ss F-test ## INFERENCE AND INTERPRETATION myLm4 <- update(myLm3, ~ . - Type) # Reduced model...with no effect of Type anova(myLm4, myLm3) # Test for Type effect myType <- factor(Type, levels=c("non-echolocating bats","echolocating bats","non-echolocating birds")) myLm3a <- lm(logEnergy ~ logMass + myType) summary(myLm3a) 100*(exp(myLm3a$coef[3]) - 1) 100*(exp(confint(myLm3a,3))-1) # Conclusion: Adjusted for body mass, the median energy expenditure for # echo-locating bats exceeds that for echo-locating bats by an estimated # 8.2% (95% confidence interval: 29.6% LESS to 66.3% MORE) # DISPLAY FOR PRESENTATION myPlotCode <- ifelse(Type=="non-echolocating birds",24,21) myPointColor <- ifelse(Type=="echolocating bats","green","white") plot(Energy ~ Mass, log="xy", xlab="Body Mass (g); Log Scale ", ylab="In-Flight Energy Expenditure (W); Log Scale", main="In-Flight Energy Expenditure Study", pch=myPlotCode,bg=myPointColor,lwd=2, cex=1.5) dummyMass <- seq(5,800,length=50) beta <- myLm3$coef curve1 <- exp(beta[1] + beta[2]*log(dummyMass)) curve2 <- exp(beta[1] + beta[2]*log(dummyMass) + beta[3]) curve3 <- exp(beta[1] + beta[2]*log(dummyMass) + beta[4]) lines(curve1 ~ dummyMass) lines(curve2 ~ dummyMass, lty=2) lines(curve3 ~ dummyMass, lty=3) legend(100,3, c("Echolocating Bats","Non-Echolocating Bats","Non-Echolocating Birds"), pch=c(21,21,24),lwd=2,pt.cex=c(1.5,1.5,1.5),pt.lwd=c(2,2,2), pt.bg=c("green","white","white"),lty=c(1,2,3)) detach(case1002)str(case1002) attach(case1002) ## EXPLORATION plot(Energy~Mass, case1002, log="xy", xlab = "Body Mass (g) (log scale)", ylab = "Energy Expenditure (W) (log scale)", pch = ifelse(Type=="echolocating bats", 19, ifelse(Type=="non-echolocating birds", 21, 24))) legend(7, 50, pch=c(24, 21, 19), c("Non-echolocating bats", "Non-echolocating birds","Echolocating bats")) logEnergy <- log(Energy) logMass <- log(Mass) myLm2 <- lm(logEnergy ~ logMass + Type + logMass:Type) plot(myLm2, which=1) myLm3 <- update(myLm2, ~ . - logMass:Type) anova(myLm3, myLm2) # Test for interaction with extra ss F-test ## INFERENCE AND INTERPRETATION myLm4 <- update(myLm3, ~ . - Type) # Reduced model...with no effect of Type anova(myLm4, myLm3) # Test for Type effect myType <- factor(Type, levels=c("non-echolocating bats","echolocating bats","non-echolocating birds")) myLm3a <- lm(logEnergy ~ logMass + myType) summary(myLm3a) 100*(exp(myLm3a$coef[3]) - 1) 100*(exp(confint(myLm3a,3))-1) # Conclusion: Adjusted for body mass, the median energy expenditure for # echo-locating bats exceeds that for echo-locating bats by an estimated # 8.2% (95% confidence interval: 29.6% LESS to 66.3% MORE) # DISPLAY FOR PRESENTATION myPlotCode <- ifelse(Type=="non-echolocating birds",24,21) myPointColor <- ifelse(Type=="echolocating bats","green","white") plot(Energy ~ Mass, log="xy", xlab="Body Mass (g); Log Scale ", ylab="In-Flight Energy Expenditure (W); Log Scale", main="In-Flight Energy Expenditure Study", pch=myPlotCode,bg=myPointColor,lwd=2, cex=1.5) dummyMass <- seq(5,800,length=50) beta <- myLm3$coef curve1 <- exp(beta[1] + beta[2]*log(dummyMass)) curve2 <- exp(beta[1] + beta[2]*log(dummyMass) + beta[3]) curve3 <- exp(beta[1] + beta[2]*log(dummyMass) + beta[4]) lines(curve1 ~ dummyMass) lines(curve2 ~ dummyMass, lty=2) lines(curve3 ~ dummyMass, lty=3) legend(100,3, c("Echolocating Bats","Non-Echolocating Bats","Non-Echolocating Birds"), pch=c(21,21,24),lwd=2,pt.cex=c(1.5,1.5,1.5),pt.lwd=c(2,2,2), pt.bg=c("green","white","white"),lty=c(1,2,3)) detach(case1002)
These data were collected on 18 women and 14 men to investigate a certain theory on why women exhibit a lower tolerance for alcohol and develop alcohol–related liver disease more readily than men.
case1101case1101
A data frame with 32 observations on the following 5 variables.
subject number in the study
first–pass metabolism of alcohol in the stomach (in mmol/liter-hour)
gastric alcohol dehydrogenase activity in the stomach
(in mol/min/g of tissue)
sex of the subject
whether the subject is alcoholic or not
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(case1101) attach(case1101) ## EXPLORATION library(lattice) xyplot(Metabol~Gastric|Sex*Alcohol, case1101) myPch <- ifelse(Sex=="Female",24,21) myBg <- ifelse(Alcohol=="Alcoholic","gray","white") plot(Metabol~Gastric, pch=myPch,bg=myBg,cex=1.5) legend(1,12, pch=c(24,24,21,21), pt.cex=c(1.5,1.5,1.5,1.5), pt.bg=c("white","gray", "white", "gray"), c("Non-alcoholic Females", "Alcoholic Females", "Non-alcoholic Males", "Alcoholic Males")) identify(Metabol ~ Gastric) # Left click on outliers to show case number; Esc when finished. myLm1 <- lm(Metabol ~ Gastric + Sex + Gastric:Sex) plot(myLm1, which=1) plot(myLm1, which=4) # Show Cook's Distance; note cases 31 and 32. plot(myLm1, which=5) # Note leverage and studentized residual for cases 31 and 32. subject <- 1:32 # Create ID number from 1 to 32 # Refit model without cases 31 and 32: myLm2 <- update(myLm1, ~ ., subset = (subject !=31 & subject !=32)) plot(myLm2,which=1) plot(myLm2,which=4) plot(myLm2,which=5) summary(myLm1) summary(myLm2) # Significance of interaction terms hinges on cases 31 and 32. myLm3 <- update(myLm2, ~ . - Gastric:Sex) #Drop interaction (without 31,32). summary(myLm3) if(require(car)){ # Use the car library crPlots(myLm3) # Show partial residual (component + residual) plots. } ## INFERENCE AND INTERPRETATION summary(myLm3) confint(myLm3,2:3) ## DISPLAY FOR PRESENTATION myCol <- ifelse(Sex=="Male","blue","red") plot(Metabol ~ Gastric, xlab=expression("Gastric Alcohol Dehydrogenase Activity in Stomach ("*mu*"mol/min/g of Tissue)"), ylab="First-pass Metabolism in the Stomach (mmol/liter-hour)", main="First-Pass Alcohol Metabolism and Enzyme Activity for 18 Females and 14 Males", pch=myPch, bg=myBg,cex=1.75, col=myCol, lwd=1) legend(0.8,12.2, c("Females", "Males"), lty=c(1,2), pch=c(24,21), pt.cex=c(1.75,1.75), col=c("red", "blue")) dummyGastric <- seq(min(Gastric),3,length=100) beta <- myLm3$coef curveF <- beta[1] + beta[2]*dummyGastric curveM <- beta[1] + beta[2]*dummyGastric + beta[3] lines(curveF ~ dummyGastric, col="red") lines(curveM ~ dummyGastric, col="blue",lty=2) text(.8,10,"gray indicates alcoholic",cex = .8, adj=0) detach(case1101)str(case1101) attach(case1101) ## EXPLORATION library(lattice) xyplot(Metabol~Gastric|Sex*Alcohol, case1101) myPch <- ifelse(Sex=="Female",24,21) myBg <- ifelse(Alcohol=="Alcoholic","gray","white") plot(Metabol~Gastric, pch=myPch,bg=myBg,cex=1.5) legend(1,12, pch=c(24,24,21,21), pt.cex=c(1.5,1.5,1.5,1.5), pt.bg=c("white","gray", "white", "gray"), c("Non-alcoholic Females", "Alcoholic Females", "Non-alcoholic Males", "Alcoholic Males")) identify(Metabol ~ Gastric) # Left click on outliers to show case number; Esc when finished. myLm1 <- lm(Metabol ~ Gastric + Sex + Gastric:Sex) plot(myLm1, which=1) plot(myLm1, which=4) # Show Cook's Distance; note cases 31 and 32. plot(myLm1, which=5) # Note leverage and studentized residual for cases 31 and 32. subject <- 1:32 # Create ID number from 1 to 32 # Refit model without cases 31 and 32: myLm2 <- update(myLm1, ~ ., subset = (subject !=31 & subject !=32)) plot(myLm2,which=1) plot(myLm2,which=4) plot(myLm2,which=5) summary(myLm1) summary(myLm2) # Significance of interaction terms hinges on cases 31 and 32. myLm3 <- update(myLm2, ~ . - Gastric:Sex) #Drop interaction (without 31,32). summary(myLm3) if(require(car)){ # Use the car library crPlots(myLm3) # Show partial residual (component + residual) plots. } ## INFERENCE AND INTERPRETATION summary(myLm3) confint(myLm3,2:3) ## DISPLAY FOR PRESENTATION myCol <- ifelse(Sex=="Male","blue","red") plot(Metabol ~ Gastric, xlab=expression("Gastric Alcohol Dehydrogenase Activity in Stomach ("*mu*"mol/min/g of Tissue)"), ylab="First-pass Metabolism in the Stomach (mmol/liter-hour)", main="First-Pass Alcohol Metabolism and Enzyme Activity for 18 Females and 14 Males", pch=myPch, bg=myBg,cex=1.75, col=myCol, lwd=1) legend(0.8,12.2, c("Females", "Males"), lty=c(1,2), pch=c(24,21), pt.cex=c(1.75,1.75), col=c("red", "blue")) dummyGastric <- seq(min(Gastric),3,length=100) beta <- myLm3$coef curveF <- beta[1] + beta[2]*dummyGastric curveM <- beta[1] + beta[2]*dummyGastric + beta[3] lines(curveF ~ dummyGastric, col="red") lines(curveM ~ dummyGastric, col="blue",lty=2) text(.8,10,"gray indicates alcoholic",cex = .8, adj=0) detach(case1101)
The human brain is protected from bacteria and toxins, which course through the blood–stream, by a single layer of cells called the blood–brain barrier. These data come from an experiment (on rats, which possess a similar barrier) to study a method of disrupting the barrier by infusing a solution of concentrated sugars.
case1102case1102
A data frame with 34 observations on the following 9 variables.
Brain tumor count (per gm)
Liver count (per gm)
Sacrifice time (in hours)
Treatment received
Days post inoculation
Sex of the rat
Initial weight (in grams)
Weight loss (in grams)
Tumor weight (in 10 grams)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(case1102) attach(case1102) ## EXPLORATION logRatio <- log(Brain/Liver) logTime <- log(Time) myMatrix <- cbind (logRatio, Days, Weight, Loss, Tumor, logTime) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix,groups=Treatment, smooth=FALSE, diagonal="histogram", col=c("green","blue"), pch=c(16,17), cex=1.5) } myLm1 <- lm(logRatio ~ Treatment + logTime + Days + Sex + Weight + Loss + Tumor) plot(myLm1, which=1) if(require(car)){ # Use the car library crPlots(myLm1) # Draw partial resdual plots. } myLm2 <- lm(logRatio ~ Treatment + factor(Time) + Days + Sex + Weight + Loss + Tumor) # Include Time as a factor. anova(myLm1,myLm2) if(require(car)){ # Use the car library crPlots(myLm2) # Draw partial resdual plots. } summary(myLm2) # Use backard elimination myLm3 <- update(myLm2, ~ . - Days) summary(myLm3) myLm4 <- update(myLm3, ~ . - Sex) summary(myLm4) myLm5 <- update(myLm4, ~ . - Weight) summary(myLm5) myLm6 <- update(myLm5, ~ . - Tumor) summary(myLm6) myLm7 <- update(myLm6, ~ . - Loss) summary(myLm7) # Final model for inference ## INFERENCE AND INTERPRETATION myTreatment <- factor(Treatment,levels=c("NS","BD")) # Change level ordering myLm7a <- lm(logRatio ~ factor(Time) + myTreatment) summary(myLm7a) beta <- myLm7a$coef exp(beta[5]) exp(confint(myLm7a,5)) # Interpetation: The median ratio of brain to liver tumor counts for barrier- # disrupted rats is estimated to be 2.2 times the median ratio for control rats # (95% CI: 1.5 times to 3.2 times as large). ## DISPLAY FOR PRESENTATION ratio <- Brain/Liver jTime <- exp(jitter(logTime,.2)) # Back-transform a jittered version of logTime plot(ratio ~ jTime, log="xy", xlab="Sacrifice Time (Hours), jittered; Log Scale", ylab="Effectiveness: Brain Tumor Count Relative To Liver Tumor Count; Log Scale", main="Blood Brain Barrier Disruption Effectiveness in 34 Rats", pch= ifelse(Treatment=="BD",21,24), bg=ifelse(Treatment=="BD","green","orange"), lwd=2, cex=2) dummyTime <- c(0.5, 3, 24, 72) controlTerm <- beta[1] + beta[2]*(dummyTime==3) + beta[3]*(dummyTime==24) + beta[4]*(dummyTime==72) controlCurve <- exp(controlTerm) lines(controlCurve ~ dummyTime, lty=1,lwd=2) BDTerm <- controlTerm + beta[5] BDCurve <- exp(BDTerm) lines(BDCurve ~ dummyTime,lty=2,lwd=2) legend(0.5,10,c("Barrier disruption","Saline control"),pch=c(21,22), pt.bg=c("green","orange"),pt.lwd=c(2,2),pt.cex=c(2,2), lty=c(2,1),lwd=c(2,2)) detach(case1102)str(case1102) attach(case1102) ## EXPLORATION logRatio <- log(Brain/Liver) logTime <- log(Time) myMatrix <- cbind (logRatio, Days, Weight, Loss, Tumor, logTime) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix,groups=Treatment, smooth=FALSE, diagonal="histogram", col=c("green","blue"), pch=c(16,17), cex=1.5) } myLm1 <- lm(logRatio ~ Treatment + logTime + Days + Sex + Weight + Loss + Tumor) plot(myLm1, which=1) if(require(car)){ # Use the car library crPlots(myLm1) # Draw partial resdual plots. } myLm2 <- lm(logRatio ~ Treatment + factor(Time) + Days + Sex + Weight + Loss + Tumor) # Include Time as a factor. anova(myLm1,myLm2) if(require(car)){ # Use the car library crPlots(myLm2) # Draw partial resdual plots. } summary(myLm2) # Use backard elimination myLm3 <- update(myLm2, ~ . - Days) summary(myLm3) myLm4 <- update(myLm3, ~ . - Sex) summary(myLm4) myLm5 <- update(myLm4, ~ . - Weight) summary(myLm5) myLm6 <- update(myLm5, ~ . - Tumor) summary(myLm6) myLm7 <- update(myLm6, ~ . - Loss) summary(myLm7) # Final model for inference ## INFERENCE AND INTERPRETATION myTreatment <- factor(Treatment,levels=c("NS","BD")) # Change level ordering myLm7a <- lm(logRatio ~ factor(Time) + myTreatment) summary(myLm7a) beta <- myLm7a$coef exp(beta[5]) exp(confint(myLm7a,5)) # Interpetation: The median ratio of brain to liver tumor counts for barrier- # disrupted rats is estimated to be 2.2 times the median ratio for control rats # (95% CI: 1.5 times to 3.2 times as large). ## DISPLAY FOR PRESENTATION ratio <- Brain/Liver jTime <- exp(jitter(logTime,.2)) # Back-transform a jittered version of logTime plot(ratio ~ jTime, log="xy", xlab="Sacrifice Time (Hours), jittered; Log Scale", ylab="Effectiveness: Brain Tumor Count Relative To Liver Tumor Count; Log Scale", main="Blood Brain Barrier Disruption Effectiveness in 34 Rats", pch= ifelse(Treatment=="BD",21,24), bg=ifelse(Treatment=="BD","green","orange"), lwd=2, cex=2) dummyTime <- c(0.5, 3, 24, 72) controlTerm <- beta[1] + beta[2]*(dummyTime==3) + beta[3]*(dummyTime==24) + beta[4]*(dummyTime==72) controlCurve <- exp(controlTerm) lines(controlCurve ~ dummyTime, lty=1,lwd=2) BDTerm <- controlTerm + beta[5] BDCurve <- exp(BDTerm) lines(BDCurve ~ dummyTime,lty=2,lwd=2) legend(0.5,10,c("Barrier disruption","Saline control"),pch=c(21,22), pt.bg=c("green","orange"),pt.lwd=c(2,2),pt.cex=c(2,2), lty=c(2,1),lwd=c(2,2)) detach(case1102)
Data on the average SAT scores for US states in 1982 and possible associated factors.
case1201case1201
A data frame with 50 observations on the following 8 variables.
US state
state averages of the total SAT (verbal + quantitative) scores
the percentage of the total eligible students (high school seniors) in the state who took the exam
the median income of families of test–takers (in hundreds of dollars)
the average number of years that the test–takers had formal studies in social sciences, natural sciences and humanities
the percentage of the test–takers who attended public secondary schools
the total state expenditure on secondary schools (in hundreds of dollars per student)
the median percentile ranking of the test–takers within their secondary school classes
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(case1201) attach(case1201) ## EXPLORATION logTakers <- log(Takers) myMatrix <- cbind(SAT, logTakers,Income, Years, Public, Expend, Rank) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix, diagonal="histogram", smooth=FALSE) } State[Public < 50] # Identify state with low Public (Louisiana) State[Expend > 40] # Alaska myLm1 <- lm(SAT ~ logTakers + Income+ Years + Public + Expend + Rank) plot(myLm1,which=1) plot(myLm1,which=4) # Cook's Distance State[29] # Identify State number 29? ([1] Alaska) plot(myLm1,which=5) if(require(car)){ # Use the car library crPlots(myLm1) # Partial residual plot } myLm2 <- update(myLm1, ~ . ,subset=(State != "Alaska")) plot(myLm2,which=1) plot(myLm2,which=4) if(require(car)){ # Use the car library crPlots(myLm2) # Partial residual plot } ## RANK STATES ON SAT SCORES, ADJUSTED FOR Takers AND Rank myLm3 <- lm(SAT ~ logTakers + Rank) myResiduals <- myLm3$res myOrder <- order(myResiduals) State[myOrder] ## DISPLAY FOR PRESENTATION dotchart(myResiduals[myOrder], labels=State[myOrder], xlab="SAT Scores, Adjusted for Percent Takers and HS Ranks (Deviation From Average)", main="States Ranked by Adjusted SAT Scores", bg="green", cex=.8) abline(v=0, col="gray") ## VARIABLE SELECTION (FOR RANKING STATES AFTER ACCOUNTING FOR ALL VARIABLES) expendSquared <- Expend^2 if(require(leaps)){ # Use the leaps library mySubsets <- regsubsets(SAT ~ logTakers + Income+ Years + Public + Expend + Rank + expendSquared, nvmax=8, data=case1201, subset=(State != "Alaska")) mySummary <- summary(mySubsets) p <- apply(mySummary$which, 1, sum) plot(p, mySummary$bic, ylab = "BIC") cbind(p,mySummary$bic) mySummary$which[4,] myLm4 <- lm(SAT ~ logTakers + Years + Expend + Rank, subset=(State != "Alaska")) summary(myLm4) ## DISPLAY FOR PRESENTATION myResiduals2 <- myLm4$res myOrder2 <- order(myResiduals2) newState <- State[State != "Alaska"] newState[myOrder2] dotchart(myResiduals2[myOrder2], labels=State[myOrder2], xlab="Adjusted SAT Scores (Deviation From Average Adjusted Value)", main=paste("States Ranked by SAT Scores Adjusted for Demographics", "of Takers and Education Expenditure", sep = " "), bg="green", cex = .8) abline(v=0, col="gray") } detach(case1201)str(case1201) attach(case1201) ## EXPLORATION logTakers <- log(Takers) myMatrix <- cbind(SAT, logTakers,Income, Years, Public, Expend, Rank) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix, diagonal="histogram", smooth=FALSE) } State[Public < 50] # Identify state with low Public (Louisiana) State[Expend > 40] # Alaska myLm1 <- lm(SAT ~ logTakers + Income+ Years + Public + Expend + Rank) plot(myLm1,which=1) plot(myLm1,which=4) # Cook's Distance State[29] # Identify State number 29? ([1] Alaska) plot(myLm1,which=5) if(require(car)){ # Use the car library crPlots(myLm1) # Partial residual plot } myLm2 <- update(myLm1, ~ . ,subset=(State != "Alaska")) plot(myLm2,which=1) plot(myLm2,which=4) if(require(car)){ # Use the car library crPlots(myLm2) # Partial residual plot } ## RANK STATES ON SAT SCORES, ADJUSTED FOR Takers AND Rank myLm3 <- lm(SAT ~ logTakers + Rank) myResiduals <- myLm3$res myOrder <- order(myResiduals) State[myOrder] ## DISPLAY FOR PRESENTATION dotchart(myResiduals[myOrder], labels=State[myOrder], xlab="SAT Scores, Adjusted for Percent Takers and HS Ranks (Deviation From Average)", main="States Ranked by Adjusted SAT Scores", bg="green", cex=.8) abline(v=0, col="gray") ## VARIABLE SELECTION (FOR RANKING STATES AFTER ACCOUNTING FOR ALL VARIABLES) expendSquared <- Expend^2 if(require(leaps)){ # Use the leaps library mySubsets <- regsubsets(SAT ~ logTakers + Income+ Years + Public + Expend + Rank + expendSquared, nvmax=8, data=case1201, subset=(State != "Alaska")) mySummary <- summary(mySubsets) p <- apply(mySummary$which, 1, sum) plot(p, mySummary$bic, ylab = "BIC") cbind(p,mySummary$bic) mySummary$which[4,] myLm4 <- lm(SAT ~ logTakers + Years + Expend + Rank, subset=(State != "Alaska")) summary(myLm4) ## DISPLAY FOR PRESENTATION myResiduals2 <- myLm4$res myOrder2 <- order(myResiduals2) newState <- State[State != "Alaska"] newState[myOrder2] dotchart(myResiduals2[myOrder2], labels=State[myOrder2], xlab="Adjusted SAT Scores (Deviation From Average Adjusted Value)", main=paste("States Ranked by SAT Scores Adjusted for Demographics", "of Takers and Education Expenditure", sep = " "), bg="green", cex = .8) abline(v=0, col="gray") } detach(case1201)
Data on employees from one job category (skilled, entry–level clerical) of a bank that was sued for sex discrimination. The data are on 32 male and 61 female employees, hired between 1965 and 1975.
case1202case1202
A data frame with 93 observations on the following 7 variables.
Annual salary at time of hire
Salary as of March 1975
Sex of employee
Seniority (months since first hired)
Age of employee (in months)
Education (in years)
Work experience prior to employment with the bank (months)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Roberts, H.V. (1979). Harris Trust and Savings Bank: An Analysis of Employee Compensation, Report 7946, Center for Mathematical Studies in Business and Economics, University of Chicago Graduate School of Business.
str(case1202) attach(case1202) ## EXPLORATION logSal <- log(Bsal) myMatrix <- cbind (logSal, Senior,Age, Educ, Exper) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix, smooth=FALSE, diagonal="histogram", groups=Sex, col=c("red","blue") ) } myLm1 <- lm(logSal ~ Senior + Age + Educ + Exper + Sex) plot(myLm1, which=1) plot(myLm1, which=4) # Cook's Distance if(require(car)){ # Use the car library crPlots(myLm1) # Partial residual plots } ageSquared <- Age^2 ageCubed <- Age^3 experSquared <- Exper^2 experCubed <- Exper^3 myLm2 <- lm(logSal ~ Senior + Age + ageSquared + ageCubed + Educ + Exper + experSquared + experCubed + Sex) plot(myLm2, which=1) # Residual plot plot(myLm1, which=4) # Cook's distance if(require(leaps)){ # Use the leaps library mySubsets <- regsubsets(logSal ~ (Senior + Age + Educ + Exper + ageSquared + experSquared)^2, nvmax=25, data=case1202) mySummary <- summary(mySubsets) p <- apply(mySummary$which, 1, sum) plot(mySummary$bic ~ p, ylab = "BIC") cbind(p,mySummary$bic) mySummary$which[8,] # Note that Age:ageSquared = ageCubed } myLm3 <- lm(logSal ~ Age + Educ + ageSquared + Senior:Educ + Age:Exper + ageCubed + Educ:Exper + Exper:ageSquared) summary(myLm3) myLm4 <- update(myLm3, ~ . + Sex) summary(myLm4) myLm5 <- update(myLm4, ~ . + Sex:Age + Sex:Educ + Sex:Senior + Sex:Exper + Sex:ageSquared) anova(myLm4, myLm5) ## INFERENCE AND INTERPRETATION summary(myLm4) beta <- myLm4$coef exp(beta[6]) exp(confint(myLm4,6)) # Conclusion: The median beginning salary for males was estimated to be 12% # higher than the median salary for females with similar values of the available # qualification variables (95% confidence interval: 7% to 17% higher). ## DISPLAY FOR PRESENTATION years <- Exper/12 # Change months to years plot(Bsal ~ years, log="y", xlab="Previous Work Experience (Years)", ylab="Beginning Salary (Dollars); Log Scale", main="Beginning Salaries and Experience for 61 Female and 32 Male Employees", pch= ifelse(Sex=="Male",24,21), bg = "gray", col= ifelse(Sex=="Male","blue","red"), lwd=2, cex=1.8 ) myLm6 <- lm(logSal ~ Exper + experSquared + experCubed + Sex) beta <- myLm6$coef dummyExper <- seq(min(Exper),max(Exper),length=50) curveF <- beta[1] + beta[2]*dummyExper + beta[3]*dummyExper^2 + beta[4]*dummyExper^3 curveM <- curveF + beta[5] dummyYears <- dummyExper/12 lines(exp(curveF) ~ dummyYears, lty=1, lwd=2,col="red") lines(exp(curveM) ~ dummyYears, lty = 2, lwd=2, col="blue") legend(28,8150, c("Male","Female"),pch=c(24,21), pt.cex=1.8, pt.lwd=2, pt.bg=c("gray","gray"), col=c("blue","red"), lty=c(2,1), lwd=2) detach(case1202)str(case1202) attach(case1202) ## EXPLORATION logSal <- log(Bsal) myMatrix <- cbind (logSal, Senior,Age, Educ, Exper) if(require(car)){ # Use the car library scatterplotMatrix(myMatrix, smooth=FALSE, diagonal="histogram", groups=Sex, col=c("red","blue") ) } myLm1 <- lm(logSal ~ Senior + Age + Educ + Exper + Sex) plot(myLm1, which=1) plot(myLm1, which=4) # Cook's Distance if(require(car)){ # Use the car library crPlots(myLm1) # Partial residual plots } ageSquared <- Age^2 ageCubed <- Age^3 experSquared <- Exper^2 experCubed <- Exper^3 myLm2 <- lm(logSal ~ Senior + Age + ageSquared + ageCubed + Educ + Exper + experSquared + experCubed + Sex) plot(myLm2, which=1) # Residual plot plot(myLm1, which=4) # Cook's distance if(require(leaps)){ # Use the leaps library mySubsets <- regsubsets(logSal ~ (Senior + Age + Educ + Exper + ageSquared + experSquared)^2, nvmax=25, data=case1202) mySummary <- summary(mySubsets) p <- apply(mySummary$which, 1, sum) plot(mySummary$bic ~ p, ylab = "BIC") cbind(p,mySummary$bic) mySummary$which[8,] # Note that Age:ageSquared = ageCubed } myLm3 <- lm(logSal ~ Age + Educ + ageSquared + Senior:Educ + Age:Exper + ageCubed + Educ:Exper + Exper:ageSquared) summary(myLm3) myLm4 <- update(myLm3, ~ . + Sex) summary(myLm4) myLm5 <- update(myLm4, ~ . + Sex:Age + Sex:Educ + Sex:Senior + Sex:Exper + Sex:ageSquared) anova(myLm4, myLm5) ## INFERENCE AND INTERPRETATION summary(myLm4) beta <- myLm4$coef exp(beta[6]) exp(confint(myLm4,6)) # Conclusion: The median beginning salary for males was estimated to be 12% # higher than the median salary for females with similar values of the available # qualification variables (95% confidence interval: 7% to 17% higher). ## DISPLAY FOR PRESENTATION years <- Exper/12 # Change months to years plot(Bsal ~ years, log="y", xlab="Previous Work Experience (Years)", ylab="Beginning Salary (Dollars); Log Scale", main="Beginning Salaries and Experience for 61 Female and 32 Male Employees", pch= ifelse(Sex=="Male",24,21), bg = "gray", col= ifelse(Sex=="Male","blue","red"), lwd=2, cex=1.8 ) myLm6 <- lm(logSal ~ Exper + experSquared + experCubed + Sex) beta <- myLm6$coef dummyExper <- seq(min(Exper),max(Exper),length=50) curveF <- beta[1] + beta[2]*dummyExper + beta[3]*dummyExper^2 + beta[4]*dummyExper^3 curveM <- curveF + beta[5] dummyYears <- dummyExper/12 lines(exp(curveF) ~ dummyYears, lty=1, lwd=2,col="red") lines(exp(curveM) ~ dummyYears, lty = 2, lwd=2, col="blue") legend(28,8150, c("Male","Female"),pch=c(24,21), pt.cex=1.8, pt.lwd=2, pt.bg=c("gray","gray"), col=c("blue","red"), lty=c(2,1), lwd=2) detach(case1202)
To study the influence of ocean grazers on regeneration rates of seaweed in the intertidal zone, a researcher scraped rock plots free of seaweed and observed the degree of regeneration when certain types of seaweed-grazing animals were denied access. The grazers were limpets (L), small fishes (f) and large fishes (F). Each plot received one of six treatments named by which grazers were allowed access. In addition, the researcher applied the treatments in eight blocks of 12 plots each. Within each block she randomly assigned treatments to plots. The blocks covered a wide range of tidal conditions.
case1301case1301
A data frame with 96 observations on the following 3 variables.
percent of regenerated seaweed cover
a factor with levels "B1", "B2",
"B3", "B4", "B5", "B6", "B7"
and "B8"
a factor indicating treatment, with levels
"C", "f", "fF", "L", "Lf" and
"LfF"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Olson, A. (1993). Evolutionary and Ecological Interactions Affecting Seaweeds, Ph.D. Thesis. Oregon State University.
str(case1301) attach(case1301) ## EXPLORATION AND MODEL DEVELOPMENT plot(Cover ~ Treat,xlab="Animals Present",ylab="Remaining Seaweed Coverage (%)") myLm1 <- lm(Cover ~ Block + Treat + Block:Treat) plot(myLm1,which=1) ratio <- Cover/(100 - Cover) logRatio <- log(ratio) myLm2 <- lm(logRatio ~ Block + Treat + Block:Treat) plot(myLm2, which=1) myLm3 <- lm(logRatio ~ Block + Treat) anova(myLm3, myLm2) # Test for interaction with extra ss F-test if(require(car)){ # Use the car library crPlots(myLm3) # Partial residual plots myLm4 <- lm(logRatio ~ Treat) anova(myLm4, myLm3) # Test for Block effect myLm5 <- lm(logRatio ~ Block) anova(myLm5, myLm3) # Test for Treatment effect lmp <- factor(ifelse(Treat %in% c("L", "Lf", "LfF"), "yes", "no")) sml <- factor(ifelse(Treat %in% c("f", "fF", "Lf", "LfF"), "yes","no")) big <- factor(ifelse(Treat %in% c("fF", "LfF"), "yes","no")) myLm6 <- lm(logRatio ~ Block + lmp + sml + big) crPlots(myLm6) myLm7 <- lm(logRatio ~ Block + (lmp + sml + big)^2) anova(myLm6, myLm7) # Test for interactions of lmp, sml, and big ## INFERENCE AND INTERPRETATION summary(myLm6) # Get p-values for lmp, sml, and big effects; R^2 = .8522 beta <- myLm6$coef exp(beta[9:11]) exp(confint(myLm6,9:11) ) myLm7 <- update(myLm6, ~ . - lmp) summary(myLm7) # R^2 = .4568; Note .8522-.4580 = 39.54# (explained by limpets) myLm8 <- update(myLm6, ~ . - big) summary(myLm8) # R^2 = .8225; Note .8522-.8255= 2.97# (explained by big fish) myLm9 <- update(myLm6, ~ . - sml) summary(myLm9) # R^2: .8400; Note .8522-.8400 = 1.22# (explained by small fish) ## DISPLAY FOR PRESENTATION myYLab <- "Adjusted Seaweed Regeneration (Log Scale; Deviation from Average)" crPlots(myLm6, ylab=myYLab, ylim=c(-2.2,2.2), main="Effects of Blocks and Treatments on Log Regeneration Ratio, Adjusted for Other Factors") } detach(case1301)str(case1301) attach(case1301) ## EXPLORATION AND MODEL DEVELOPMENT plot(Cover ~ Treat,xlab="Animals Present",ylab="Remaining Seaweed Coverage (%)") myLm1 <- lm(Cover ~ Block + Treat + Block:Treat) plot(myLm1,which=1) ratio <- Cover/(100 - Cover) logRatio <- log(ratio) myLm2 <- lm(logRatio ~ Block + Treat + Block:Treat) plot(myLm2, which=1) myLm3 <- lm(logRatio ~ Block + Treat) anova(myLm3, myLm2) # Test for interaction with extra ss F-test if(require(car)){ # Use the car library crPlots(myLm3) # Partial residual plots myLm4 <- lm(logRatio ~ Treat) anova(myLm4, myLm3) # Test for Block effect myLm5 <- lm(logRatio ~ Block) anova(myLm5, myLm3) # Test for Treatment effect lmp <- factor(ifelse(Treat %in% c("L", "Lf", "LfF"), "yes", "no")) sml <- factor(ifelse(Treat %in% c("f", "fF", "Lf", "LfF"), "yes","no")) big <- factor(ifelse(Treat %in% c("fF", "LfF"), "yes","no")) myLm6 <- lm(logRatio ~ Block + lmp + sml + big) crPlots(myLm6) myLm7 <- lm(logRatio ~ Block + (lmp + sml + big)^2) anova(myLm6, myLm7) # Test for interactions of lmp, sml, and big ## INFERENCE AND INTERPRETATION summary(myLm6) # Get p-values for lmp, sml, and big effects; R^2 = .8522 beta <- myLm6$coef exp(beta[9:11]) exp(confint(myLm6,9:11) ) myLm7 <- update(myLm6, ~ . - lmp) summary(myLm7) # R^2 = .4568; Note .8522-.4580 = 39.54# (explained by limpets) myLm8 <- update(myLm6, ~ . - big) summary(myLm8) # R^2 = .8225; Note .8522-.8255= 2.97# (explained by big fish) myLm9 <- update(myLm6, ~ . - sml) summary(myLm9) # R^2: .8400; Note .8522-.8400 = 1.22# (explained by small fish) ## DISPLAY FOR PRESENTATION myYLab <- "Adjusted Seaweed Regeneration (Log Scale; Deviation from Average)" crPlots(myLm6, ylab=myYLab, ylim=c(-2.2,2.2), main="Effects of Blocks and Treatments on Log Regeneration Ratio, Adjusted for Other Factors") } detach(case1301)
One company of soldiers in each of 10 platoons was assigned to a Pygmalion treatment group, with remaining companies in the platoon assigned to a control group. Leaders of the Pygmalion platoons were told their soldiers had done particularly well on a battery of tests which were, in fact, non-existent. In this randomised block experiment, platoons are experimental units, companies are blocks, and average Practical Specialty test score for soldiers in a platoon is the response. The researchers wished to see if the platoon response was affected by the artificially-induced expectations of the platoon leader.
case1302case1302
A data frame with 29 observations on the following 3 variables.
a factor indicating company identification,
with levels "C1", "C2", ..., "C10"
a factor indicating treatment with two levels,
"Pygmalion" and "Control"
average score on practical specialty test of all soldiers in the platoon
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Eden, D. (1990). Pygmalion Without Interpersonal Contrast Effects: Whole Groups Gain from Raising Manager Expectations, Journal of Applied Psychology 75(4): 395–398.
str(case1302) attach(case1302) ## EXPLORATION AND MODEL DEVELOPMENT plot(Score ~ as.numeric(Company),cex=1.5, pch=21, bg=ifelse(Treat=="Pygmalion","blue","light gray")) myLm1 <- lm(Score ~ Company + Treat + Company:Treat) # Fit with interaction. plot(myLm1,which=1) # Plot residuals. myLm2 <- update(myLm1, ~ . - Company:Treat) # Refit, without interaction. anova(myLm2, myLm1) # Show extra-ss-F-test p-value (for interaction effect). if(require(car)){ # Use the car library crPlots(myLm2) } ## INFERENCE myLm3 <- update(myLm2, ~ . - Company) # Fit reduced model without Company. anova(myLm3, myLm2) # Test for Company effect. summary(myLm2) # Show estimate and p-value for Pygmalion effect. confint(myLm2,11) # Show 95% CI for Pygmalion effect. ## DISPLAY FOR PRESENTATION beta <- myLm2$coef partialRes <- myLm2$res + beta[11]*ifelse(Treat=="Pygmalion",1,0) # partial res boxplot(partialRes ~ Treat, # Boxplots of partial residuals for each treatment ylab="Average Test Score, Adjusted for Company Effect (Deviation from Company Average)", names=c("19 Control Platoons","10 Pygmalion Treated Platoons"), col="green", boxlwd=2, medlwd=2,whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5 ) detach(case1302)str(case1302) attach(case1302) ## EXPLORATION AND MODEL DEVELOPMENT plot(Score ~ as.numeric(Company),cex=1.5, pch=21, bg=ifelse(Treat=="Pygmalion","blue","light gray")) myLm1 <- lm(Score ~ Company + Treat + Company:Treat) # Fit with interaction. plot(myLm1,which=1) # Plot residuals. myLm2 <- update(myLm1, ~ . - Company:Treat) # Refit, without interaction. anova(myLm2, myLm1) # Show extra-ss-F-test p-value (for interaction effect). if(require(car)){ # Use the car library crPlots(myLm2) } ## INFERENCE myLm3 <- update(myLm2, ~ . - Company) # Fit reduced model without Company. anova(myLm3, myLm2) # Test for Company effect. summary(myLm2) # Show estimate and p-value for Pygmalion effect. confint(myLm2,11) # Show 95% CI for Pygmalion effect. ## DISPLAY FOR PRESENTATION beta <- myLm2$coef partialRes <- myLm2$res + beta[11]*ifelse(Treat=="Pygmalion",1,0) # partial res boxplot(partialRes ~ Treat, # Boxplots of partial residuals for each treatment ylab="Average Test Score, Adjusted for Company Effect (Deviation from Company Average)", names=c("19 Control Platoons","10 Pygmalion Treated Platoons"), col="green", boxlwd=2, medlwd=2,whisklty=1, whisklwd=2, staplewex=.2, staplelwd=2, outlwd=2, outpch=21, outbg="green", outcex=1.5 ) detach(case1302)
Researchers taught each of 4 chimps to learn 10 words in American sign language and recorded the learning time for each word for each chimp. They wished to describe chimp differences and word differences.
case1401case1401
A data frame with 40 observations on the following 4 variables.
learning time in minutes
a factor indicating chimp, with four levels
"Booee", "Cindy", "Bruno" and "Thelma"
a factor indicating word taught, with 10 levels
the order in which the sign was taught
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Fouts, R.S. (1973). Acquisition and Testing of Gestural Signs in Four Young Chimpanzees, Science 180: 978–980.
str(case1401) attach(case1401) ## EXPLORATION AND MODEL DEVELOPMENT plot(Minutes ~ Sign) myLm1 <- lm(Minutes ~ Chimp + Sign) plot(myLm1,which=1) # Plot residuals (indicates a need for transformation). logMinutes <- log(Minutes) myLm2 <- lm(logMinutes ~ Chimp + Sign) plot(myLm2, which=1) # This looks fine. if(require(car)){ # Use the car library crPlots(myLm2) # Partial residual plots } ## INFERENCE AND INTERPRETATION myLm3 <- update(myLm2, ~ . - Chimp) # Fit reduced model without Chimp. anova(myLm3, myLm2) # Test for Chimp effect. myLm4 <- update(myLm2, ~ . - Sign) # Fit reduced model without Sign. anova(myLm4, myLm2) # Test for Sign effect. # Fit 2-way model without intercept to order signs from easiest to hardest myAov1 <- aov(logMinutes ~ Sign + Chimp - 1) sort(myAov1$coef[1:10]) # Show the ordering of Signs orderedSign <- factor(Sign,levels=c("listen","drink","shoe","key","more", "food","fruit","hat","look","string") ) # Re-order signs, easiest 1st myAov2 <- aov(logMinutes ~ orderedSign + Chimp - 1) # Refit opar <- par(no.readonly=TRUE) # Store current graphics parameters settings par(mar=c(4.1,7.1,4.1,2.1)) # Adjust margins to allow room for y-axis labels ## takes too long to run if(require(multcomp)){ # Use the multcomp library myMultComp <- glht(myAov2, linfct = mcp(orderedSign = "Tukey")) plot(myMultComp) # Plot Tukey-adjusted confidence intervals. summary(myMultComp) # Show Tukey-adjusted p-values pairwise comparisons confint(myMultComp) # Show Tukey-adjusted 95% confidence intervals } par(opar) # Restore original graphics parameters settings ## DISPLAY FOR PRESENTATION myYLab <- "Log Learning Time, Adjusted for Chimp Effect" myXLab <- "Sign Learned" if(require(car)){ # Use the car library crPlots(myAov2, ylab=myYLab, xlab=myXLab, main="Learning Times by Sign, Adjusted for Chimp Effects", layout=c(1,1)) # Click on graph area to show next page (Just use first one.) } detach(case1401)str(case1401) attach(case1401) ## EXPLORATION AND MODEL DEVELOPMENT plot(Minutes ~ Sign) myLm1 <- lm(Minutes ~ Chimp + Sign) plot(myLm1,which=1) # Plot residuals (indicates a need for transformation). logMinutes <- log(Minutes) myLm2 <- lm(logMinutes ~ Chimp + Sign) plot(myLm2, which=1) # This looks fine. if(require(car)){ # Use the car library crPlots(myLm2) # Partial residual plots } ## INFERENCE AND INTERPRETATION myLm3 <- update(myLm2, ~ . - Chimp) # Fit reduced model without Chimp. anova(myLm3, myLm2) # Test for Chimp effect. myLm4 <- update(myLm2, ~ . - Sign) # Fit reduced model without Sign. anova(myLm4, myLm2) # Test for Sign effect. # Fit 2-way model without intercept to order signs from easiest to hardest myAov1 <- aov(logMinutes ~ Sign + Chimp - 1) sort(myAov1$coef[1:10]) # Show the ordering of Signs orderedSign <- factor(Sign,levels=c("listen","drink","shoe","key","more", "food","fruit","hat","look","string") ) # Re-order signs, easiest 1st myAov2 <- aov(logMinutes ~ orderedSign + Chimp - 1) # Refit opar <- par(no.readonly=TRUE) # Store current graphics parameters settings par(mar=c(4.1,7.1,4.1,2.1)) # Adjust margins to allow room for y-axis labels ## takes too long to run if(require(multcomp)){ # Use the multcomp library myMultComp <- glht(myAov2, linfct = mcp(orderedSign = "Tukey")) plot(myMultComp) # Plot Tukey-adjusted confidence intervals. summary(myMultComp) # Show Tukey-adjusted p-values pairwise comparisons confint(myMultComp) # Show Tukey-adjusted 95% confidence intervals } par(opar) # Restore original graphics parameters settings ## DISPLAY FOR PRESENTATION myYLab <- "Log Learning Time, Adjusted for Chimp Effect" myXLab <- "Sign Learned" if(require(car)){ # Use the car library crPlots(myAov2, ylab=myYLab, xlab=myXLab, main="Learning Times by Sign, Adjusted for Chimp Effects", layout=c(1,1)) # Click on graph area to show next page (Just use first one.) } detach(case1401)
In a completely randomized design with a 2x3x5 factorial treatment structure, researchers randomly assigned one of 30 treatment combinations to open-topped growing chambers, in which two soybean cultivars were planted. The responses for each chamber were the yields of the two types of soybean.
case1402case1402
A data frame with 30 observations on the following 5 variables.
a factor indicating treatment, with two levels
"Well-watered" and "Stressed"
a quantitative treatment with three levels 0, 0.02 and 0.06
a quantitative treatment with five levels 0.02, 0.05, 0.07, 0.08 and 0.10
the yield of the Forrest cultivar of soybean (in kg/ha)
the yield of the Williams cultivar of soybean (in kg/ha)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Heggestad, H.E. and Lesser, V.M. (1990). Effects of Chronic Doses of Sulfur Dioxide, Ozone, and Drought on Yields and Growth of Soybeans Under Field Conditions, Journal of Environmental Quality 19: 488–495.
str(case1402) attach(case1402) ## EXPLORATION AND MODEL DEVELOPMENT; FORREST CULTIVAR logForrest <- log(Forrest) # Fit model without interactions first--to examine partial residual plots. myLm1 <- lm(logForrest ~ Stress + SO2 + O3) if(require(car)){ # Use the car library crPlots(myLm1) # Partial res plots => linear effects of SO2 and O3 look ok. } myLm2 <- lm(logForrest ~ (Stress + SO2 + O3)^2) # all 2-factor interactions. plot(myLm1, which=1) # Residual plot looks ok. anova(myLm1,myLm2) # Test for interactive effects. ## INFERENCE AND INTERPRETATION; FORREST CULTIVAR summary(myLm1) betaF <- myLm1$coef # Effect of 0.01 increase in SO2 (note coeff is negative): 100*(1 - exp(0.01*betaF[3])) #1.655701; a 1.65% decrease in median yield 100*(1-exp(0.01*confint(myLm1,"SO2"))) #3.772277 -0.5074294: 95% onfidence interval for effect of 0.01 increase in SO2 # Effect of 0.01 increase in O3 (note coeff is negative): 100*(1 - exp(0.01*betaF[4])) # 5.585979 I.e. a 5.6% decrease in median yield 100*(1-exp(0.01*confint(myLm1,"O3"))) #7.445966 3.688613: 95% confidence interval for effect of 0.01 increase in O3 # Effect of water stress (note coeff is positive for effect of well-watered): 100*(1 - exp(-betaF[2])) # Get estimate for negative of this beta #3.220556: a 3.2% decrease in median yield due to water stress 100*(1-exp(-confint(myLm1,2))) #-7.875521 13.17529: 95% confidence interval ## DISPLAY FOR PRESENTATION; FORREST CULTIVAR jO3 <- jitter(O3,factor=.25) # Jitter for plotting; jittering factor 0.25. jS <- jitter(SO2,factor=.25) # Jitter SO2 for plotting. cexfac <- 1.75 # Use character expansion factor of 1.75 for plotting symbols. opar <- par(no.readonly=TRUE) # Store current graphics parameters settings par(mfrow=c(1,2)) # Make a panel of 2 plots in 1 row. myPointCode <- ifelse(Stress=="Well-watered",21,24) myPointColor <- ifelse(Stress=="Well-watered","green","orange") par(mar=c(4.1,4.1,2.1,0.1)) plot(Forrest ~ jO3, log="y", ylab="Yield of Forrest Cultivar (kg/ha)", xlab=expression(paste(italic("Ozone ("),mu,"L/L), jittered")), pch=myPointCode, lwd=2, bg=myPointColor, cex=cexfac) legend(.02,2400, c("Well-watered","Water Stressed"), pch=c(21,24), pt.cex=cexfac, pt.bg=c("green","orange"), pt.lwd=2, lty=c(3,1), lwd=c(2,2)) dummyO <- seq(min(O3), max(O3), length=2) curve1 <- exp(betaF[1] + betaF[3]*mean(SO2) + betaF[4]*dummyO) curve2 <- exp(betaF[1] + betaF[2] + betaF[3]*mean(SO2)+ betaF[4]*dummyO) lines(curve1 ~ dummyO,lwd=2) lines(curve2 ~ dummyO,lwd=2,lty=3) # PLOT FORREST VS SO2 par(mar=c(4.1,2.1,2.1,2.1)) # Change margins plot(Forrest ~ jS, log="y", ylab="", xlab=expression(paste(italic("Sulfur Dioxide ("),mu,"L/L), jittered")), yaxt="n", pch=myPointCode, lwd=2, bg=myPointColor, cex=cexfac) dummyS <- seq(min(SO2),max(SO2),length=2) curve1 <- exp(betaF[1] + betaF[3]*dummyS + betaF[4]*mean(O3)) curve2 <- exp(betaF[1] + betaF[2] + betaF[3]*dummyS + betaF[4]*mean(O3)) lines(curve1 ~ dummyS,lwd=2) lines(curve2 ~ dummyS,lwd=2,lty=3) par(opar) # Restore previous graphics parameter settings detach(case1402)str(case1402) attach(case1402) ## EXPLORATION AND MODEL DEVELOPMENT; FORREST CULTIVAR logForrest <- log(Forrest) # Fit model without interactions first--to examine partial residual plots. myLm1 <- lm(logForrest ~ Stress + SO2 + O3) if(require(car)){ # Use the car library crPlots(myLm1) # Partial res plots => linear effects of SO2 and O3 look ok. } myLm2 <- lm(logForrest ~ (Stress + SO2 + O3)^2) # all 2-factor interactions. plot(myLm1, which=1) # Residual plot looks ok. anova(myLm1,myLm2) # Test for interactive effects. ## INFERENCE AND INTERPRETATION; FORREST CULTIVAR summary(myLm1) betaF <- myLm1$coef # Effect of 0.01 increase in SO2 (note coeff is negative): 100*(1 - exp(0.01*betaF[3])) #1.655701; a 1.65% decrease in median yield 100*(1-exp(0.01*confint(myLm1,"SO2"))) #3.772277 -0.5074294: 95% onfidence interval for effect of 0.01 increase in SO2 # Effect of 0.01 increase in O3 (note coeff is negative): 100*(1 - exp(0.01*betaF[4])) # 5.585979 I.e. a 5.6% decrease in median yield 100*(1-exp(0.01*confint(myLm1,"O3"))) #7.445966 3.688613: 95% confidence interval for effect of 0.01 increase in O3 # Effect of water stress (note coeff is positive for effect of well-watered): 100*(1 - exp(-betaF[2])) # Get estimate for negative of this beta #3.220556: a 3.2% decrease in median yield due to water stress 100*(1-exp(-confint(myLm1,2))) #-7.875521 13.17529: 95% confidence interval ## DISPLAY FOR PRESENTATION; FORREST CULTIVAR jO3 <- jitter(O3,factor=.25) # Jitter for plotting; jittering factor 0.25. jS <- jitter(SO2,factor=.25) # Jitter SO2 for plotting. cexfac <- 1.75 # Use character expansion factor of 1.75 for plotting symbols. opar <- par(no.readonly=TRUE) # Store current graphics parameters settings par(mfrow=c(1,2)) # Make a panel of 2 plots in 1 row. myPointCode <- ifelse(Stress=="Well-watered",21,24) myPointColor <- ifelse(Stress=="Well-watered","green","orange") par(mar=c(4.1,4.1,2.1,0.1)) plot(Forrest ~ jO3, log="y", ylab="Yield of Forrest Cultivar (kg/ha)", xlab=expression(paste(italic("Ozone ("),mu,"L/L), jittered")), pch=myPointCode, lwd=2, bg=myPointColor, cex=cexfac) legend(.02,2400, c("Well-watered","Water Stressed"), pch=c(21,24), pt.cex=cexfac, pt.bg=c("green","orange"), pt.lwd=2, lty=c(3,1), lwd=c(2,2)) dummyO <- seq(min(O3), max(O3), length=2) curve1 <- exp(betaF[1] + betaF[3]*mean(SO2) + betaF[4]*dummyO) curve2 <- exp(betaF[1] + betaF[2] + betaF[3]*mean(SO2)+ betaF[4]*dummyO) lines(curve1 ~ dummyO,lwd=2) lines(curve2 ~ dummyO,lwd=2,lty=3) # PLOT FORREST VS SO2 par(mar=c(4.1,2.1,2.1,2.1)) # Change margins plot(Forrest ~ jS, log="y", ylab="", xlab=expression(paste(italic("Sulfur Dioxide ("),mu,"L/L), jittered")), yaxt="n", pch=myPointCode, lwd=2, bg=myPointColor, cex=cexfac) dummyS <- seq(min(SO2),max(SO2),length=2) curve1 <- exp(betaF[1] + betaF[3]*dummyS + betaF[4]*mean(O3)) curve2 <- exp(betaF[1] + betaF[2] + betaF[3]*dummyS + betaF[4]*mean(O3)) lines(curve1 ~ dummyS,lwd=2) lines(curve2 ~ dummyS,lwd=2,lty=3) par(opar) # Restore previous graphics parameter settings detach(case1402)
Data from an observational study of nitrate levels measured at three week intervals for five years in two watersheds. One of the watersheds was undisturbed and the other had been logged with a patchwork pattern.
case1501case1501
A data frame with 88 observations on the following 3 variables.
week after the start of the study
natural logarithm of nitrate level (ppm) in the logged watershed (ppm)
natural logarithm of nitrate level in the undisturbed watershed (ppm)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learnings.
Harr, R.D., Friderksen, R.L., and Rothacher, J. (1979). Changes in Streamflow Following Timber Harvests in Southwestern Oregon, USDA/USFS Research Paper PNW-249, Pacific NW Forest and Range Experiment Station, Portland, Oregon.
str(case1501) attach(case1501) ## EXPLORATION opar <- par(no.readonly=TRUE) # Store current graphics parameters settings par(mfrow=c(2,1)) # Set graphics parameters: 2 row, 1 column layout plot(NoCut ~ Week, type="b", ylab="Log of Nitrate Concentration; NoCut") abline(h=mean(NoCut)) # Horizontal line at the mean plot(Patch ~ Week, type="b", ylab="Log of Nitrate Concentration; Patch Cut") abline(h=mean(Patch)) par(opar) # Restore previous graphics settings lag.plot(NoCut,do.lines=FALSE) # Lag plot for NoCut lag.plot(Patch,do.lines=FALSE) # Lag plot for Patch pacf(NoCut) # partial autocorrelation function plot; noCut pacf(Patch) # partial autocorrelation function plot; Patch ## INFERENCE (2-sample comparison, accounting for first serial correlation) diff <- mean(Patch) - mean(NoCut) nPatch <- length(Patch) # length of Patch series nNoCut <- length(NoCut) # length of NoCut series acfPatch <- acf(Patch, type="covariance") # auto covariances for Patch series c0Patch <- acfPatch$acf[1]*nPatch/(nPatch-1) # variance; n-1 divisor (Patch) c1Patch <- acfPatch$acf[2]*nPatch/(nPatch-1) # autocov; n-1 divisor (Patch) acfNoCut <- acf(NoCut, type="covariance") # auto covariances for NoCut series c0NoCut <- acfNoCut$acf[1]*nNoCut/(nNoCut - 1) # variance; n-1 divisor (NoCut) c1NoCut <- acfNoCut$acf[2]*nNoCut/(nNoCut - 1) # autocov; n-1 divisor (NoCut) dfPatch <- nPatch - 1 # DF (n-1); Patch dfNoCut <- nNoCut - 1 # DF (n-1); NoCut c0Pooled <- (dfPatch*c0Patch + dfNoCut*c0NoCut)/(dfPatch + dfNoCut) c0Pooled #[1] 1.413295 = pooled estimate of variance c1Pooled <- (dfPatch*c1Patch + dfNoCut*c1NoCut)/(dfPatch + dfNoCut) c1Pooled #[1] 0.9103366 = pooled estimate of lag 1 covariance # Pooled estimate of first serial correlation coefficient: r1 <- c1Pooled/c0Pooled #[1] 0.6441233 SEdiff <- sqrt((1 + r1)/(1-r1))*sqrt(c0Pooled*(1/nPatch + 1/nNoCut)) # t-test and confidence interval tStat <- diff/SEdiff #[1] 0.2713923 pValue <- 1 - pt(tStat,dfPatch + dfNoCut) # One-sided p-value halfWidth <- qt(.975,dfPatch + dfNoCut)*SEdiff # half width of 95% CI diff + c(-1,1)*halfWidth #95% CI -0.6557578 0.8648487 ## GRAPHICAL DISPLAY FOR PRESENTATION par(mfrow=c(1,1)) # Reset mfrow to a single plot per page plot(exp(Patch) ~ Week, # Use exp(Patch) to show results in original units log="y", type="b", xlab="Weeks After Logging", ylab="Nitrate Concentration in Watershed Runoff (ppm)", main="Nitrate Series in Patch-Cut and Undisturbed Watersheds", pch=21, col="dark green", lwd=3, bg="green", cex=1.3 ) points(exp(NoCut) ~ Week, pch=24, col="dark blue", lwd=3, bg="orange",cex=1.3) lines(exp(NoCut) ~ Week, lwd=3, col="dark blue",lty=3) abline(h=exp(mean(Patch)),col="dark green",lwd=2) abline(h=exp(mean(NoCut)),col="dark blue", lwd=2,lty=2) legend(205,100,legend=c("Patch Cut", "Undisturbed"), pch=c(21,24), col=c("dark green","dark blue"), pt.bg = c("green","orange"), pt.cex=c(1.3,1.3), lty=c(1,3), lwd=c(3,3)) text(-1, 8.5, "Mean",col="dark green") text(-1,6.3,"Mean", col="dark blue") detach(case1501)str(case1501) attach(case1501) ## EXPLORATION opar <- par(no.readonly=TRUE) # Store current graphics parameters settings par(mfrow=c(2,1)) # Set graphics parameters: 2 row, 1 column layout plot(NoCut ~ Week, type="b", ylab="Log of Nitrate Concentration; NoCut") abline(h=mean(NoCut)) # Horizontal line at the mean plot(Patch ~ Week, type="b", ylab="Log of Nitrate Concentration; Patch Cut") abline(h=mean(Patch)) par(opar) # Restore previous graphics settings lag.plot(NoCut,do.lines=FALSE) # Lag plot for NoCut lag.plot(Patch,do.lines=FALSE) # Lag plot for Patch pacf(NoCut) # partial autocorrelation function plot; noCut pacf(Patch) # partial autocorrelation function plot; Patch ## INFERENCE (2-sample comparison, accounting for first serial correlation) diff <- mean(Patch) - mean(NoCut) nPatch <- length(Patch) # length of Patch series nNoCut <- length(NoCut) # length of NoCut series acfPatch <- acf(Patch, type="covariance") # auto covariances for Patch series c0Patch <- acfPatch$acf[1]*nPatch/(nPatch-1) # variance; n-1 divisor (Patch) c1Patch <- acfPatch$acf[2]*nPatch/(nPatch-1) # autocov; n-1 divisor (Patch) acfNoCut <- acf(NoCut, type="covariance") # auto covariances for NoCut series c0NoCut <- acfNoCut$acf[1]*nNoCut/(nNoCut - 1) # variance; n-1 divisor (NoCut) c1NoCut <- acfNoCut$acf[2]*nNoCut/(nNoCut - 1) # autocov; n-1 divisor (NoCut) dfPatch <- nPatch - 1 # DF (n-1); Patch dfNoCut <- nNoCut - 1 # DF (n-1); NoCut c0Pooled <- (dfPatch*c0Patch + dfNoCut*c0NoCut)/(dfPatch + dfNoCut) c0Pooled #[1] 1.413295 = pooled estimate of variance c1Pooled <- (dfPatch*c1Patch + dfNoCut*c1NoCut)/(dfPatch + dfNoCut) c1Pooled #[1] 0.9103366 = pooled estimate of lag 1 covariance # Pooled estimate of first serial correlation coefficient: r1 <- c1Pooled/c0Pooled #[1] 0.6441233 SEdiff <- sqrt((1 + r1)/(1-r1))*sqrt(c0Pooled*(1/nPatch + 1/nNoCut)) # t-test and confidence interval tStat <- diff/SEdiff #[1] 0.2713923 pValue <- 1 - pt(tStat,dfPatch + dfNoCut) # One-sided p-value halfWidth <- qt(.975,dfPatch + dfNoCut)*SEdiff # half width of 95% CI diff + c(-1,1)*halfWidth #95% CI -0.6557578 0.8648487 ## GRAPHICAL DISPLAY FOR PRESENTATION par(mfrow=c(1,1)) # Reset mfrow to a single plot per page plot(exp(Patch) ~ Week, # Use exp(Patch) to show results in original units log="y", type="b", xlab="Weeks After Logging", ylab="Nitrate Concentration in Watershed Runoff (ppm)", main="Nitrate Series in Patch-Cut and Undisturbed Watersheds", pch=21, col="dark green", lwd=3, bg="green", cex=1.3 ) points(exp(NoCut) ~ Week, pch=24, col="dark blue", lwd=3, bg="orange",cex=1.3) lines(exp(NoCut) ~ Week, lwd=3, col="dark blue",lty=3) abline(h=exp(mean(Patch)),col="dark green",lwd=2) abline(h=exp(mean(NoCut)),col="dark blue", lwd=2,lty=2) legend(205,100,legend=c("Patch Cut", "Undisturbed"), pch=c(21,24), col=c("dark green","dark blue"), pt.bg = c("green","orange"), pt.cex=c(1.3,1.3), lty=c(1,3), lwd=c(3,3)) text(-1, 8.5, "Mean",col="dark green") text(-1,6.3,"Mean", col="dark blue") detach(case1501)
The data are the temperatures (in degrees Celsius) averaged for the northern hemisphere over a full year, for years 1850 to 2010. The 161-year average temperature has been subtracted, so each observation is the temperature difference from the series average.
case1502case1502
A data frame with 161 observations on the following 2 variables.
year in which yearly average temperature was computed, from 1850 to 2010
northern hemisphere temperature minus the 161-year average (degrees Celsius)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Jones, P.D., D. E. Parker, T. J. Osborn, and K. R. Briffa, (2011) Global and Hemispheric Temperature Anomalies and and Marine Instrumental Records, CDIAC, http://cdiac.ornl.gov/trends/temp/jonescru/jones.html, August 4, 2011.
str(case1502) attach(case1502) ## EXPLORATION AND MODEL BUILDING plot(Temperature ~ Year, type="b") # Type = "b" for *both* points and lines yearSquared <- Year^2 yearCubed <- Year^3 myLm1 <- lm(Temperature ~ Year + yearSquared + yearCubed) res1 <- myLm1$res myPacf <- pacf(res1) # Partial autocorrelation from residuals r1 <- myPacf$acf[1] #First serial correlation coefficient n <- length(Temperature) # Series length = 161 v <- Temperature[2:n] - r1*Temperature[1:(n-1)] # Filtered response ones <- rep(1-r1, n-1) # make a variable of all 1's u1 <- Year[2:n] - r1*Year[1:(n-1)] # Filtered "ones" u2 <- yearSquared[2:n] - r1*yearSquared[1:(n-1)] # Filtered X1 u3 <- yearCubed[2:n] - r1*yearCubed[1:(n-1)] # Filtered X2 myLm2 <- lm(v ~ u1 + u2 + u3 ) res2 <- myLm2$res pacf(res2) # Looks fine; don't worry about lag 4 marginal significance plot(myLm2, which=1) # Residual plot summary(myLm2) # Cubic term isn't needed. myLm3 <- update(myLm2, ~ . - u3) # Drop cubic term ## INFERENCE summary(myLm3) # Everything remaining is statistically significant. ## GRAPHICAL DISPLAY FOR PRESENTATION plot(Temperature ~ Year, xlab="Year", ylab=expression(paste("Annual Average Temperature (Difference From Average), ", degree,"C")),main="Annual Average Temperature in Northern Hemisphere; 1850-2010", type="b", pch=21, lwd=2, bg="green", cex=1.5) myFits <- myLm3$fit lines(myFits ~ Year[2:161], col="blue", lwd=2) legend(1850,0.6,"Quadratic Regression Fit, Adjusted for AR(1) Serial Correlation", col="blue", lwd=2, box.lty=0) detach(case1502)str(case1502) attach(case1502) ## EXPLORATION AND MODEL BUILDING plot(Temperature ~ Year, type="b") # Type = "b" for *both* points and lines yearSquared <- Year^2 yearCubed <- Year^3 myLm1 <- lm(Temperature ~ Year + yearSquared + yearCubed) res1 <- myLm1$res myPacf <- pacf(res1) # Partial autocorrelation from residuals r1 <- myPacf$acf[1] #First serial correlation coefficient n <- length(Temperature) # Series length = 161 v <- Temperature[2:n] - r1*Temperature[1:(n-1)] # Filtered response ones <- rep(1-r1, n-1) # make a variable of all 1's u1 <- Year[2:n] - r1*Year[1:(n-1)] # Filtered "ones" u2 <- yearSquared[2:n] - r1*yearSquared[1:(n-1)] # Filtered X1 u3 <- yearCubed[2:n] - r1*yearCubed[1:(n-1)] # Filtered X2 myLm2 <- lm(v ~ u1 + u2 + u3 ) res2 <- myLm2$res pacf(res2) # Looks fine; don't worry about lag 4 marginal significance plot(myLm2, which=1) # Residual plot summary(myLm2) # Cubic term isn't needed. myLm3 <- update(myLm2, ~ . - u3) # Drop cubic term ## INFERENCE summary(myLm3) # Everything remaining is statistically significant. ## GRAPHICAL DISPLAY FOR PRESENTATION plot(Temperature ~ Year, xlab="Year", ylab=expression(paste("Annual Average Temperature (Difference From Average), ", degree,"C")),main="Annual Average Temperature in Northern Hemisphere; 1850-2010", type="b", pch=21, lwd=2, bg="green", cex=1.5) myFits <- myLm3$fit lines(myFits ~ Year[2:161], col="blue", lwd=2) legend(1850,0.6,"Quadratic Regression Fit, Adjusted for AR(1) Serial Correlation", col="blue", lwd=2, box.lty=0) detach(case1502)
Researchers taught 18 monkeys to distinguish each of 100 pairs of objects, 20 pairs each at 16, 12, 8, 4, and 2 weeks prior to a treatment. After this training, they blocked access to the hippocampal formation in 11 of the monkeys. All monkeys were then tested on their ability to distinguish the objects. The five-dimensional response for each monkey is the number of correct objects distinguished among those taught at 16, 12, 8, 4, and 2 weeks prior to treatment.
case1601case1601
A data frame with 18 observations on the following 7 variables.
Monkey name
a treatment factor with levels "Control" and
"Treated"
percentage of 20 objects taught 2 weeks prior to treatment that were correctly distinguished in the test
percentage of 20 objects taught 4 weeks prior to treatment that were correctly distinguished in the test
percentage of 20 objects taught 8 weeks prior to treatment that were correctly distinguished in the test
percentage of 20 objects taught 12 weeks prior to treatment that were correctly distinguished in the test
percentage of 20 objects taught 16 weeks prior to treatment that were correctly distinguished in the test
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Sola-Morgan, S. M. and Squire, L. R. (1990). The Primate Hippocampal Formation: Evidence for a Time-limited Role in Memory Storage, Science 250: 288–290.
str(case1601) attach(case1601) ## EXPLORATION short <- (Week2 + Week4)/2 long <- (Week8 + Week12 + Week16)/3 myPointCode <- ifelse(Treatment=="Control",15,16) myPointColor <- ifelse(Treatment=="Control","orange","green") plot(long ~ short, pch=myPointCode, col=myPointColor, cex=2) abline(h=mean(long),lty=2) abline(v=mean(short),lty=2) identify(short,long,labels=Monkey) # Identify outliers; press Esc when done ## INFERENCE USING HOTELLING's T-SQUARED TEST myLm1 <- lm(cbind(short,long) ~ Treatment) # Full model myLm2 <- lm(cbind(short,long) ~ 1) # Reduced model, with only intercept anova(myLm2, myLm1, test="Hotelling") # p-value for Treatment effect # confidence intervals n1 <- sum(Treatment=="Control") # 7 control monkeys n2 <- sum(Treatment=="Treated") # 11 treated monkeys multiplier <- sqrt(2*((n1+n2-2)/(n1+n2-3))*qf(.95,2,n1+n2-3)) # Sleuth p. 492 summary(myLm1) shortEffect <- myLm1$coef[2,1] # Difference in sample averages; Short seShortEffect <- 3.352 # Read this from summary(myLm1) halfWidth <- multiplier*seShortEffect # Half width of 95% confidence interval shortEffect + c(-1,1)*halfWidth #95% CI for effect of treatment on Short longEffect <- myLm1$coef[2,2] # Difference in sample averages; Long seLongEffect <- 3.2215 # Read this from summary(myLm1) halfWidth <- multiplier*seLongEffect # Half width of 95% confidence interval longEffect + c(-1,1)*halfWidth #95% CI for effect of treatment on Long ## GRAPHICAL DISPLAY FOR PRESENTATION myPointCode <- ifelse(Treatment=="Control",21,22) myPointColor <- ifelse(Treatment=="Control","green","orange") plot(long ~ jitter(short), xlab="Short-Term Memory Score (Percent Correct)", ylab="Long-Term Memory Score (Percent Correct)", main="Memory Scores for 11 Hippocampus-Blocked and 7 Control Monkeys", pch=myPointCode, bg=myPointColor, cex=2.5, lwd=3) identify(short,long,labels=Monkey) # Label the outliers; press Esc when done legend(52,54,legend=c("Control","Hippocampus Blocked"), pch=c(21,22), pt.bg=c("green","orange"), pt.cex=c(2.5,2.5), pt.lwd=c(3,3), cex=1.5) ## ADVANCED: RANDOMIZATION TEST FOR EQUALITY OF BIVARIATE RESPONSES myAnova <- anova(myLm2, myLm1, test="Hotelling") #Hotelling Test for Treatment myAnova$approx[2] #[1] 12.32109: F-statistic numRep <- 50 # Number of random regroupings (change to 50,000) FStats <- rep(0,numRep) # Initialize a variable for storing the F-statistics myLmReduced <- lm(cbind(short,long) ~ 1)# Fit the reduced model once for (rep in 1:numRep) { # Do the following commands in parenthese num.rep times randomGroup <- rep("Group1",18) # Set randomGroup initially to all "Group1" randomGroup[sample(1:18,7)] <- "Group2" # Change 7 at random to "Group2" randomGroup <- factor(randomGroup) # Make the character variable a factor myLmFull <- lm(cbind(short,long) ~ randomGroup) # Fit full model myAnova2 <- anova(myLmReduced, myLmFull, test="Hotelling") # Hotelling's test FStats[rep] <- myAnova2$approx[2] # Store the F-statistic } # If numRep = 50,000, go get a cup of coffee while you wait for this. hist(FStats, main="Approx. Randomizatin Dist of F-stat if No Treatment Effect") abline(v=12.32109) # Show actually observed Hotelling F-statistic pValue <- sum(FStats >= 12.32109)/numRep pValue # Approximate randomization test p-value (no distributional assumptions) detach(case1601)str(case1601) attach(case1601) ## EXPLORATION short <- (Week2 + Week4)/2 long <- (Week8 + Week12 + Week16)/3 myPointCode <- ifelse(Treatment=="Control",15,16) myPointColor <- ifelse(Treatment=="Control","orange","green") plot(long ~ short, pch=myPointCode, col=myPointColor, cex=2) abline(h=mean(long),lty=2) abline(v=mean(short),lty=2) identify(short,long,labels=Monkey) # Identify outliers; press Esc when done ## INFERENCE USING HOTELLING's T-SQUARED TEST myLm1 <- lm(cbind(short,long) ~ Treatment) # Full model myLm2 <- lm(cbind(short,long) ~ 1) # Reduced model, with only intercept anova(myLm2, myLm1, test="Hotelling") # p-value for Treatment effect # confidence intervals n1 <- sum(Treatment=="Control") # 7 control monkeys n2 <- sum(Treatment=="Treated") # 11 treated monkeys multiplier <- sqrt(2*((n1+n2-2)/(n1+n2-3))*qf(.95,2,n1+n2-3)) # Sleuth p. 492 summary(myLm1) shortEffect <- myLm1$coef[2,1] # Difference in sample averages; Short seShortEffect <- 3.352 # Read this from summary(myLm1) halfWidth <- multiplier*seShortEffect # Half width of 95% confidence interval shortEffect + c(-1,1)*halfWidth #95% CI for effect of treatment on Short longEffect <- myLm1$coef[2,2] # Difference in sample averages; Long seLongEffect <- 3.2215 # Read this from summary(myLm1) halfWidth <- multiplier*seLongEffect # Half width of 95% confidence interval longEffect + c(-1,1)*halfWidth #95% CI for effect of treatment on Long ## GRAPHICAL DISPLAY FOR PRESENTATION myPointCode <- ifelse(Treatment=="Control",21,22) myPointColor <- ifelse(Treatment=="Control","green","orange") plot(long ~ jitter(short), xlab="Short-Term Memory Score (Percent Correct)", ylab="Long-Term Memory Score (Percent Correct)", main="Memory Scores for 11 Hippocampus-Blocked and 7 Control Monkeys", pch=myPointCode, bg=myPointColor, cex=2.5, lwd=3) identify(short,long,labels=Monkey) # Label the outliers; press Esc when done legend(52,54,legend=c("Control","Hippocampus Blocked"), pch=c(21,22), pt.bg=c("green","orange"), pt.cex=c(2.5,2.5), pt.lwd=c(3,3), cex=1.5) ## ADVANCED: RANDOMIZATION TEST FOR EQUALITY OF BIVARIATE RESPONSES myAnova <- anova(myLm2, myLm1, test="Hotelling") #Hotelling Test for Treatment myAnova$approx[2] #[1] 12.32109: F-statistic numRep <- 50 # Number of random regroupings (change to 50,000) FStats <- rep(0,numRep) # Initialize a variable for storing the F-statistics myLmReduced <- lm(cbind(short,long) ~ 1)# Fit the reduced model once for (rep in 1:numRep) { # Do the following commands in parenthese num.rep times randomGroup <- rep("Group1",18) # Set randomGroup initially to all "Group1" randomGroup[sample(1:18,7)] <- "Group2" # Change 7 at random to "Group2" randomGroup <- factor(randomGroup) # Make the character variable a factor myLmFull <- lm(cbind(short,long) ~ randomGroup) # Fit full model myAnova2 <- anova(myLmReduced, myLmFull, test="Hotelling") # Hotelling's test FStats[rep] <- myAnova2$approx[2] # Store the F-statistic } # If numRep = 50,000, go get a cup of coffee while you wait for this. hist(FStats, main="Approx. Randomizatin Dist of F-stat if No Treatment Effect") abline(v=12.32109) # Show actually observed Hotelling F-statistic pValue <- sum(FStats >= 12.32109)/numRep pValue # Approximate randomization test p-value (no distributional assumptions) detach(case1601)
In a randomized, double-blind, crossover experiment, researchers randomly assigned 20 volunteer hospital employees to either a low-fiber or low-fiber treatment group. The subjects followed the diets for six weeks. After two weeks on their normal diet, all patients crossed over to the other treatment group for another six weeks. The total serum cholesterol (in mg/dl) was measured on each patient before the first treatment, at the end of the first six week treatment, and at the end of the second six week treatment.
case1602case1602
A data frame with 20 observations on the following 4 variables.
total serum cholesterol before treatment
total serum cholesterol after the high fiber diet
total serum cholesterol after the low fiber diet
factor to identify order of treatment, with two
levels "HL" and "LH"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Swain, J.F., Rouse, I.L., Curley, C.B., and Sacks, F.M. (1990). Comparison of the Effects of Oat Bran and Low-fiber Wheat on Serum Lipoprotein Levels and Blood Pressure, New England Journal of Medicine 320: 1746–1747.
str(case1602) attach(case1602) ## EXPLORATION highMinusBase <- HiFiber-Baseline highMinusLow <- HiFiber-LoFiber plot(highMinusBase ~ highMinusLow) abline(h=0) # Horizontal line at 0 abline(v=0) # Vertical line at 0 # Hotelling 2-sample t-test for order effect on bivariate response: myLm1 <- lm(cbind(highMinusBase,highMinusLow) ~ Order ) # Full model myLm2 <- update(myLm1, ~ . - Order) # Reduced model withour Order effect anova(myLm2, myLm1, test="Hotelling") # p-value for Order effect ## INFERENCE: HOTELLING ONE-SAMPLE TEST THAT MEAN OF BIVARIATE RESPONSE IS (0,0) myLm3 <- lm(cbind(highMinusBase, highMinusLow) ~ 1) # Full model myLm4 <- update(myLm3, ~ . - 1) # Reduced model (with both means = 0) anova(myLm4, myLm3, test="Hotelling") # test that the bivariate mean is (0,0) # Confidence intervals summary(myLm3) HighMinusBase <- myLm3$coef[1] # -13.850 seHighMinusBase <- 3.533 # Standard error, read from summary(myLm3) HighMinusLow <- myLm3$coef[2] # -0.850 seHighMinusLow <- 3.527 # Standard error, read from summary(myLm3) n <- length(highMinusBase) # 20: sample size multiplier <- sqrt(2*((n-1)/(n-2))*qf(.95,2,n-2)) # See Sleuth, page 495 HighMinusBase + c(-1,1)*multiplier*seHighMinusBase # 95% CI for High - Base HighMinusLow + c(-1,2)*multiplier*seHighMinusLow # 95% CI for High - Low ## GRAPHICAL DISPLAY FOR PRESENTATION lowMinusBase <- LoFiber - Baseline myPointCode <- ifelse(Order== "HL",21,22) myPointColor <- ifelse(Order== "HL","green","light blue") plot(highMinusBase ~ lowMinusBase, xlab="Cholesterol Change (from Baseline) After High Fiber Diet (mg/dl)", ylab="Cholesterol Change (From Baseline) After Low Fiber Diet (mg/dl)", main="Cholesterol Effects of High- and Low-Cholesterol Diets on 20 Subjects", ylim=c(-45,20), pch=myPointCode, bg=myPointColor, cex=2.5, lwd=2) abline(h=0) # Horizontal line at 0 abline(v=0) # Vertical line at 0 legend(-43,22,c("High Fiber Given First","Low Fiber Given First"), pch=c(21,22), pt.bg =c("green","light blue"),pt.cex =c(2.5,2.5), lw = c(2,2), lty=c(0,0)) detach(case1602)str(case1602) attach(case1602) ## EXPLORATION highMinusBase <- HiFiber-Baseline highMinusLow <- HiFiber-LoFiber plot(highMinusBase ~ highMinusLow) abline(h=0) # Horizontal line at 0 abline(v=0) # Vertical line at 0 # Hotelling 2-sample t-test for order effect on bivariate response: myLm1 <- lm(cbind(highMinusBase,highMinusLow) ~ Order ) # Full model myLm2 <- update(myLm1, ~ . - Order) # Reduced model withour Order effect anova(myLm2, myLm1, test="Hotelling") # p-value for Order effect ## INFERENCE: HOTELLING ONE-SAMPLE TEST THAT MEAN OF BIVARIATE RESPONSE IS (0,0) myLm3 <- lm(cbind(highMinusBase, highMinusLow) ~ 1) # Full model myLm4 <- update(myLm3, ~ . - 1) # Reduced model (with both means = 0) anova(myLm4, myLm3, test="Hotelling") # test that the bivariate mean is (0,0) # Confidence intervals summary(myLm3) HighMinusBase <- myLm3$coef[1] # -13.850 seHighMinusBase <- 3.533 # Standard error, read from summary(myLm3) HighMinusLow <- myLm3$coef[2] # -0.850 seHighMinusLow <- 3.527 # Standard error, read from summary(myLm3) n <- length(highMinusBase) # 20: sample size multiplier <- sqrt(2*((n-1)/(n-2))*qf(.95,2,n-2)) # See Sleuth, page 495 HighMinusBase + c(-1,1)*multiplier*seHighMinusBase # 95% CI for High - Base HighMinusLow + c(-1,2)*multiplier*seHighMinusLow # 95% CI for High - Low ## GRAPHICAL DISPLAY FOR PRESENTATION lowMinusBase <- LoFiber - Baseline myPointCode <- ifelse(Order== "HL",21,22) myPointColor <- ifelse(Order== "HL","green","light blue") plot(highMinusBase ~ lowMinusBase, xlab="Cholesterol Change (from Baseline) After High Fiber Diet (mg/dl)", ylab="Cholesterol Change (From Baseline) After Low Fiber Diet (mg/dl)", main="Cholesterol Effects of High- and Low-Cholesterol Diets on 20 Subjects", ylim=c(-45,20), pch=myPointCode, bg=myPointColor, cex=2.5, lwd=2) abline(h=0) # Horizontal line at 0 abline(v=0) # Vertical line at 0 legend(-43,22,c("High Fiber Given First","Low Fiber Given First"), pch=c(21,22), pt.bg =c("green","light blue"),pt.cex =c(2.5,2.5), lw = c(2,2), lty=c(0,0)) detach(case1602)
Engineers manipulated three factors (with 3, 2, and 4 levels each) in the construction and operation of printer rods, to see if they influenced the magnetic force around the rod.
case1701case1701
A data frame with 44 observations on the following 14 variables.
the magnetic force at each of the equally-spaces positions 1, 2, ..., 11 on the printer rod
electric current passing through the rod, with
three levels "0", "250" and "500"
(milliamperes)
a factor identifying the configuration, with
two levels "0" and "1"
a factor identifying the type of metal from
which the rod was made, with four levels
"1", "2", "3" and "4"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(case1701) attach(case1701) ## EXPLORATION MagneticForces <- cbind(L1,L2,L3,L4,L5,L6,L7,L8,L9,L10,L11) mfCor <- cor(MagneticForces) round(mfCor,2) # Show correlations, rounded to two digits mfPca <- prcomp(MagneticForces) # principal components summary(mfPca) # Show the proportion of variance explained by each PC plot(mfPca) # Graph proportion of variances explained by each PC (Scree Plot) mfCoefs <- mfPca$rotation # Extract the coefficients dim(mfCoefs) # #11 rows and 11 columns round(mfCoefs[,1:3],3) # Show the first 3 columns of the score matrix, rounded # Explore a possible meaningful linear combination suggested by first PC round(mfCoefs[,1],1) # Show coefficients of 1st pc, rounded to 1 digit # Coefficients are all very similar, suggesting a constant coefficient; use 1/11 mfAve <- (L1 + L2 + L3 + L4 + L5 + L6 + L7 + L8 + L9 + L10 + L11)/11 mfScores <- mfPca$x pc1 <- mfScores[,1] #Values for first principal component of MagneticForces cor(mfAve,pc1) # correlation of the average and the first PC (=0.999567) plot(pc1 ~ mfAve) # Explore a possible meningful linear combination suggested by second PC round(mfCoefs[,2],1) # Show coefficients of 2nd pc, rounded to 1 digit # Second set of coefficients are negative on the left end of the rod and # positive on the right end. Try Ave(L9 + L10 + L11) - Ave(L1 + L2 + L3). mfEnds <- (L9 + L10 + L11)/3 - (L1 + L2 + L3)/3 pc2 <- mfScores[,2] residualEnds <- lm(mfEnds ~ mfAve)$residual # Ends with average effect removed plot(pc2 ~ residualEnds) cor(residualEnds, pc2) # 0.973 # Explore a possible meaningful linear combination suggested by third PC round(mfCoefs[,3],1) # Show doefficients of 3rd pc, rounded to 1 digit # Try a contrast between the first 4 positions and the 6th position mfPeak <- (L1 + L2 + L3 + L4)/4 - L6 pc3 <- mfScores[,3] residualPeak <- lm(mfPeak ~ mfAve + mfEnds)$residual plot(pc3 ~ residualPeak) cor(residualPeak,pc3) # 0.971 # Note: the variation explained by the third PC seems to be due almost entirely # to one printer rod. (Keep this in mind.) ## INFERENCE: ANALYSIS OF EXPERIMENTAL FACTORS ON 3-DIMENSIONAL RESPONSE myResponse <- cbind(mfAve, mfEnds, mfPeak) cor(myResponse) myLm1 <- lm(myResponse ~ Current + Config + Material) anova(myLm1) # Noticeable effect of Current but not Config or Material plot(mfAve ~ Current) myLm2 <- lm(mfAve ~ Current) abline(myLm2) summary(myLm2) # No evidence of an effect of current on average magnetic force plot(mfEnds ~ Current) myLm3 <- lm(mfEnds ~ Current) abline(myLm3) summary(myLm3) # Evidence that Current effects the difference in MF at the ends plot(mfPeak ~ Current) myLm4 <- lm(mfPeak ~ Current) abline(myLm4) summary(myLm4) # No evidence of an effect of Current on peak MF in the center ## GRAPHICAL DISPLAY FOR PRESENTATION plot(mfEnds ~ jitter(Current,.1), xlab="Electrical Current Used in Printer Rod Manufacture (milliamperes)" , ylab="Magnetic Force at Positions 9-11 Minus Magnetic Force at Positions 1-3", main="Effect of Electrical Current on Magnetic Force Surrounding Printer Rod", col="black", pch=21, lwd=2, bg="green", cex=2 ) abline(myLm3, lwd=2) detach(case1701)str(case1701) attach(case1701) ## EXPLORATION MagneticForces <- cbind(L1,L2,L3,L4,L5,L6,L7,L8,L9,L10,L11) mfCor <- cor(MagneticForces) round(mfCor,2) # Show correlations, rounded to two digits mfPca <- prcomp(MagneticForces) # principal components summary(mfPca) # Show the proportion of variance explained by each PC plot(mfPca) # Graph proportion of variances explained by each PC (Scree Plot) mfCoefs <- mfPca$rotation # Extract the coefficients dim(mfCoefs) # #11 rows and 11 columns round(mfCoefs[,1:3],3) # Show the first 3 columns of the score matrix, rounded # Explore a possible meaningful linear combination suggested by first PC round(mfCoefs[,1],1) # Show coefficients of 1st pc, rounded to 1 digit # Coefficients are all very similar, suggesting a constant coefficient; use 1/11 mfAve <- (L1 + L2 + L3 + L4 + L5 + L6 + L7 + L8 + L9 + L10 + L11)/11 mfScores <- mfPca$x pc1 <- mfScores[,1] #Values for first principal component of MagneticForces cor(mfAve,pc1) # correlation of the average and the first PC (=0.999567) plot(pc1 ~ mfAve) # Explore a possible meningful linear combination suggested by second PC round(mfCoefs[,2],1) # Show coefficients of 2nd pc, rounded to 1 digit # Second set of coefficients are negative on the left end of the rod and # positive on the right end. Try Ave(L9 + L10 + L11) - Ave(L1 + L2 + L3). mfEnds <- (L9 + L10 + L11)/3 - (L1 + L2 + L3)/3 pc2 <- mfScores[,2] residualEnds <- lm(mfEnds ~ mfAve)$residual # Ends with average effect removed plot(pc2 ~ residualEnds) cor(residualEnds, pc2) # 0.973 # Explore a possible meaningful linear combination suggested by third PC round(mfCoefs[,3],1) # Show doefficients of 3rd pc, rounded to 1 digit # Try a contrast between the first 4 positions and the 6th position mfPeak <- (L1 + L2 + L3 + L4)/4 - L6 pc3 <- mfScores[,3] residualPeak <- lm(mfPeak ~ mfAve + mfEnds)$residual plot(pc3 ~ residualPeak) cor(residualPeak,pc3) # 0.971 # Note: the variation explained by the third PC seems to be due almost entirely # to one printer rod. (Keep this in mind.) ## INFERENCE: ANALYSIS OF EXPERIMENTAL FACTORS ON 3-DIMENSIONAL RESPONSE myResponse <- cbind(mfAve, mfEnds, mfPeak) cor(myResponse) myLm1 <- lm(myResponse ~ Current + Config + Material) anova(myLm1) # Noticeable effect of Current but not Config or Material plot(mfAve ~ Current) myLm2 <- lm(mfAve ~ Current) abline(myLm2) summary(myLm2) # No evidence of an effect of current on average magnetic force plot(mfEnds ~ Current) myLm3 <- lm(mfEnds ~ Current) abline(myLm3) summary(myLm3) # Evidence that Current effects the difference in MF at the ends plot(mfPeak ~ Current) myLm4 <- lm(mfPeak ~ Current) abline(myLm4) summary(myLm4) # No evidence of an effect of Current on peak MF in the center ## GRAPHICAL DISPLAY FOR PRESENTATION plot(mfEnds ~ jitter(Current,.1), xlab="Electrical Current Used in Printer Rod Manufacture (milliamperes)" , ylab="Magnetic Force at Positions 9-11 Minus Magnetic Force at Positions 1-3", main="Effect of Electrical Current on Magnetic Force Surrounding Printer Rod", col="black", pch=21, lwd=2, bg="green", cex=2 ) abline(myLm3, lwd=2) detach(case1701)
Thirty couples participated in a study of love and marriage. Wives and husbands responded separately to four questions:
What is the level of passionate love you feel for your spouse?
What is the level of passionate love your spouse feels for you?
What is the level of compassionate love you feel for your spouse?
What is the level of compassionate love your spouse feels for you?
Each response was recorded on a five-point scale: 1=None, 2=Very Little, 3=Some, 4=A Great Deal and 5=A Tremendous Amount.
case1702case1702
A data frame with 30 observations on the following 9 variables.
couple identification number
level of passionate love husband feels for spouse
level of passionate love wife feels for spouse
level of compassionate love husband feels for spouse
level of compassionate love wife feels for spouse
level of passionate love husband perceives spouse to have for him
level of passionate love wife perceives spouse to have for her
level of compassionate love husband perceives spouse to have for him
level of compassionate love husband perceives spouse to have for her
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Johnson, R.A. and Wichern, D.W. (1988). Applied Multivariate Statistical Analysis (2nd ed), Prentice-Hall.
str(case1702) attach(case1702) ## EXPLORATION x <- cbind(HP,WP,HC,WC) # 4 components of level of love you feel for spouse y <- cbind(PW,PH,CW,CH) # 4 components of level of love you perceive from spouse if(require(CCA)){ # Use the CCA library myCCA <- cc(x,y) # Store canonical correlation computations canCor <- myCCA$cor # Extract the canonical correlations canCor #[1] 0.9506990 0.8665601 0.5571876 0.1106555 # Make a function to test the number of canonical correlations (advanced). # Bartlett modification of likelihood ratio test # Reference: Mardia, Kent, and Bibby, 1980, Multivariate Analysis, myTest <- function(xMatrix,yMatrix) { if(require(CCA)){ # Use the CCA library myCCA <- cc(xMatrix,yMatrix) # Store CCA computations canCor <- myCCA$cor # extract the canonical correlations n <- dim(x)[1] # number of rows of x,= sample size p <- dim(y)[2] # number of component variables in y q <- dim(x)[2] # number of component variables in x k <- min(p,q) # the maximum number of canonical pairs testStat <- rep(0,k) # store the test statistics; initially set to 0 degFr <- rep(0,k) # store the associated degrees of freedom canCor2 <- canCor[k:1] # Reverse order for the following calculations productTerm <- 1 for (i in 1:k) { productTerm <- productTerm*(1 - canCor2[i]^2) degFr[i] <- (p + 1 - i)*(q + 1 - i) testStat[i] <- -(n -(p+q+3)/2)*log(productTerm) } pair <- 1:k testStat <- round(testStat[k:1],2) # Revert to original order; round pValue <- round(1 - pchisq(testStat,degFr),4) # p-value to 4 digits canCor <- round(canCor,4) # Round to 4 digits cbind(pair,canCor, testStat,degFr, pValue) # Show the results; } } myTest(x,y) # Explore possible meaningful linear combination suggested by first pair of CCs round(myCCA$xcoef,1) round(myCCA$ycoef,1) # The 1st column of xcoef is almost entirely HC; 1st column of ycoef is CW ccX1 <- myCCA$scores$xscores[,1] plot(ccX1 ~ jitter(HC)) # See if HC is a good substitute for 1st X canonical var cor(ccX1,HC) #[1] 0.9719947 ccY1 <- myCCA$scores$yscores[,1] plot(ccY1 ~ jitter(CW)) # See if CW is a good substitute for 1st y canonical var cor(ccY1, CW) #[1] 0.9975468 # Analyze the correlation of the meaningful substitute variables cor(HC,CW) # [1] 0.9280323 myLm1 <- lm(HC ~ CW) summary(myLm1) # p-value < 0.0001 (test for slope= 0 equiv to test that corr = 0) # Explore possible meaningful linear combination suggested by 2nd pair of CCs # Suggested substitutes from 2nd columns above are WC and CH ccX2 <- myCCA$scores$xscores[,2] WCres <- lm(WC ~ HC)$res # WC with effect of HC removed plot(ccX2 ~ WCres) cor(ccX2,WCres) #[1] 0.9045225 ccY2 <- myCCA$scores$yscores[,2] CHres <- lm(CH ~ CW)$res # CH with effect of CW removed plot(ccY2 ~ CHres) cor(ccY2,CHres) # [1] 0.9280248 cor(WC,CH) #[1] 0.8134213 myLm2 <- lm(WC ~ CH) summary(myLm2) # p-value < 0.0001 for test that correlation = 0 # Explore cannonical correlations from other groupings of variables x <- cbind(HP, HC, PW, CW) # husband's responses y <- cbind(WP, WC, PH, CH) # wife's responses myTest(x,y) # No evidence that husbands' responses are correlated with wifes' x <- cbind(HP,PW,WP,PH) # passionate responses y <- cbind(HC,CW,WC,CH) # compassionate responses myTest(x,y) #No evidence that passionate anc compassionate responses are correlated ## GRAPHICAL DISPLAYS FOR PRESENTATION jFactor <- 0.3 # Jittering factor (try different values to see what works) jHC <- jitter(HC,factor=jFactor) jCW <- jitter(CW,factor=jFactor) jCH <- jitter(CH,factor=jFactor) jWC <- jitter(WC,factor=jFactor) opar <- par(no.readonly=TRUE) # Store current graphical parameter settings par(mfrow=c(2,2)) # Prepare to make a 2x2 panel of graphs par(mar=c(1.1,4.1,1.1,1.1) ) # Adjust margins plot(jHC ~ jCW, ylab="Husband's Compassionate Love For His Wife (Jittered)", xlab="", ylim=c(3,5.1), xlim=c(3,5.1), col="black", pch=21, lwd=2, bg="green", cex=2 ) text(3,5.1,"correlation = 0.93",adj=0) text(3,5.0,"p-value < 0.0001",adj=0) abline(myLm1) par(mar=c(1.1,1.1,1.1,4.1)) plot(jHC ~ jCH, xlab="", ylab="", ylim=c(3,5.1), xlim=c(3,5.1), col="black", pch=21, lwd=2, bg="green", cex=2 ) cor(HC,CH) #[1] 0.274204 myLm3 <- lm(HC ~ CH) summary(myLm3)# p-value = 0.143 text(3,5.1,"correlation = 0.27",adj=0) text(3,5.0,"p-value = 0.14",adj=0) abline(myLm3) par(mar=c(4.1,4.1,1.1,1.1)) plot(jWC ~ jCW, xlab="Husband's Perceived Compassionate Love From His Wife (Jittered)", ylab="Wife's Compassionate Love For Her Husband (Jittered)", ylim=c(3,5.1), xlim=c(3,5.1), col="black", pch=21, lwd=2, bg="green", cex=2 ) cor(WC,CW) #[1] 0.04171195 myLm4 <- lm(WC ~ CW) summary(myLm4) # p-value = 0.827 text(3,3.1,"correlation = 0.04",adj=0) text(3,3,"p-value = 0.8",adj=0) abline(myLm4) par(mar=c(4.1,1.1,1.1,4.1)) plot(jWC ~ jCH, ylab="", xlab="Wife's Perceived Compassionate Love From Her Husband (Jittered)", ylim=c(3,5.1), xlim=c(3,5.1), col="black", pch=21, lwd=2, bg="green", cex=2 ) text(3,3.1,"correlation = 0.81",adj=0) text(3,3,"p-value < 0.0001",adj=0) abline(myLm2) par(opar) # Restore previous graphics parameter settings } detach(case1702)str(case1702) attach(case1702) ## EXPLORATION x <- cbind(HP,WP,HC,WC) # 4 components of level of love you feel for spouse y <- cbind(PW,PH,CW,CH) # 4 components of level of love you perceive from spouse if(require(CCA)){ # Use the CCA library myCCA <- cc(x,y) # Store canonical correlation computations canCor <- myCCA$cor # Extract the canonical correlations canCor #[1] 0.9506990 0.8665601 0.5571876 0.1106555 # Make a function to test the number of canonical correlations (advanced). # Bartlett modification of likelihood ratio test # Reference: Mardia, Kent, and Bibby, 1980, Multivariate Analysis, myTest <- function(xMatrix,yMatrix) { if(require(CCA)){ # Use the CCA library myCCA <- cc(xMatrix,yMatrix) # Store CCA computations canCor <- myCCA$cor # extract the canonical correlations n <- dim(x)[1] # number of rows of x,= sample size p <- dim(y)[2] # number of component variables in y q <- dim(x)[2] # number of component variables in x k <- min(p,q) # the maximum number of canonical pairs testStat <- rep(0,k) # store the test statistics; initially set to 0 degFr <- rep(0,k) # store the associated degrees of freedom canCor2 <- canCor[k:1] # Reverse order for the following calculations productTerm <- 1 for (i in 1:k) { productTerm <- productTerm*(1 - canCor2[i]^2) degFr[i] <- (p + 1 - i)*(q + 1 - i) testStat[i] <- -(n -(p+q+3)/2)*log(productTerm) } pair <- 1:k testStat <- round(testStat[k:1],2) # Revert to original order; round pValue <- round(1 - pchisq(testStat,degFr),4) # p-value to 4 digits canCor <- round(canCor,4) # Round to 4 digits cbind(pair,canCor, testStat,degFr, pValue) # Show the results; } } myTest(x,y) # Explore possible meaningful linear combination suggested by first pair of CCs round(myCCA$xcoef,1) round(myCCA$ycoef,1) # The 1st column of xcoef is almost entirely HC; 1st column of ycoef is CW ccX1 <- myCCA$scores$xscores[,1] plot(ccX1 ~ jitter(HC)) # See if HC is a good substitute for 1st X canonical var cor(ccX1,HC) #[1] 0.9719947 ccY1 <- myCCA$scores$yscores[,1] plot(ccY1 ~ jitter(CW)) # See if CW is a good substitute for 1st y canonical var cor(ccY1, CW) #[1] 0.9975468 # Analyze the correlation of the meaningful substitute variables cor(HC,CW) # [1] 0.9280323 myLm1 <- lm(HC ~ CW) summary(myLm1) # p-value < 0.0001 (test for slope= 0 equiv to test that corr = 0) # Explore possible meaningful linear combination suggested by 2nd pair of CCs # Suggested substitutes from 2nd columns above are WC and CH ccX2 <- myCCA$scores$xscores[,2] WCres <- lm(WC ~ HC)$res # WC with effect of HC removed plot(ccX2 ~ WCres) cor(ccX2,WCres) #[1] 0.9045225 ccY2 <- myCCA$scores$yscores[,2] CHres <- lm(CH ~ CW)$res # CH with effect of CW removed plot(ccY2 ~ CHres) cor(ccY2,CHres) # [1] 0.9280248 cor(WC,CH) #[1] 0.8134213 myLm2 <- lm(WC ~ CH) summary(myLm2) # p-value < 0.0001 for test that correlation = 0 # Explore cannonical correlations from other groupings of variables x <- cbind(HP, HC, PW, CW) # husband's responses y <- cbind(WP, WC, PH, CH) # wife's responses myTest(x,y) # No evidence that husbands' responses are correlated with wifes' x <- cbind(HP,PW,WP,PH) # passionate responses y <- cbind(HC,CW,WC,CH) # compassionate responses myTest(x,y) #No evidence that passionate anc compassionate responses are correlated ## GRAPHICAL DISPLAYS FOR PRESENTATION jFactor <- 0.3 # Jittering factor (try different values to see what works) jHC <- jitter(HC,factor=jFactor) jCW <- jitter(CW,factor=jFactor) jCH <- jitter(CH,factor=jFactor) jWC <- jitter(WC,factor=jFactor) opar <- par(no.readonly=TRUE) # Store current graphical parameter settings par(mfrow=c(2,2)) # Prepare to make a 2x2 panel of graphs par(mar=c(1.1,4.1,1.1,1.1) ) # Adjust margins plot(jHC ~ jCW, ylab="Husband's Compassionate Love For His Wife (Jittered)", xlab="", ylim=c(3,5.1), xlim=c(3,5.1), col="black", pch=21, lwd=2, bg="green", cex=2 ) text(3,5.1,"correlation = 0.93",adj=0) text(3,5.0,"p-value < 0.0001",adj=0) abline(myLm1) par(mar=c(1.1,1.1,1.1,4.1)) plot(jHC ~ jCH, xlab="", ylab="", ylim=c(3,5.1), xlim=c(3,5.1), col="black", pch=21, lwd=2, bg="green", cex=2 ) cor(HC,CH) #[1] 0.274204 myLm3 <- lm(HC ~ CH) summary(myLm3)# p-value = 0.143 text(3,5.1,"correlation = 0.27",adj=0) text(3,5.0,"p-value = 0.14",adj=0) abline(myLm3) par(mar=c(4.1,4.1,1.1,1.1)) plot(jWC ~ jCW, xlab="Husband's Perceived Compassionate Love From His Wife (Jittered)", ylab="Wife's Compassionate Love For Her Husband (Jittered)", ylim=c(3,5.1), xlim=c(3,5.1), col="black", pch=21, lwd=2, bg="green", cex=2 ) cor(WC,CW) #[1] 0.04171195 myLm4 <- lm(WC ~ CW) summary(myLm4) # p-value = 0.827 text(3,3.1,"correlation = 0.04",adj=0) text(3,3,"p-value = 0.8",adj=0) abline(myLm4) par(mar=c(4.1,1.1,1.1,4.1)) plot(jWC ~ jCH, ylab="", xlab="Wife's Perceived Compassionate Love From Her Husband (Jittered)", ylim=c(3,5.1), xlim=c(3,5.1), col="black", pch=21, lwd=2, bg="green", cex=2 ) text(3,3.1,"correlation = 0.81",adj=0) text(3,3,"p-value < 0.0001",adj=0) abline(myLm2) par(opar) # Restore previous graphics parameter settings } detach(case1702)
To better understand whether the relationship between heart disease and obesity could be due to the social stigma associated with obesity, researchers examined cardiovascular disease rates of obese and non-obese women in American Samoa, where obesity was considered socially desirable. 3,112 women were categorized according to whether they were obese or not and whether they died from cardiovascular disease (CVD).
case1801case1801
A data frame with 2 observations on the following 3 variables.
a factor with levels "NotObese" and
"obese"
the number of women who died from CVD
the number that died from other causes
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cenage Learning.
Crews, D.E. (1988). Cardiovascular Mortality in American Samoa, Human Biology 60: 417–433.
str(case1801) attach(case1801) ## EXPLORATION myTable <- cbind(Deaths,NonDeaths) # Form a 2 by 2 table of counts row.names(myTable) <- Obesity # Assign the levels of Obesity as row names myTable # Show the table ## INFERENCE (4 methods for getting p-values and confidence intervals) prop.test(myTable, alternative="greater", correct=FALSE) # Compare 2 proportions prop.test(myTable, alternative="greater", correct=TRUE) # ...with cont. corect. prop.test(myTable,correct=TRUE) # 2-sided alternative (default) to get CI chisq.test(myTable) # Pearson's Chi-Squared Test fisher.test(myTable, alternative="greater") # Fisher's exact test fisher.test(myTable) # 2-sided alternative to get CI for odds ratio myGlm1 <- glm(myTable ~ Obesity, family=binomial) # Logistic reg (CH 21) summary(myGlm1) # Get p-value-- 0.734 beta <- myGlm1$coef exp(beta[2]) #Odds of death are estimated to be 17% higher for obese women exp(confint(myGlm1,2)) # 95% confidence interval ## GRAPHICAL DISPLAY FOR PRESENTATION myTable # Deaths NonDeaths #Obese 16 2045 #NotObese 7 1044 prop.test(16,(16+2045)) #For one proportion, est: 0.0078 95% CI: 0.0046 to 0.013 prop.test(7,(7+1044)) #For one proportion, est: 0067 95% CI: 0.0029 to 0.014 pHat <- c(0.007763222, 0.006660324)*1000 # Get estimated deaths per 1,000 women lower95 <- c(0.00459943, 0.002921568)*1000 upper95 <- c(0.01287243, 0.014318321)*1000 if(require(Hmisc)) { # Use Hmisc library myObj <- Cbind(pHat,lower95,upper95) Dotplot(Obesity ~ myObj, # Draw a dot plot of estimates and CIs xlab="Estimated CVD Deaths Per 1,000 Women (and 95% Confidence Intervals)", ylab="Weight Category", ylim=c(.5,2.5), cex=2) } detach(case1801)str(case1801) attach(case1801) ## EXPLORATION myTable <- cbind(Deaths,NonDeaths) # Form a 2 by 2 table of counts row.names(myTable) <- Obesity # Assign the levels of Obesity as row names myTable # Show the table ## INFERENCE (4 methods for getting p-values and confidence intervals) prop.test(myTable, alternative="greater", correct=FALSE) # Compare 2 proportions prop.test(myTable, alternative="greater", correct=TRUE) # ...with cont. corect. prop.test(myTable,correct=TRUE) # 2-sided alternative (default) to get CI chisq.test(myTable) # Pearson's Chi-Squared Test fisher.test(myTable, alternative="greater") # Fisher's exact test fisher.test(myTable) # 2-sided alternative to get CI for odds ratio myGlm1 <- glm(myTable ~ Obesity, family=binomial) # Logistic reg (CH 21) summary(myGlm1) # Get p-value-- 0.734 beta <- myGlm1$coef exp(beta[2]) #Odds of death are estimated to be 17% higher for obese women exp(confint(myGlm1,2)) # 95% confidence interval ## GRAPHICAL DISPLAY FOR PRESENTATION myTable # Deaths NonDeaths #Obese 16 2045 #NotObese 7 1044 prop.test(16,(16+2045)) #For one proportion, est: 0.0078 95% CI: 0.0046 to 0.013 prop.test(7,(7+1044)) #For one proportion, est: 0067 95% CI: 0.0029 to 0.014 pHat <- c(0.007763222, 0.006660324)*1000 # Get estimated deaths per 1,000 women lower95 <- c(0.00459943, 0.002921568)*1000 upper95 <- c(0.01287243, 0.014318321)*1000 if(require(Hmisc)) { # Use Hmisc library myObj <- Cbind(pHat,lower95,upper95) Dotplot(Obesity ~ myObj, # Draw a dot plot of estimates and CIs xlab="Estimated CVD Deaths Per 1,000 Women (and 95% Confidence Intervals)", ylab="Weight Category", ylim=c(.5,2.5), cex=2) } detach(case1801)
In a randomized experiment, researchers assigned 407 volunteers to receive 1,000 mg of Vitamin C daily throughout the cold season and 411 to receive a placebo. A physician interviewed the volunteers at the end of the study to determine whether or not they had suffered any colds during the study period.
case1802case1802
A data frame with 2 observations on the following 3 variables.
a factor with levels "Placebo" and
"VitC"
the number of who got colds
the number that did not get any colds
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cenage Learning.
Anderson, T.W., Reid, D.B.W. and Beaton, G. H. (1972). Vitamin C and the Common Cold, Canadian Medial Association Journal 107: 503–508.
str(case1802) attach(case1802) library(MASS) ## INFERENCE (4 methods) myTable <- cbind(Cold,NoCold) row.names(myTable) <- c("Placebo","Vitamin C") myTable prop.test(myTable, alternative="greater") # Compare 2 binomial proportions # Alternative: pop prop. of first column (cold) in larger in first row (placebo) prop.test(myTable, alternative="greater", correct=TRUE) prop.test(myTable,correct=TRUE) # Use 2-sided alternative to get CI chisq.test(myTable) # Chi-square test fisher.test(myTable, alternative="greater") fisher.test(myTable) # 2-sided alternative to get CI for odds ratio myGlm1 <- glm(myTable ~ Treatment, family=binomial) # logistic reg (Ch 21) summary(myGlm1) beta <- myGlm1$coef 1 - exp(beta[2]) # 0.3474911 1 - exp(confint(myGlm1,2)) # 0.53365918 0.09042098 # Interpretation: The odds of getting a cold are 35% less on Vitamin C than # Placebo (95% confidence interval: 9% to 53% less). detach(case1802)str(case1802) attach(case1802) library(MASS) ## INFERENCE (4 methods) myTable <- cbind(Cold,NoCold) row.names(myTable) <- c("Placebo","Vitamin C") myTable prop.test(myTable, alternative="greater") # Compare 2 binomial proportions # Alternative: pop prop. of first column (cold) in larger in first row (placebo) prop.test(myTable, alternative="greater", correct=TRUE) prop.test(myTable,correct=TRUE) # Use 2-sided alternative to get CI chisq.test(myTable) # Chi-square test fisher.test(myTable, alternative="greater") fisher.test(myTable) # 2-sided alternative to get CI for odds ratio myGlm1 <- glm(myTable ~ Treatment, family=binomial) # logistic reg (Ch 21) summary(myGlm1) beta <- myGlm1$coef 1 - exp(beta[2]) # 0.3474911 1 - exp(confint(myGlm1,2)) # 0.53365918 0.09042098 # Interpretation: The odds of getting a cold are 35% less on Vitamin C than # Placebo (95% confidence interval: 9% to 53% less). detach(case1802)
In a retrospective case-control study, researchers identified 86 lung cancer patients and 86 controls (without lung cancer), and categorized them according to whether they were smokers or non-smokers. The goal is to see whether the odds of lung cancer are greater for smokers than for non-smokers.
case1803case1803
A data frame with 2 observations on the following 3 variables.
a factor with levels "NonSmokers" and
"Smokers"
the number of who were lung cancer patients
the number who were controls
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cenage Learning.
Anderson, T.W., Reid, D.B.W. and Beaton, G. H. (1972). Vitamin C and the Common Cold, Canadian Medial Association Journal 107: 503–508.
str(case1803) attach(case1803) ## INFERENCE myTable <- cbind(Cancer,Control) # Make a 2-by-2 table of counts row.names(myTable) <- Smoking # Assign the levels of Smoking as row names myTable fisher.test(myTable, alternative="greater") # Alternative: that odds of Cancer # in first row are greater. fisher.test(myTable) # 2-sided alternative to get CI for odds ratio myGlm1 <- glm(myTable ~ Smoking, family=binomial) # logistic reg (Ch 21) summary(myGlm1) exp(myGlm1$coef[2]) # 5.37963 : Estimated odds ratio exp(confint(myGlm1)[2,]) # 1.675169 24.009510: Approximate confidence interval # Interpretation: The odds of cancer ar 5.4 times as large for smokers as for # non-smokers (95% confidence interval: 1.7 to 24.0 times as large). detach(case1803)str(case1803) attach(case1803) ## INFERENCE myTable <- cbind(Cancer,Control) # Make a 2-by-2 table of counts row.names(myTable) <- Smoking # Assign the levels of Smoking as row names myTable fisher.test(myTable, alternative="greater") # Alternative: that odds of Cancer # in first row are greater. fisher.test(myTable) # 2-sided alternative to get CI for odds ratio myGlm1 <- glm(myTable ~ Smoking, family=binomial) # logistic reg (Ch 21) summary(myGlm1) exp(myGlm1$coef[2]) # 5.37963 : Estimated odds ratio exp(confint(myGlm1)[2,]) # 1.675169 24.009510: Approximate confidence interval # Interpretation: The odds of cancer ar 5.4 times as large for smokers as for # non-smokers (95% confidence interval: 1.7 to 24.0 times as large). detach(case1803)
Researchers gave 48 male bank supervisors attending a management institute hypothetical personnel files and asked them whether they would promote the applicant based on the file. The personnel files were identical except that 24 of them listed a male and 24 listed a female applicant. The assignment of managers to receive either a male or female applicant file was carried out at random.
case1901case1901
A data frame with 2 observations on the following 3 variables.
a factor with levels "Female" and "Male"
the number of managers who promoted the applicant
the number of managers who did not promote the applicant
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Rosen, B. and Jerdee, J (1974). Influence of Sex Role Steroetypes on Personnel Decisions, Journal of Applied Psychology 59: 9–14.
str(case1901) attach(case1901) ## INFERENCE myTable <- cbind(Promoted,NotPromoted) row.names(myTable) <- Gender myTable fisher.test(myTable, alternative="greater") # Alternative: that odds of Promotion in first row (Males) are greater. fisher.test(myTable) # Use 2-sided to get confidence interval for odds ratio prop.test(myTable) # Compare two binomial proportions ## GRAPHICAL DISPLAY FOR PRESENTATION myTable # Promoted NotPromoted #Male 21 3 #Female 14 10 prop.test(21,(21+3)) # Est = .875; CI = .665 to .967 prop.test(14,(14+10))# Est = .583; CI = .369 to .772 pHat <- c(0.875,0.583) lower95 <- c(0.665, 0.369) upper95 <- c(0.967, 0.772) if(require(Hmisc)) { # Use Hmisc library myObj<- Cbind(pHat,lower95,upper95) # Cbind: a form of cbind needed for Dotplot Dotplot(Gender ~ myObj, xlab="Probability of Promotion Based on Applicant File (and 95% Confidence Intervals)", ylab="Gender Listed in Applicant File", ylim=c(.5,2.5), cex=2) } detach(case1901)str(case1901) attach(case1901) ## INFERENCE myTable <- cbind(Promoted,NotPromoted) row.names(myTable) <- Gender myTable fisher.test(myTable, alternative="greater") # Alternative: that odds of Promotion in first row (Males) are greater. fisher.test(myTable) # Use 2-sided to get confidence interval for odds ratio prop.test(myTable) # Compare two binomial proportions ## GRAPHICAL DISPLAY FOR PRESENTATION myTable # Promoted NotPromoted #Male 21 3 #Female 14 10 prop.test(21,(21+3)) # Est = .875; CI = .665 to .967 prop.test(14,(14+10))# Est = .583; CI = .369 to .772 pHat <- c(0.875,0.583) lower95 <- c(0.665, 0.369) upper95 <- c(0.967, 0.772) if(require(Hmisc)) { # Use Hmisc library myObj<- Cbind(pHat,lower95,upper95) # Cbind: a form of cbind needed for Dotplot Dotplot(Gender ~ myObj, xlab="Probability of Promotion Based on Applicant File (and 95% Confidence Intervals)", ylab="Gender Listed in Applicant File", ylim=c(.5,2.5), cex=2) } detach(case1901)
Lawyers collected data on convicted black murderers in the state of Georgia to see whether convicted black murderers whose victim was white were more likely to receive the death penalty than those whose victim was black, after accounting for aggravation level of the murder. They categorized murders into 6 progressively more serious types. Category 1 comprises barroom brawls, liquor-induced arguments lovers' quarrels, and similar crimes. Category 6 includes the most vicious, cruel, cold=blooded, unprovoked crimes.
case1902case1902
A data frame with 12 observations on the following 4 variables.
the aggravation level of the crime, a numerical variable ranging from 1 to 6
a factor indicating race of murder victim, with
levels "White" and "Black"
number in the aggravation and victim category who received the death penalty
number in the aggravation and victim category who did not receive the death penalty
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Woodworth, G.C. (1989). Statistics and the Death Penalty, Stats 2: 9–12.
str(case1902) attach(case1902) ## EXPLORATION proportionDeath <- Death/(Death + NoDeath) myPointCode <- ifelse(Victim=="White",22,24) myPointColor <- ifelse(Victim=="White","white","black") plot(proportionDeath ~ Aggravation, pch=myPointCode, bg=myPointColor) oddsOfDeath <- Death/(NoDeath + .5) # Add .5 to the demoninator to avoid 0's plot(oddsOfDeath ~ Aggravation, pch=myPointCode, bg=myPointColor) plot(oddsOfDeath ~ Aggravation, log="y", pch=myPointCode, bg=myPointColor) # Use logistic regression (Ch 21) to see if the 6 odds ratios are constant myGlm1 <- glm(cbind(Death,NoDeath) ~ Aggravation + Victim + Aggravation:Victim, family=binomial) # Logistic reg with interaction myGlm2 <- update(myGlm1, ~ . - Aggravation:Victim) # without interaction anova(myGlm2, myGlm1) # no evidence of interaction. ## INFERENCE # Mantel Haenszel myTable <- array(rbind(Death, NoDeath), dim=c(2,2,6), dimnames=list(Penalty=c("Death","No Death"), Victim=c("White","Black"), Aggravation=c("1","2","3","4","5","6"))) myTable # Show the 6 2x2 tables mantelhaen.test(myTable, alternative="greater", correct=FALSE) # 1-sided p-value mantelhaen.test(myTable, alternative="greater") # with continuity correction mantelhaen.test(myTable) # two.sided (default) for confidence interval # Logistic Regression (Ch 21) (treating aggravation level as numerical) summary(myGlm2) beta <- myGlm2$coef exp(beta[3]) # 6.1144 exp(confint(myGlm2,3)) # 2.23040 18.72693 # Interpretation: The odds of death penalty for white victim murderers are # estimated to be 6 times the odds of death penalty for black victim murderers # with similar aggravation level(95% confidence interval: 2.2 to 18.7 times). ## GRAPHICAL DISPLAY FOR PRESENTATION myPointColor <- ifelse(Victim=="White","green", "orange") plot(jitter(proportionDeath,.1) ~ jitter(Aggravation,.1), xlab="Aggravation Level of the Murder", ylab="Proportion of Murderers Who Received Death Penalty", pch=myPointCode, bg=myPointColor, cex=2, lwd=2) legend(1,1, c("White Victim Murderers","Black Victim Murderers"), pch=c(21,22), pt.cex=c(2,2), pt.bg=c("green","orange"), pt.lw=c(2,2)) # Include logistic regression fit on plot dummyAg <- seq(min(Aggravation),max(Aggravation),length=50) etaB <- beta[1] + beta[2]*dummyAg etaW <- etaB + beta[3] pB <- exp(etaB)/(1 + exp(etaB)) # Estimated prob of DP; Black victim pW <- exp(etaW)/(1 + exp(etaW)) # Estimated prob of DP; White victim lines(pB ~ dummyAg,lty=1) lines(pW ~ dummyAg,lty=2) detach(case1902)str(case1902) attach(case1902) ## EXPLORATION proportionDeath <- Death/(Death + NoDeath) myPointCode <- ifelse(Victim=="White",22,24) myPointColor <- ifelse(Victim=="White","white","black") plot(proportionDeath ~ Aggravation, pch=myPointCode, bg=myPointColor) oddsOfDeath <- Death/(NoDeath + .5) # Add .5 to the demoninator to avoid 0's plot(oddsOfDeath ~ Aggravation, pch=myPointCode, bg=myPointColor) plot(oddsOfDeath ~ Aggravation, log="y", pch=myPointCode, bg=myPointColor) # Use logistic regression (Ch 21) to see if the 6 odds ratios are constant myGlm1 <- glm(cbind(Death,NoDeath) ~ Aggravation + Victim + Aggravation:Victim, family=binomial) # Logistic reg with interaction myGlm2 <- update(myGlm1, ~ . - Aggravation:Victim) # without interaction anova(myGlm2, myGlm1) # no evidence of interaction. ## INFERENCE # Mantel Haenszel myTable <- array(rbind(Death, NoDeath), dim=c(2,2,6), dimnames=list(Penalty=c("Death","No Death"), Victim=c("White","Black"), Aggravation=c("1","2","3","4","5","6"))) myTable # Show the 6 2x2 tables mantelhaen.test(myTable, alternative="greater", correct=FALSE) # 1-sided p-value mantelhaen.test(myTable, alternative="greater") # with continuity correction mantelhaen.test(myTable) # two.sided (default) for confidence interval # Logistic Regression (Ch 21) (treating aggravation level as numerical) summary(myGlm2) beta <- myGlm2$coef exp(beta[3]) # 6.1144 exp(confint(myGlm2,3)) # 2.23040 18.72693 # Interpretation: The odds of death penalty for white victim murderers are # estimated to be 6 times the odds of death penalty for black victim murderers # with similar aggravation level(95% confidence interval: 2.2 to 18.7 times). ## GRAPHICAL DISPLAY FOR PRESENTATION myPointColor <- ifelse(Victim=="White","green", "orange") plot(jitter(proportionDeath,.1) ~ jitter(Aggravation,.1), xlab="Aggravation Level of the Murder", ylab="Proportion of Murderers Who Received Death Penalty", pch=myPointCode, bg=myPointColor, cex=2, lwd=2) legend(1,1, c("White Victim Murderers","Black Victim Murderers"), pch=c(21,22), pt.cex=c(2,2), pt.bg=c("green","orange"), pt.lw=c(2,2)) # Include logistic regression fit on plot dummyAg <- seq(min(Aggravation),max(Aggravation),length=50) etaB <- beta[1] + beta[2]*dummyAg etaW <- etaB + beta[3] pB <- exp(etaB)/(1 + exp(etaB)) # Estimated prob of DP; Black victim pW <- exp(etaW)/(1 + exp(etaW)) # Estimated prob of DP; White victim lines(pB ~ dummyAg,lty=1) lines(pW ~ dummyAg,lty=2) detach(case1902)
This data frame contains the ages and sexes of the adult (over 15 years) survivors and nonsurvivors of the Donner party.
case2001case2001
A data frame with 45 observations on the following 3 variables.
Age of person
Sex of person
Whether the person survived or died
In 1846 the Donner and Reed families left Springfield, Illinois, for California by covered wagon. In July, the Donner Party, as it became known, reached Fort Bridger, Wyoming. There its leaders decided to attempt a new and untested rote to the Sacramento Valley. Having reached its full size of 87 people and 20 wagons, the party was delayed by a difficult crossing of the Wasatch Range and again in the crossing of the desert west of the Great Salt Lake. The group became stranded in the eastern Sierra Nevada mountains when the region was hit by heavy snows in late October. By the time the last survivor was rescued on April 21, 1847, 40 of the 87 members had died from famine and exposure to extreme cold.
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Grayson, D.K. (1990). Donner Party Deaths: A Demographic Assessment, Journal of Anthropological Research 46: 223–242.
str(case2001) attach(case2001) ## EXPLORATION AND MODEL BUILDING myPointCode <- ifelse(Sex=="Female",22,24) myPointColor <- ifelse(Sex=="Female","green","orange") survivalIndicator <- ifelse(Status=="Survived",1,0) jFactor <- 0.1 # jittering factor plot(jitter(survivalIndicator,jFactor) ~ jitter(Age, jFactor), pch=myPointCode, bg=myPointColor, cex=1.5) # Logistic regression. Start with a rich model; use backward elimination ageSquared <- Age^2 myGlm1 <- glm(Status ~ Age + ageSquared + Sex + Age:Sex + ageSquared:Sex, family=binomial) # Use backward elimination, but remove interaction and squared terms 1st summary(myGlm1) myGlm2 <- update(myGlm1, ~ . - ageSquared:Sex) summary(myGlm2) myGlm3 <- update(myGlm2, ~ . - ageSquared) summary(myGlm3) # Wald test p-value for interaction of Age and Sex is: 0.0865 # More accurate likelihood ratio (drop in deviance) test: myGlm4 <-update(myGlm3, ~ . - Age:Sex) anova(myGlm4, myGlm3) # Drop-in-devaince chi-square stat = 3.9099 on 1 d.f. 1 - pchisq(3.9099,1) # 2-sided p-value = 0.048 ## INFERENCE AND INTERPRETATION # Proceed by ignoring interaction (for a casual and approximate analysis) myGlm5 <- update(myGlm4, ~ . - Sex) anova(myGlm5, myGlm4) # Drop-in-deviance chi-square statistic = 5.0344 on 1 d.f. 1 - pchisq(5.0344,1) # 2-sided p-value 0.02484869: Highly suggestive 0.0248869/2 # 1-sided p-value = half the 2-sided p-value = 0.01244345 # Interpretation and confidence interval Sex <- factor(Sex,levels=c("Male","Female")) # Reorder levels so "Male" is ref myGlm4b <- glm(Status ~ Age + Sex, family=binomial) beta <- myGlm4b$coef exp(beta[3]) # 4.939645 exp(confint(myGlm4b,3)) # 25.246069 1.215435 # Interpretation: The odds of survival for females are estimated to be 5 times # the odds of survival of similarly-aged mean (95% CI: 1.2 times to 25.2 times). ## GRAPHICAL DISPLAY FOR PRESENTATION myPointCode <- ifelse(Sex=="Female",22,24) myPointColor <- ifelse(Sex=="Female","green","orange") myLineColor <- ifelse(Sex=="Female","dark green","blue") survivalIndicator <- ifelse(Status=="Survived",1,0) jFactor <- 0.1 plot(jitter(survivalIndicator,jFactor) ~ jitter(Age, jFactor), ylab="Estimated Survival Probability", xlab="Age (years)", main=c("Donner Party Survival by Sex and Age"), xlim=c(15,75), pch=myPointCode, bg=myPointColor, col=myLineColor, cex=2, lwd=3) beta <- myGlm4b$coef dummyAge <- seq(15,65,length=50) linearMale <- beta[1] + beta[2]*dummyAge #log odds of survival for males linearFemale <- linearMale + beta[3] #log odds of survival for females pCurveMale <- exp(linearMale)/(1 + exp(linearMale) ) # survival prob; males pCurveFemale <- exp(linearFemale)/(1 + exp(linearFemale)) # females lines(pCurveMale ~ dummyAge,lty=2,col="blue",lwd=3) lines(pCurveFemale[dummyAge <= 50] ~ dummyAge[dummyAge <= 50],lty=1, col="dark green",lwd=3) legend(63,.5,legend=c("Females","Males"), pch=c(22,24), pt.bg = c("green","orange"), pt.cex=c(2,2), lty=c(1,2), lwd=c(3,3), col=c("dark green","blue")) text(72,1,"Survived (20)") text(72,0,"Died (25)") detach(case2001)str(case2001) attach(case2001) ## EXPLORATION AND MODEL BUILDING myPointCode <- ifelse(Sex=="Female",22,24) myPointColor <- ifelse(Sex=="Female","green","orange") survivalIndicator <- ifelse(Status=="Survived",1,0) jFactor <- 0.1 # jittering factor plot(jitter(survivalIndicator,jFactor) ~ jitter(Age, jFactor), pch=myPointCode, bg=myPointColor, cex=1.5) # Logistic regression. Start with a rich model; use backward elimination ageSquared <- Age^2 myGlm1 <- glm(Status ~ Age + ageSquared + Sex + Age:Sex + ageSquared:Sex, family=binomial) # Use backward elimination, but remove interaction and squared terms 1st summary(myGlm1) myGlm2 <- update(myGlm1, ~ . - ageSquared:Sex) summary(myGlm2) myGlm3 <- update(myGlm2, ~ . - ageSquared) summary(myGlm3) # Wald test p-value for interaction of Age and Sex is: 0.0865 # More accurate likelihood ratio (drop in deviance) test: myGlm4 <-update(myGlm3, ~ . - Age:Sex) anova(myGlm4, myGlm3) # Drop-in-devaince chi-square stat = 3.9099 on 1 d.f. 1 - pchisq(3.9099,1) # 2-sided p-value = 0.048 ## INFERENCE AND INTERPRETATION # Proceed by ignoring interaction (for a casual and approximate analysis) myGlm5 <- update(myGlm4, ~ . - Sex) anova(myGlm5, myGlm4) # Drop-in-deviance chi-square statistic = 5.0344 on 1 d.f. 1 - pchisq(5.0344,1) # 2-sided p-value 0.02484869: Highly suggestive 0.0248869/2 # 1-sided p-value = half the 2-sided p-value = 0.01244345 # Interpretation and confidence interval Sex <- factor(Sex,levels=c("Male","Female")) # Reorder levels so "Male" is ref myGlm4b <- glm(Status ~ Age + Sex, family=binomial) beta <- myGlm4b$coef exp(beta[3]) # 4.939645 exp(confint(myGlm4b,3)) # 25.246069 1.215435 # Interpretation: The odds of survival for females are estimated to be 5 times # the odds of survival of similarly-aged mean (95% CI: 1.2 times to 25.2 times). ## GRAPHICAL DISPLAY FOR PRESENTATION myPointCode <- ifelse(Sex=="Female",22,24) myPointColor <- ifelse(Sex=="Female","green","orange") myLineColor <- ifelse(Sex=="Female","dark green","blue") survivalIndicator <- ifelse(Status=="Survived",1,0) jFactor <- 0.1 plot(jitter(survivalIndicator,jFactor) ~ jitter(Age, jFactor), ylab="Estimated Survival Probability", xlab="Age (years)", main=c("Donner Party Survival by Sex and Age"), xlim=c(15,75), pch=myPointCode, bg=myPointColor, col=myLineColor, cex=2, lwd=3) beta <- myGlm4b$coef dummyAge <- seq(15,65,length=50) linearMale <- beta[1] + beta[2]*dummyAge #log odds of survival for males linearFemale <- linearMale + beta[3] #log odds of survival for females pCurveMale <- exp(linearMale)/(1 + exp(linearMale) ) # survival prob; males pCurveFemale <- exp(linearFemale)/(1 + exp(linearFemale)) # females lines(pCurveMale ~ dummyAge,lty=2,col="blue",lwd=3) lines(pCurveFemale[dummyAge <= 50] ~ dummyAge[dummyAge <= 50],lty=1, col="dark green",lwd=3) legend(63,.5,legend=c("Females","Males"), pch=c(22,24), pt.bg = c("green","orange"), pt.cex=c(2,2), lty=c(1,2), lwd=c(3,3), col=c("dark green","blue")) text(72,1,"Survived (20)") text(72,0,"Died (25)") detach(case2001)
A 1972–1981 health survey in The Hague, Netherlands, discovered an association between keeping pet birds and increased risk of lung cancer. To investigate birdkeeping as a risk factor, researchers conducted a case–control study of patients in 1985 at four hospitals in The Hague (population 450,000). They identified 49 cases of lung cancer among the patients who were registered with a general practice, who were age 65 or younger and who had resided in the city since 1965. They also selected 98 controls from a population of residents having the same general age structure.
case2002case2002
A data frame with 147 observations on the following 7 variables.
Whether subject has lung cancer
Sex of subject
Socioeconomic status, determined by occupation of the household's principal wage earner
Indicator for birdkeeping (caged birds in the home for more that 6 consecutive months from 5 to 14 years before diagnosis (cases) or examination (control))
Age of subject (in years)
Years of smoking prior to diagnosis or examination
Average rate of smoking (in cigarettes per day)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Holst, P.A., Kromhout, D. and Brand, R. (1988). For Debate: Pet Birds as an Independent Risk Factor for Lung Cancer, British Medical Journal 297: 13–21.
str(case2002) attach(case2002) ## EXPLORATION AND MODEL BUILDING myCode <- ifelse(BK=="Bird" & LC=="LungCancer","Bird & Cancer", ifelse(BK=="Bird" & LC=="NoCancer","Bird & No Cancer", ifelse(BK=="NoBird" & LC=="LungCancer","No Bird & Cancer", "No Bird & No Cancer"))) table(myCode) if(require(car)){ # Use the car library scatterplotMatrix(cbind(AG,YR,CD), groups=myCode, diagonal="none",reg.line=FALSE, pch=c(15,21,15,21), col=c("dark green","dark green","purple","purple"), var.labels=c("Age","Years Smoked","Cigarettes per Day"), cex=1.5) } # Reorder the levels so that the model is for log odds of cancer LC <- factor(LC, levels=c("NoCancer","LungCancer")) myGlm <- glm(LC ~ FM + SS + AG + YR + CD + BK, family=binomial) if(require(car)){ # Use the car library crPlots(myGlm) } # It appears that there's an effect of Years of Smoking and of Bird Keeping # after accounting for other variables; no obvious effects of other variables # Logistic regression model building using backward elimination (witholding BK) myGlm1 <- glm(LC ~ FM + SS + AG + YR + CD, family=binomial) summary(myGlm1) myGlm2 <- update(myGlm1, ~ . - SS) summary(myGlm2) myGlm3 <- update(myGlm2, ~ . - CD) summary(myGlm3) myGlm4 <- update(myGlm3, ~ . - FM) summary(myGlm4) # Everything left has a small p-value (retain the intercept) ## INFERENCE AND INTERPRETATION myGlm5 <- update(myGlm4, ~ . + BK) # Now add bird keeping summary(myGlm5) myGlm6 <- update(myGlm5, ~ . + BK:YR + AG:YR) # Try interaction terms anova(myGlm6,myGlm5) # Drop-in-deviance = 1.61 on 2 d.f. 1 - pchisq(1.61,2) # p-value = .45: no evidence of interaction anova(myGlm4,myGlm5) # Test for bird keeping effect (1 - pchisq(12.612,1))/2 # 1-sided p-value: 0.0001916391 BK <- factor(BK, levels=c("NoBird", "Bird")) # Make "no bird" the ref level myGlm5b <- glm(LC ~ AG + YR + BK, family=binomial) beta <- myGlm5b$coef # Extract estimated coefficients exp(beta[4]) # 3.961248 exp(confint(myGlm5b,4)) # 1.836764 8.900840 # Interpretation: The odds of lung cancer for people who kept birds were # estimated to be 4 times the odds of lung cancer for people of similar age, sex, # smoking history, and socio-economic status who didn't keep birds # (95% confidence interval for this adjusted odds ratio: 1.8 times to 8.9 times). # See bestglm library for an alternative variable selection technique. detach(case2002)str(case2002) attach(case2002) ## EXPLORATION AND MODEL BUILDING myCode <- ifelse(BK=="Bird" & LC=="LungCancer","Bird & Cancer", ifelse(BK=="Bird" & LC=="NoCancer","Bird & No Cancer", ifelse(BK=="NoBird" & LC=="LungCancer","No Bird & Cancer", "No Bird & No Cancer"))) table(myCode) if(require(car)){ # Use the car library scatterplotMatrix(cbind(AG,YR,CD), groups=myCode, diagonal="none",reg.line=FALSE, pch=c(15,21,15,21), col=c("dark green","dark green","purple","purple"), var.labels=c("Age","Years Smoked","Cigarettes per Day"), cex=1.5) } # Reorder the levels so that the model is for log odds of cancer LC <- factor(LC, levels=c("NoCancer","LungCancer")) myGlm <- glm(LC ~ FM + SS + AG + YR + CD + BK, family=binomial) if(require(car)){ # Use the car library crPlots(myGlm) } # It appears that there's an effect of Years of Smoking and of Bird Keeping # after accounting for other variables; no obvious effects of other variables # Logistic regression model building using backward elimination (witholding BK) myGlm1 <- glm(LC ~ FM + SS + AG + YR + CD, family=binomial) summary(myGlm1) myGlm2 <- update(myGlm1, ~ . - SS) summary(myGlm2) myGlm3 <- update(myGlm2, ~ . - CD) summary(myGlm3) myGlm4 <- update(myGlm3, ~ . - FM) summary(myGlm4) # Everything left has a small p-value (retain the intercept) ## INFERENCE AND INTERPRETATION myGlm5 <- update(myGlm4, ~ . + BK) # Now add bird keeping summary(myGlm5) myGlm6 <- update(myGlm5, ~ . + BK:YR + AG:YR) # Try interaction terms anova(myGlm6,myGlm5) # Drop-in-deviance = 1.61 on 2 d.f. 1 - pchisq(1.61,2) # p-value = .45: no evidence of interaction anova(myGlm4,myGlm5) # Test for bird keeping effect (1 - pchisq(12.612,1))/2 # 1-sided p-value: 0.0001916391 BK <- factor(BK, levels=c("NoBird", "Bird")) # Make "no bird" the ref level myGlm5b <- glm(LC ~ AG + YR + BK, family=binomial) beta <- myGlm5b$coef # Extract estimated coefficients exp(beta[4]) # 3.961248 exp(confint(myGlm5b,4)) # 1.836764 8.900840 # Interpretation: The odds of lung cancer for people who kept birds were # estimated to be 4 times the odds of lung cancer for people of similar age, sex, # smoking history, and socio-economic status who didn't keep birds # (95% confidence interval for this adjusted odds ratio: 1.8 times to 8.9 times). # See bestglm library for an alternative variable selection technique. detach(case2002)
In a study of the Krunnit Islands archipelago, researchers presented results of extensive bird surveys taken over four decades. They visited each island several times, cataloguing species. If a species was found on a specific island in 1949, it was considered to be at risk of extinction for the next survey of the island in 1959. If it was not found in 1959, it was counted as an “extinction”, even though it might reappear later. This data frame contains data on island size, number of species at risk to become extinct and number of extinctions.
case2101case2101
A data frame with 18 observations on the following 4 variables.
Name of Island
Area of Island
Number of species at risk
Number of extinctions
Scientists agree that preserving certain habitats in their natural states is necessary to slow the accelerating rate of species extinctions. But they are divided on how to construct such reserves. Given a finite amount of available land, is it better to have many small reserves or a few large one? Central to the debate on this question are observational studies of what has happened in island archipelagos, where nearly the same fauna tries to survive on islands of different sizes.
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Väisänen, R.A. and Järvinen, O. (1977). Dynamics of Protected Bird Communities in a Finnish Archipelago, Journal of Animal Ecology 46: 891–908.
str(case2101) attach(case2101) ## EXPLORATION AND MODEL BUILDING proportionExtinct <- Extinct/AtRisk oddsExtinct <- proportionExtinct/(1 - proportionExtinct) logitExtinct <- log(oddsExtinct) # Logit = Log Odds logArea <- log(Area) plot(logitExtinct ~ logArea) binResponse <- cbind(Extinct,AtRisk-Extinct) myGlm1 <- glm(binResponse ~ logArea, family=binomial) summary(myGlm1) logArea2 <- logArea^2 myGlm2 <- update(myGlm1, ~ . + logArea2) summary(myGlm2) # p-value for quadratic term: 0.77 ## INFERENCE AND INTERPRETATION myGlm3 <- update(myGlm1, ~ . - logArea) anova(myGlm3, myGlm1) # Drop in deviance statistic = 33.277 on 1 d.f. 1 - pchisq(33.277,1) # p-value = 7.992234e-09 beta <- myGlm1$coef 1 - 2^beta[2] # 0.1861153 1 - 2^confint(myGlm1,2) #0.2462041 0.1247743 # Interpretation: Associated with each doubling of island area is a 19% # reduction in the odds of extinction (95% confidence interval: 12% to 25% # reduction). ## GRAPHICAL DISPLAY FOR PRESENTATION plot(oddsExtinct ~ Area, log="xy", ylab="Observed Odds of Extinction; log scale", xlab=expression(paste("Island Area (km"^"2","); log scale")), main="Extinctions of Bird Species in the Krunnit Island Archipelago", pch=21, lwd=2, bg="green", cex=2) # Plot odds of extinction vs island area dummyArea <- seq(min(Area),max(Area),length=50) lp <- beta[1] + beta[2]*log(dummyArea) odds <- exp(lp) lines(odds ~ dummyArea,lwd=2) plot(proportionExtinct ~ Area, log="xy", ylab="Proportions of 1949 Species not Found in 1959", xlab=expression(paste("Island Area (km"^"2","); log scale")), main="Proportions of 1949 Bird Species Extinct in 1959 on 18 Krunnit Archipelago Islands", pch=21, lwd=2, bg="green", cex=2) # Plot probability of extinction vs area dummyArea <- seq(min(Area),max(Area),length=50) lp <- beta[1] + beta[2]*log(dummyArea) myProbability <- exp(lp)/(1 + exp(lp)) lines(myProbability ~ dummyArea,lwd=2,col="blue") legend(.08,.055,legend="Estimated Probability of Extinction",lty=1,lwd=2,col="blue") detach(case2101)str(case2101) attach(case2101) ## EXPLORATION AND MODEL BUILDING proportionExtinct <- Extinct/AtRisk oddsExtinct <- proportionExtinct/(1 - proportionExtinct) logitExtinct <- log(oddsExtinct) # Logit = Log Odds logArea <- log(Area) plot(logitExtinct ~ logArea) binResponse <- cbind(Extinct,AtRisk-Extinct) myGlm1 <- glm(binResponse ~ logArea, family=binomial) summary(myGlm1) logArea2 <- logArea^2 myGlm2 <- update(myGlm1, ~ . + logArea2) summary(myGlm2) # p-value for quadratic term: 0.77 ## INFERENCE AND INTERPRETATION myGlm3 <- update(myGlm1, ~ . - logArea) anova(myGlm3, myGlm1) # Drop in deviance statistic = 33.277 on 1 d.f. 1 - pchisq(33.277,1) # p-value = 7.992234e-09 beta <- myGlm1$coef 1 - 2^beta[2] # 0.1861153 1 - 2^confint(myGlm1,2) #0.2462041 0.1247743 # Interpretation: Associated with each doubling of island area is a 19% # reduction in the odds of extinction (95% confidence interval: 12% to 25% # reduction). ## GRAPHICAL DISPLAY FOR PRESENTATION plot(oddsExtinct ~ Area, log="xy", ylab="Observed Odds of Extinction; log scale", xlab=expression(paste("Island Area (km"^"2","); log scale")), main="Extinctions of Bird Species in the Krunnit Island Archipelago", pch=21, lwd=2, bg="green", cex=2) # Plot odds of extinction vs island area dummyArea <- seq(min(Area),max(Area),length=50) lp <- beta[1] + beta[2]*log(dummyArea) odds <- exp(lp) lines(odds ~ dummyArea,lwd=2) plot(proportionExtinct ~ Area, log="xy", ylab="Proportions of 1949 Species not Found in 1959", xlab=expression(paste("Island Area (km"^"2","); log scale")), main="Proportions of 1949 Bird Species Extinct in 1959 on 18 Krunnit Archipelago Islands", pch=21, lwd=2, bg="green", cex=2) # Plot probability of extinction vs area dummyArea <- seq(min(Area),max(Area),length=50) lp <- beta[1] + beta[2]*log(dummyArea) myProbability <- exp(lp)/(1 + exp(lp)) lines(myProbability ~ dummyArea,lwd=2,col="blue") legend(.08,.055,legend="Estimated Probability of Extinction",lty=1,lwd=2,col="blue") detach(case2101)
This data was collected by J.A. Bishop. Bishop selected seven locations progressively farther from Liverpool. At each location, Bishop chose eight trees at random. Equal number of dead (frozen) light (Typicals) and dark (Carbonaria) moths were glued to the trunks in lifelike positions. After 24 hours, a count was taken of the numbers of each morph that had been removed—presumably by predators.
case2102case2102
A data frame with 14 observations on the following 4 variables.
Morph, a factor with levels "light" and "dark"
Distance from Liverpool (in km)
Number of moths placed
Number of moths removed
Population geneticists consider clines particularly favourable situations for investigating evolutionary phenomena. A cline is a region where two colour morphs of one species arrange themselves at opposite ends of an environmental gradient, with increasing mixtures occurring between. Such a cline exists near Liverpool, England, where a dark morph of a local moth has flourished in response to the blackening of tree trunks by air pollution from the mills. The moths are nocturnal, resting during the day on tree trunks, where their coloration acts as camouflage against predatory birds. In Liverpool, where tree trunks are blackened by smoke, a high percentage of the moths are of the dark morph. One encounters a higher percentage of the typical (pepper–and–salt) morph as one travels from the city into the Welsh countryside, where tree trunks are lighter. J.A. Bishop used this cline to study the intensity of natural selection.
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Bishop, J.A. (1972). An Experimental Study of the Cline of Industrial Melanism in Biston betularia [Lepidoptera] Between Urban Liverpool and Rural North Wales, Journal of Animal Ecology 41: 209–243.
str(case2102) attach(case2102) ## EXPLORATION AND MODEL BUILDING proportionRemoved <- Removed/Placed myPointCode <- ifelse(Morph=="dark",21,22) myPointColor <- ifelse(Morph=="dark","blue","gray") plot(proportionRemoved ~ Distance, pch=myPointCode, bg=myPointColor, cex=2, lwd=2) binResponse <- cbind(Removed, Placed-Removed) Morph <- factor(Morph, levels=c("light","dark")) # Make "light" the ref level myGlm1 <- glm(binResponse ~ Distance + Morph + Distance:Morph, family=binomial) summary(myGlm1) # Residual deviance: 13.230 on 10 degrees of freedom 1 - pchisq(13.230,10) # No evidence of overdispersion myGlm2 <- update(myGlm1, ~ . - Distance:Morph) anova(myGlm2, myGlm1) # Drop in deviance statistic = 11.931 on 1 d.f. 1 - pchisq(11.931,1) # p-value = 0.0005520753 => strong evidence of interaction # It appears that the intercepts are the same for both light and dark morphs, # that there is no effect of Distance for light morphs, but there is an effect # of Distance for dark morphs. ## INFERENCE AND INTERPREATION myTerm <- Distance*ifelse(Morph=="dark",1,0) # Create indicator var for "dark" myGlm3 <- glm(binResponse ~ myTerm, family=binomial) summary(myGlm3) ## GRAPHICAL DISPLAY FOR PRESENTATION myPointCode <- ifelse(Morph=="dark",22,24) myPointColor <- ifelse(Morph=="dark","blue","orange") plot(proportionRemoved ~ Distance, ylab="Proportion of Moths Taken", main="Proportions of Moths Taken by Predators at Seven Locations", xlab="Distance from Liverpool (km)", pch=myPointCode, bg=myPointColor, cex=2, lwd=2) beta <- myGlm3$coef dummyDist <- seq(0,55,length=50) lp <- beta[1] + beta[2]*dummyDist propDark <- exp(lp)/(1 + exp(lp)) lines(propDark ~ dummyDist,lwd=2,col="blue") propLight <- rep(exp(beta[1])/(1 + exp(beta[1])),length(dummyDist)) lines(propLight ~ dummyDist,lwd=2,col="orange") legend(0,0.47,legend=c("Dark Morph","Light Morph"), pch=c(22,24),pt.bg=c("blue","orange"),pt.cex=c(2,2),pt.lwd=c(2,2)) detach(case2102)str(case2102) attach(case2102) ## EXPLORATION AND MODEL BUILDING proportionRemoved <- Removed/Placed myPointCode <- ifelse(Morph=="dark",21,22) myPointColor <- ifelse(Morph=="dark","blue","gray") plot(proportionRemoved ~ Distance, pch=myPointCode, bg=myPointColor, cex=2, lwd=2) binResponse <- cbind(Removed, Placed-Removed) Morph <- factor(Morph, levels=c("light","dark")) # Make "light" the ref level myGlm1 <- glm(binResponse ~ Distance + Morph + Distance:Morph, family=binomial) summary(myGlm1) # Residual deviance: 13.230 on 10 degrees of freedom 1 - pchisq(13.230,10) # No evidence of overdispersion myGlm2 <- update(myGlm1, ~ . - Distance:Morph) anova(myGlm2, myGlm1) # Drop in deviance statistic = 11.931 on 1 d.f. 1 - pchisq(11.931,1) # p-value = 0.0005520753 => strong evidence of interaction # It appears that the intercepts are the same for both light and dark morphs, # that there is no effect of Distance for light morphs, but there is an effect # of Distance for dark morphs. ## INFERENCE AND INTERPREATION myTerm <- Distance*ifelse(Morph=="dark",1,0) # Create indicator var for "dark" myGlm3 <- glm(binResponse ~ myTerm, family=binomial) summary(myGlm3) ## GRAPHICAL DISPLAY FOR PRESENTATION myPointCode <- ifelse(Morph=="dark",22,24) myPointColor <- ifelse(Morph=="dark","blue","orange") plot(proportionRemoved ~ Distance, ylab="Proportion of Moths Taken", main="Proportions of Moths Taken by Predators at Seven Locations", xlab="Distance from Liverpool (km)", pch=myPointCode, bg=myPointColor, cex=2, lwd=2) beta <- myGlm3$coef dummyDist <- seq(0,55,length=50) lp <- beta[1] + beta[2]*dummyDist propDark <- exp(lp)/(1 + exp(lp)) lines(propDark ~ dummyDist,lwd=2,col="blue") propLight <- rep(exp(beta[1])/(1 + exp(beta[1])),length(dummyDist)) lines(propLight ~ dummyDist,lwd=2,col="orange") legend(0,0.47,legend=c("Dark Morph","Light Morph"), pch=c(22,24),pt.bg=c("blue","orange"),pt.cex=c(2,2),pt.lwd=c(2,2)) detach(case2102)
Although male elephants are capable of reproducing by 14 to 17 years of age, your adult males are usually unsuccessful in competing with their larger elders for the attention of receptive females. Since male elephants continue to grow throughout their lifetimes, and since larger males tend to be more successful at mating, the males most likely to pass their genes to future generations are those whose characteristics enable them to live long lives. Joyce Poole studied a population of African elephants in Amboseli National Park, Kenya, for 8 years. This data frame contains the number of successful matings and ages (at the study's beginning) of 41 male elephants.
case2201case2201
A data frame with 41 observations on the following 2 variables.
Age of elephant at beginning of study
Number of successful matings
Ramsey, F.L. and Schafer, D.W. (2002). The Statistical Sleuth: A Course in Methods of Data Analysis (2nd ed), Duxbury.
Poole, J.H. (1989). Mate Guarding, Reproductive Success and Female Choice in African Elephants, Animal Behavior 37: 842–849.
str(case2201) attach(case2201) ## EXPLORATION AND MODEL BUILDING plot(Matings ~ Age, log="y") ageSquared <- Age^2 myGlm1 <- glm(Matings ~ Age + ageSquared, family=poisson) summary(myGlm1) # No evidence of a need for ageSquared ## INFERENCE AND INTERPRETATION myGlm2 <- update(myGlm1, ~ . - ageSquared) summary(myGlm2) beta <- myGlm2$coef exp(beta[2]) #1.071107 exp(confint(myGlm2,2)) #1.042558 1.100360 # Interpretation: Associated with each 1 year increase in age is a 7% increase # in the mean number of matings (95% confidence interval 4% to 10% increase). ## GRAPHICAL DISPLAY FOR PRESENTATION plot(Matings ~ Age, ylab="Number of Successful Matings", xlab="Age of Male Elephant (Years)", main="Age and Number of Successful Matings for 41 African Elephants", pch=21, bg="green", cex=2, lwd=2) dummyAge <- seq(min(Age),max(Age), length=50) lp <- beta[1] + beta[2]*dummyAge curve <- exp(lp) lines(curve ~ dummyAge,lwd=2) detach(case2201)str(case2201) attach(case2201) ## EXPLORATION AND MODEL BUILDING plot(Matings ~ Age, log="y") ageSquared <- Age^2 myGlm1 <- glm(Matings ~ Age + ageSquared, family=poisson) summary(myGlm1) # No evidence of a need for ageSquared ## INFERENCE AND INTERPRETATION myGlm2 <- update(myGlm1, ~ . - ageSquared) summary(myGlm2) beta <- myGlm2$coef exp(beta[2]) #1.071107 exp(confint(myGlm2,2)) #1.042558 1.100360 # Interpretation: Associated with each 1 year increase in age is a 7% increase # in the mean number of matings (95% confidence interval 4% to 10% increase). ## GRAPHICAL DISPLAY FOR PRESENTATION plot(Matings ~ Age, ylab="Number of Successful Matings", xlab="Age of Male Elephant (Years)", main="Age and Number of Successful Matings for 41 African Elephants", pch=21, bg="green", cex=2, lwd=2) dummyAge <- seq(min(Age),max(Age), length=50) lp <- beta[1] + beta[2]*dummyAge curve <- exp(lp) lines(curve ~ dummyAge,lwd=2) detach(case2201)
The Del Norte Salamander (plethodon elongates) is a small (5–7 cm) salamander found among rock rubble, rock outcrops and moss-covered talus in a narrow range of northwest California. To study the habitat characteristics of the species and particularly the tendency of these salamanders to reside in dwindling old-growth forests, researchers selected 47 sites from plausible salamander habitat in national forest and parkland. Randomly chosen grid points were searched for the presence of a site with suitable rocky habitat. At each suitable site, a 7 metre by 7 metre search are was examined for the number of salamanders it contained. This data frame contains the counts of salamanders at the sites, along with the percentage of forest canopy and age of the forest in years.
case2202case2202
A data frame with 47 observations on the following 4 variables.
Investigated site
Number of salamanders found in 49 m area
Percentage of canopy cover
Forest age
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Welsh, H.H. and Lind, A.J. (1995). Journal of Herpetology 29(2): 198–210.
str(case2202) attach(case2202) ## EXPLORATION AND MODEL BUILDING logSalamanders <- log(Salamanders + .5) logForestAge <- log(ForestAge + .5) myMatrix <- cbind(PctCover,logForestAge,logSalamanders) if (require(car)) { # Use car library scatterplotMatrix(myMatrix, diagonal="histogram", reg.line=FALSE, spread=FALSE) } myGlm1 <- glm(Salamanders ~ PctCover + logForestAge + PctCover:logForestAge, family=poisson) summary(myGlm1) # Backward elimination... myGlm2 <- update(myGlm1, ~ . - PctCover:logForestAge) summary(myGlm2) myGlm3 <- update(myGlm2, ~ . - logForestAge) summary(myGlm3) # PctCover is the only explanatory variable remaining plot(Salamanders ~ PctCover) # It appears that there are 2 distributions # of Salamander counts; one for PctCover < 70 and one for PctCover > 70 # See if PctCover is associated Salamanders in each subset myGlm4 <- glm(Salamanders ~ PctCover, family=poisson,subset=(PctCover > 70)) summary(myGlm4) # No evidence of an effect for this subset myGlm5 <- glm(Salamanders ~ PctCover, family=poisson,subset=(PctCover < 70)) summary(myGlm5) # No evidence on this subset either ## INFERENCE (2 means) Group <- ifelse(PctCover > 70,"High","Low") Group <- factor(Group, levels=c("Low","High") ) # Make "Low Cover" the ref group myGlm6 <- glm(Salamanders ~ Group, family=poisson) summary(myGlm6) ## GRAPHICAL DISPLAY FOR PRESENTATION plot(Salamanders ~ PctCover, ylab="Number of Salamanders", xlab="Percentage of Canopy Covered", main="Number of Salamanders versus Percent Canopy Cover", pch=21,bg="green", cex=2, lwd=2) beta <- myGlm6$coef lines(c(0,55),exp(c(beta[1],beta[1])),lwd=2) text(56,exp(beta[1]),paste("mean= ",round(exp(beta[1]),1)),adj=0) lines(c(76,93),exp(c(beta[1]+beta[2],beta[1]+beta[2])),lwd=2) text(56,exp(beta[1]+beta[2]),paste("mean=",round((beta[1]+beta[2]),1)),adj=-1) detach(case2202)str(case2202) attach(case2202) ## EXPLORATION AND MODEL BUILDING logSalamanders <- log(Salamanders + .5) logForestAge <- log(ForestAge + .5) myMatrix <- cbind(PctCover,logForestAge,logSalamanders) if (require(car)) { # Use car library scatterplotMatrix(myMatrix, diagonal="histogram", reg.line=FALSE, spread=FALSE) } myGlm1 <- glm(Salamanders ~ PctCover + logForestAge + PctCover:logForestAge, family=poisson) summary(myGlm1) # Backward elimination... myGlm2 <- update(myGlm1, ~ . - PctCover:logForestAge) summary(myGlm2) myGlm3 <- update(myGlm2, ~ . - logForestAge) summary(myGlm3) # PctCover is the only explanatory variable remaining plot(Salamanders ~ PctCover) # It appears that there are 2 distributions # of Salamander counts; one for PctCover < 70 and one for PctCover > 70 # See if PctCover is associated Salamanders in each subset myGlm4 <- glm(Salamanders ~ PctCover, family=poisson,subset=(PctCover > 70)) summary(myGlm4) # No evidence of an effect for this subset myGlm5 <- glm(Salamanders ~ PctCover, family=poisson,subset=(PctCover < 70)) summary(myGlm5) # No evidence on this subset either ## INFERENCE (2 means) Group <- ifelse(PctCover > 70,"High","Low") Group <- factor(Group, levels=c("Low","High") ) # Make "Low Cover" the ref group myGlm6 <- glm(Salamanders ~ Group, family=poisson) summary(myGlm6) ## GRAPHICAL DISPLAY FOR PRESENTATION plot(Salamanders ~ PctCover, ylab="Number of Salamanders", xlab="Percentage of Canopy Covered", main="Number of Salamanders versus Percent Canopy Cover", pch=21,bg="green", cex=2, lwd=2) beta <- myGlm6$coef lines(c(0,55),exp(c(beta[1],beta[1])),lwd=2) text(56,exp(beta[1]),paste("mean= ",round(exp(beta[1]),1)),adj=0) lines(c(76,93),exp(c(beta[1]+beta[2],beta[1]+beta[2])),lwd=2) text(56,exp(beta[1]+beta[2]),paste("mean=",round((beta[1]+beta[2]),1)),adj=-1) detach(case2202)
Researchers used 7 red and 7 black playing cards to randomly assign 14 volunteer males with high blood pressure to one of two diets for four weeks: a fish oil diet and a standard oil diet. These data are the reductions in diastolic blood pressure.
ex0112ex0112
A data frame with 14 observations on the following 2 variables.
reduction in diastolic blood pressure (in mm of mercury)
factor variable indicating the diet that the subject
followed, with levels "FishOil" and "RegularOil"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Knapp, H.R. and FitzGerald, G.A. (1989). The Antihypertensive Effects of Fish Oil, New England Journal of Medicine 320: 1037–1043.
str(ex0112)str(ex0112)
The data are the gross domestic product per capita for 228 countries in 2010.
ex0116ex0116
A data frame with 228 observations on the following 3 variables.
rank order of country from highest to lowest GDP
name of country
per capita GDP in US
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Central Intelligence Agency, Country Comparison: GDP per capita
(PPP), The World Factbook,
https://www.cia.gov/library/publications/theworld-factbook/rankorder/2004rank.html
(retrieved June 30,2011).
str(ex0116)str(ex0116)
The data are the zinc concentrations (in mg/ml) in the blood of rats that received a dietary supplement and rats that did not receive the supplement.
ex0125ex0125
A data frame with 39 observations on the following 2 variables.
a factor representing the group, with levels "A"
for the dietary supplement group and "B" for the control group
measured zinc concentration in mg/ml
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex0125)str(ex0125)
The data are the number of pro- and anti-environmental votes, according to the League of Conservation Voters, for each member of the U.S. House of Representatives in 2005, 2006, or 2007.
ex0126ex0126
A data frame with 492 observations on the following 10 variables.
the state that the member represented
name of the representative
a factor representing political party, with levels
"R" for Republican, "D" for Democratic, and
"I" for Independent
the number of pro-environmental votes in 2005
the number of anti-environmental votes in 2005
the number of pro-environmental votes in 2006
the number of anti-environmental votes in 2006
the number of pro-environmental votes in 2007
the number of anti-environmental votes in 2007
the total percentage of a representative's votes between 2005 and 2007 that were deemd to be pro-environmental
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex0126)str(ex0126)
The data are the number of pro- and anti-environmental votes, according to the League of Conservation Voters, for each member of the U.S. Senate in 2005, 2006, or 2007.
ex0127ex0127
A data frame with 112 observations on the following 10 variables.
the state that the member represented
name of the senator
a factor representing political party, with levels "R" for Republican, "D" for Democratic, and "I" for Independent
the number of pro-environmental votes in 2005
the number of anti-environmental votes in 2005
the number of pro-environmental votes in 2006
the number of anti-environmental votes in 2006
the number of pro-environmental votes in 2007
the number of anti-environmental votes in 2007
the total percentage of a representative's votes between 2005 and 2007 that were deemd to be pro-environmental
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex0127)str(ex0127)
The data are survival times (in days) of guinea pigs that were randomly assigned either to a control group or to a treatment group that received a dose of tubercle bacilli.
ex0211ex0211
A data frame with 122 observations on the following 2 variables.
survival time of guinea pig (in days)
a factor with levels "Bacilli" and
"Control", indicating the group to which the guinea pig was
assigned
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Doksum, K. (1974). Empirical Probability Plots and Statistical Inference for Nonlinear Models in the Two–sample Case, Annals of Statistics 2: 267–277.
str(ex0211)str(ex0211)
In the 1980s, biologists Peter and Rosemary Grant caught and measured all the birds from more than 20 generations of finches on the Galapagos island of Daphne Major. In one of those years, 1977, a severe drought caused vegetation to wither, and the only remaining food source was a large, tough seed, which the finches ordinarily ignored. Were the birds with larger and stronger beaks for opening these tough seeds more likely to survive that year, and did they tend to pass this characteristic to their offspring? The data are beak depths (height of the beak at its base) of 751 finches caught the year before the drought (1976) and 89 finches captured the year after the drought (1978).
ex0218ex0218
A data frame with 840 observations on the following 2 variables.
Year the finch was caught, 1976 or 1978
Beak depth of the finch (mm)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Grant, P. (1986). Ecology and Evolution of Darwin's Finches, Princeton University Press, Princeton, N.J.
str(ex0218)str(ex0218)
As evidence in support of natural selection, Bumpus presented measurements on house sparrows brought to the Anatomical Laboratory of Brown University after an uncommonly severe winter storm. Some of these birds had survived and some had perished. Bumpus asked whether those that perished did so because they lacked physical characteristics enabling them to withstand the intensity of that particular instance of selective elimination. The data are on the the weights, in grams, for the 24 adult male sparrows that perished and for the 35 adult males that survived.
ex0221ex0221
A data frame with 59 observations on the following 2 variables.
humerus length of adult male sparrows (inches)
factor variable indicating whether the sparrow
perished or survived in a winter storm, with levels Perished
and Survived
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex0221)str(ex0221)
These data are armed Forces Qualifying Test (AFQT) score percentiles and component test scores in arithmetic reasoning, word knowledge, paragraph comprehension, and mathematical knowledge for a sample of 1,278 U.S. women and 1,306 U.S. men in 1981.
ex0222ex0222
A data frame with 2,584 observations on the following 6 variables.
a factor with levels "female" and "male"
score on the arithmetic reasoning component of the AFQT test
score on the word knowledge component
score on the paragraph comprehension component
score on the mathematical knowledge component
percentile score on the AFQT test
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
National Longitudinal Survey of Youth, U.S. Bureau of Labor Statistics, https://www.bls.gov/nls/home.htm (May 8, 2008).
ex0330, ex0331, ex0524,
ex0525, ex0828, ex0923,
ex1033, ex1223
str(ex0222)str(ex0222)
The National Highway System Designation Act was signed into law in the United States on November 28, 1995. Among other things, the act abolished the federal mandate of 55 mile per hour maximum speed limits on roads in the United States and permitted states to establish their own limits. Of the 50 states (plus the District of Columbia), 32 increased their speed limits at the beginning of 1996 or sometime during 1996. These data are the percentage changes in interstate highway traffic fatalities from 1995 to 1996.
ex0223ex0223
A data frame with 51 observations on the following 5 variables.
US state
number of traffic fatalities in 1995
number of traffic fatalities in 1996
percentage change in interstate traffic fatalities between 1995 and 1996
a factor with levels "Inc" and "Ret",
indicating whether the state increased or retained its speed limit
Ramsey, F.L. and Schafer, D.W. (2002). The Statistical Sleuth: A Course in Methods of Data Analysis (2nd ed), Duxbury.
Report to Congress: The Effect of Increased Speed Limits in the Post-NMSL Era, National Highway Traffic Safety Administration, February, 1998; available in the reports library at https://www-fars.nhtsa.dot.gov/.
str(ex0223)str(ex0223)
Researchers collected historical and current data on umpires to investigate their life expectancies following the collapse and death of a U.S. major league baseball umpire. They were investigating speculation that stress associated with the job posed a health risk. Data were found on 227 umpires who had died or had retired and were still living. The data set includes the dates of birth and death.
ex0321ex0321
A data frame with 227 observations on the following 3 variables.
observed lifetime for those umpires who had died by the time of the study or current age of those still living
0 for those who had died by the time of the study or 1 for those who were still living
length from actuarial life tables for individuals who were alive at the time the person first became an umpire
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Cohen, R.S., Kamps, C.A., Kokoska, S., Segal E.M. and Tucker, J.B.(2000). Life Expectancy of Major League Baseball Umpires, The Physician and Sportsmedicine 28(5): 83–89.
str(ex0321)str(ex0321)
Data contains yearly skin cancer rates (per 100,000 people) in Connecticut from 1938 to 1972 with a code indicating those years that came two years after higher than average sunspot activity and those years that came two years after lower than average sunspot activity.
ex0323ex0323
A data frame with 35 observations on the following 3 variables.
year
skin cancer rate per 100,000 people
a factor with levels "High" and
"Low"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Andrews, D.F. and Herzberg, A.M. (1985). Data: A Collection of Problems from many Fields for the Student and Research Worker, Springer-Verlag.
str(ex0323)str(ex0323)
As part of a study to investigate reproductive strategies in plants, biologists recorded the time spent at sources of pollen and the proportions of pollen removed by bumblebee queens and honeybee workers pollinating a species of lily.
ex0327ex0327
A data frame with 47 observations on the following 3 variables.
proportion of pollen removed
duration of visit (in seconds)
factor variable with levels "Queen" and
"Worker"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Harder, L.D. and Thompson, J.D. (1989). Evolutionary Options for Maximizing Pollen Dispersal of Animal-pollinated Plants, American Naturalist 133: 323–344.
str(ex0327)str(ex0327)
The data are incomes in U.S. dollars for 1,020 working Americans who had 12 years of education and 406 working Americans who had 16 years of education, in 2005.
ex0330ex0330
A data frame with 1,426 observations on the following 3 variables.
a subject identification number
number of years of education–either 12 or 16
income, in dollars
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
National Longitudinal Survey of Youth, U.S. Bureau of Labor Statistics, https://www.bls.gov/nls/home.htm (May 8, 2008).
ex0222, ex0331, ex0524,
ex0525, ex0828, ex0923,
ex1033, ex1223
str(ex0330)str(ex0330)
The data are incomes in U.S. dollars for 406 working Americans who had 16 years of education and 374 working Americans who had more than 16 years of education, in 2005.
ex0331ex0331
A data frame with 780 observations on the following 3 variables.
a subject identification number
factor with levels "16" and ">16"
income, in dollars
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
National Longitudinal Survey of Youth, U.S. Bureau of Labor Statistics, https://www.bls.gov/nls/home.htm (May 8, 2008).
ex0222, ex0330, ex0524,
ex0525, ex0828, ex0923,
ex1033, ex1223
str(ex0331)str(ex0331)
In-state and out-of-state tuition in dollars for random samples of 25 private and 25 public U.S. colleges and universities in 2011-2012.
ex0332ex0332
A data frame with 50 observations on the following 4 variables.
name of the college
a factor with levels "Private" and "Public"
in-state tuition in dollars
out-of-state tuition in dollars
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
College Board: http://www.collegeboard.com/student/ (11 July 2011)
str(ex0332)str(ex0332)
Relative brain weights for 51 species of mammal whose average litter size is less than 2 and for 45 species of mammal whose average litter size is greater than or equal to 2.
ex0333ex0333
A data frame with 96 observations on the following 2 variables.
relative brain sizes (1000 * Brain weight/Body weight) for 96 species of mammals
factor variable with levels "Small"
and "Large"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Sacher, G.A. and Staffeldt, E.F. (1974). Relation of Gestation Time to Brain Weight for Placental Mammals: Implications for the Theory of Vertebrate Growth, American Naturalist 108: 593–613.
str(ex0333)str(ex0333)
Plant heights (inches) for 15 pairs of plants of the same age, one of which was grown from a seed from a cross-fertilized flower and the other of which was grown from a seed from a self-fertilized flower.
ex0428ex0428
A data frame with 15 observations on the following 2 variables.
height (inches) of cross-fertilized plant
height (inches) of self-fertilized plant
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Andrews, D.F. and Herzberg, A.M. (1985). Data: A Collection of Problems from many Fields for the Student and Research Worker, Springer-Verlag.
str(ex0428)str(ex0428)
The data are the number of tree seedlings per transect in nine logged (L) and seven unlogged (U) plots affected by the Oregon Biscuit Fire, in 2004 and 2005, and the percentage of seedlings lost between 2004 and 2005. The goal is to see whether the distribution of seedlings lost differs in logged and unlogged plots.
ex0429ex0429
A data frame with 16 observations on the following 5 variables.
an identification code for plot
a factor with levels "L" for logged and "U"
for unlogged
the number of seedlings in the plot in 2004
the number of seedlings in the plot in 2005
the percentage of 2004 seedlngs that were lost
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Donato, D.C., Fontaine, J.B., Campbell, J.L., Robinson, W.D., Kauffman, J.B., and Law, B.E. (2006). Post-Wildfire Logging Hinders Regeneration and Increases Fire Risk, Science 311: 352.
str(ex0429)str(ex0429)
Tolerance to sunlight (in minutes) for 13 patients prior to and after treatment with a sunscreen.
ex0430ex0430
A data frame with 13 observations on the following 2 variables.
tolerance to sunlight (minutes) prior to sunscreen application
tolerance to sunlight (minutes) after sunscreen application
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Fusaro, R.M. and Johnson, J.A. (1974). Sunlight Protection for Erythropoietic Protoporphyria Patients, Journal of the American Medical Association 229(11): 1420.
str(ex0430)str(ex0430)
Researchers randomly assigned metastatic breast cancer patients to either a control group or a group that received weekly 90 minute sessions of group therapy and self-hypnosis, to see whether the latter treatment improved the patients' quality of life.
ex0431ex0431
A data frame with 58 observations on the following 3 variables.
months of survival after beginning of study
a factor with levels "Control" and
"Therapy"
0 if entire lifetime observed, 1 if patient known to have lived at least 122 months
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Spiegel, D., Bloom, J.R., Kraemer, H.C. and Gottheil, E. (1989). Effect of Psychosocial Treatment on Survival of Patients with Metastatic Breast Cancer, Lancet 334(8668): 888–891.
str(ex0431)str(ex0431)
To investigate the capacity of marijuana to reduce the side effects of cancer chemotherapy, researchers performed a double-blind, randomized, crossover trial. Fifteen cancer patients on chemotherapy were randomly assigned to receive either a marijuana treatment or a placebo treatment after their first three sessions of chemotherapy. They were then crossed over to the opposite treatment for their next 3 sessions.
ex0432ex0432
A data frame with 15 observations on the following 3 variables.
subject number 1–15
total number of vomiting and retching episodes under marijuana treatment
total number of vomiting and retching episodes under placebo treatment
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Chang, A.E., Shiling, D.J., Stillman, R.C., Goldberg, N.H., Seipp, C.A., Barofsky, I., Simon, R.M. and Rosenberg, S.A. (1979). Delta-9-Tetrahydrocannabinol as an Antiemetic in Cancer Patients Receiving High Dose Methotrexate, Annals of Internal Medicine 91(6): 819–824.
str(ex0432)str(ex0432)
A randomized experiment was performed to estimate the effect of a certain fatty acid CPFA on the level of a certain protein in rat livers.
ex0518ex0518
A data frame with 30 observations on the following 4 variables.
levels of protein (x 10) found in rat livers
a factor with levels "Control",
"CPFA50", "CPFA150", "CPFA300",
"CPFA450" and "CPFA600"
a factor with levels "Day1", "Day2",
"Day3", "Day4" and "Day5"
a factor with levels "Group1",
"Group2", ..., "Group10"; the observed levels of the
Treatment and Day interaction
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex0518)str(ex0518)
Data frame with measurements of oxygen isotopic composition of vertebrate bone phosphate (per mil deviations from SMOW) in 12 bones of a singe Tyrannosaurus rex specimen
ex0523ex0523
A data frame with 52 observations on the following 2 variables.
oxygen isotopic composition
a factor with levels "Bone1",
"Bone2", ..., "Bone12"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Barrick, R.E. and Showers, W.J. (1994). Thermophysiology of Tyrannosaurus rex: Evidence from Oxygen Isotopes, Science 265(5169): 222–224.
str(ex0523)str(ex0523)
These data are annual incomes in 2005 for 2,584 Americans who were selected in the National Longitudinal Study of Youth 1979, who were available for re- interview in 2006, and who had paying jobs in 2005, along with the quartile of their AFQT (IQ) test score taken in 1981. How strong is the evidence that the distributions of 2005 annual incomes differ in the four populations? By how many dollars or by what percent does the distribution of 2005 incomes for those within the highest (fourth) quartile of IQ test scores exceed the distribution for the lowest (first) quartile?
ex0524ex0524
A data frame with 2,584 observations on the following 3 variables.
subject identification number
a factor with levels "1stQuartile",
"2ndQuartile", "3rdQuartile" and "4thQuartile"
annual income in U.S. dollars, 2005
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
National Longitudinal Survey of Youth, U.S. Bureau of Labor Statistics, https://www.bls.gov/nls/home.htm (May 8, 2008).
ex0222, ex0330, ex0331,
ex0525, ex0828, ex0923,
ex1033, ex1223
str(ex0524)str(ex0524)
These data are annual incomes in 2005 of a random sample of 2,584 Americans who were selected for the National Longitudinal Survey of Youth in 1979 and who had paying jobs in 2005. The data set also includes a code for the number of years of education that each individual had completed by 2006: <12, 12, 13–15, 16, and >16. How strong is the evidence that at least one of the five population distributions (corresponding to the different years of education) is different from the others? By how many dollars or by what percent does the mean or median for each of the last four categories exceed that of the next lowest category?
ex0525ex0525
A data frame with 2,584 observations on the following 3 variables.
subject identification number
a factor for years of education category with levels
"<12", "12", "13-15", "16" and ">16"
Annual income in 2005, in U.S. dollars
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
National Longitudinal Survey of Youth, U.S. Bureau of Labor Statistics, https://www.bls.gov/nls/home.htm (May 8, 2008).
ex0222, ex0330, ex0331,
ex0524, ex0828, ex0923,
ex1033, ex1223
str(ex0525)str(ex0525)
These data are simulated to match the summary and conclusions of a real study of overweight employees who were randomly assigned to three diet groups: a low-fat diet, a low-carb diet (similar to the Atkins diet), and a Mediterranean diet. The study ran for two years, with 272 employees completing the entire protocol. Is there evidence of differences in average weight loss among these diets? If so, which diets appear to be better than which others?
ex0623ex0623
A data frame with 272 observations on the following 3 variables.
subject identification number
a factor with levels "Low-Carbohydrate",
"Low-Fat", and "Meditrranean"
the amount of weight lost in 24 months on the diet, in kg
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex0623)str(ex0623)
Is there a physiological basis for sexual preference? Researchers measured the volumes of four cell groups in the interstitial nuclei of the anterior hypothalamus in postmortem tissue from 41 subjects at autopsy from seven metropolitan hospitals in New York and California.
ex0624ex0624
A data frame with 41 observations on the following 5 variables.
volumes of INAH3 (1000
mm) cell clusters from 41 humans
a factor with levels
"Group1" |
heterosexual male with AIDS death |
"Group2" |
heterosexual male with Non-AIDS death |
"Group3" |
homosexual male with AIDS death |
"Group4" |
heterosexual female with AIDS death |
"Group5" |
heterosexual female with Non-AIDS death |
a factor with levels "Female" and "Male"
a factor with levels "Heterosexual" and
"Homosexual"
a factor with levels "AIDS" and "Non-AIDS"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
LeVay, S. (1991). A Difference in Hypothalamic Structure Between Heterosexual and Homosexual Men, Science 253(5023): 1034–1037.
str(ex0624)str(ex0624)
The first three columns are the names, orders of distance from the sun and distances from the sun (scaled so that earth is 1) of the 8 planets in our solar system and the dwarf planet, Pluto. The next three columns are the same, but also include the asteroid belt.
ex0721ex0721
A data frame with observations on the following 6 variables.
name of object in solar system, 9 objects
order of object's distance from the sun
distance of object from sun, with earth = 1
name of object in solar system, including asteroid belt
order of object's distance from the sun
distance of object from sun, with earth = 1
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex0721)str(ex0721)
As part of a study of the effects of predatory intertidal crab species on snail populations, researchers measured the mean closing forces and the propdus heights of the claws on several crabs of three species.
ex0722ex0722
A data frame with 38 observations on the following 3 variables.
closing strength of claw of the crab
propodus height of claw of the crab
species to which the crab belongs
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Yamada, S.B. and Boulding, E.G. (1992). Claw Morphology, Prey Size Selection and Foraging Efficiency in Generalist and Specialist Shell-Breaking Crabs. Journal of Experimental Marine Biology and Ecolog, 220 191–211.
str(ex0722)str(ex0722)
The data are on the proportion of male birts in Denmark, The Netherlands, Canada and the United States for a number of yeras. Notice that the proportions for Canada and the United States are only provided for the years 1970 to 1990, while Denmark and The Netherlands have data listed for 1950 to 1994.
ex0724ex0724
A data frame with 45 observations on the following 5 variables.
year of observation
male birth rate of Denmark for given year
male birth rate of The Netherlands for given year
male birth rate of Canada for given year
male birth rate of the United States for given year
Ramsey, F.L. and Schafer, D.W. (2002). The Statistical Sleuth: A Course in Methods of Data Analysis (2nd ed), Duxbury.
Davis, D.L., Gottlieb, M.B. and Stampnitzky, J.R. (1998). Reduced ratio of male to female births in several industrial countries, Journal of the American Medical Association 279(13): 1018–1023.
str(ex0724)str(ex0724)
These data are measured distances and recession velocities for 10
clusters of nebulae, much farther from earth than the nebulae reported
in case0701.
ex0727ex0727
A data frame with 10 observations on the following 2 variables.
distance from earth (in million parsec)
recession velocity (in kilometres per second)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Hubble, E. and Humason, M. (1931). The Velocity–Distance Relation Among Extra–calactic Nebulae, Astrophysics Journal 74: 43–50.
str(ex0725)str(ex0725)
These data are heights of 933 adults and their parents, as measured by Karl Pearson in 1885.
ex0726ex0726
A data frame with 933 observations on the following 5 variables.
a factor with levels "female" and "male"
an identification number for family, 1, 2,..., 205
adult height of the child, inches
height of the child's father, inches
height of the child's mother, inches
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Hubble, E. and Humason, M. (1931). The Velocity–Distance Relation Among Extra–calactic Nebulae, Astrophysics Journal 74: 43–50.
str(ex0725)str(ex0725)
Black wheatears are small birds in Spain and Morocco. Males of the species demonstrate an exaggerated sexual display by carrying many heavy stones to nesting cavities. This 35–gram bird transports, on average, 3.1 kg of stones per nesting season! Different males carry somewhat different sized stones, prompting a study on whether larger stones may be a signal of higher health status. Soler et al. calculated the average stone mass (g) carried by each of 21 male black wheatears, along with T-cell response measurements reflecting their immune systems' strengths.
ex0727ex0727
A data frame with 21 observations on the following 2 variables.
average mass of stones carried by bird (in g)
T-cell response measurement (in mm)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Soler, M., Martín-Vivaldi, M., Marín J. and Møller, A. (1999). Weight lifting and health status in the black wheatears, Behavioral Ecology 10(3): 281–286.
str(ex0727)str(ex0727)
Studies over the past two decades have shown that activity can effect the reorganisation of the human central nervous system. For example, it is known that the part of the brain associated with activity of a finger or limb is taken over for other purposes in individuals whose limb or finger has been lost. In one study, psychologists used magnetic source imaging (MSI) to measure neuronal activity in the brains of nine string players (six violinists, two cellists and one guitarist) and six controls who had never played a musical instrument, when the thumb and fifth finger of the left hand were exposed to mild stimulation. The researchers felt that stringed instrument players, who use the fingers of their left hand extensively, might show different behaviour—as a result of this extensive physical activity—than individuals who did not play stringed instruments.
ex0728ex0728
A data frame with 15 observations on the following 2 variables.
years that the individual has been playing
neuronal activity index
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Elbert, T., Pantev, C., Wienbruch, C., Rockstroh, B. and Taub E. (1995). Increased cortical representation of the fingers of the left hand in string players, Science 270(5234): 305–307.
str(ex0728)str(ex0728)
These data are the number of percentage points by which exit polls over estimated the actual vote for candidate John Kerry in the 2004 U.S. presidential election, grouped according to the distance of the exit poll interviewer from the door of the polling location. How strong is the evidence that the mean Kerry overestimate increases with increasing distance of interviewer from the door (thus lending evidence to the theory that supporters of the other candidate, George W Bush, were more inclined to avoid exit pollsters)?
ex0729ex0729
A data frame with 6 observations on the following 2 variables.
number of percentage points by which the exit poll estimate exceeded the actual percentage voting for Kerry (in all precincts with a similar distance of interviewer from the door
distance of the interviewer from the door of the polling location, in feet
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Evaluation of Edison/Mitofsky Election System 2004 prepared by Edison Media Research and Mitofsky International for the National Election Pool (NEP), January 15, 2005. https://abcnews.go.com/images/Politics/EvaluationofEdisonMitofskyElectionSystem.pdf
str(ex0729)str(ex0729)
These data are the average proportion of voters refusing to be interviewed by exit pollsters in the 2004 U.S. presidential election, grouped gby age of the interviewer, and the approximate age of the interviewer. What evidence do the data provide that the mean refusal rate decreased with incrasing age of interviewer?
ex0730ex0730
A data frame with 6 observations on the following 2 variables.
age of the exit poll interviewer, years
average proportion of voters refusing to be interviewed
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Evaluation of Edison/Mitofsky Election System 2004 prepared by Edison Media Research and Mitofsky International for the National Election Pool (NEP), January 15, 2005. https://abcnews.go.com/images/Politics/EvaluationofEdisonMitofskyElectionSystem.pdf
str(ex0730)str(ex0730)
The data in case0702 are a subset of the complete data on
postmortum pH in 12 steer carcasses.
ex0816ex0816
A data frame with 12 observations on the following 2 variables.
time after slaughter (hours)
pH level in postmortem muscle
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Schwenke, J.R. and Milliken, G.A. (1991). On the Calibration Problem Extended to Nonlinear Models, Biometrics 47(2): 563–574.
str(ex0816)str(ex0816)
In a study of the effectiveness of biological control of the exotic weed tansy ragwort, researchers manipulated the exposure to the ragwort flea beetle on 15 plots that had been planted with a high density of ragwort. Harvesting the plots the next season, they measured the average dry mass of ragwort remaining (grams/plant) and the flea beetle load (beetles/gram of ragwort dry mass) to see if the ragwort plants in plots with high flea beetle loads were smaller as a result of herbivory by the beetles.
ex0817ex0817
A data frame with 15 observations on the following 2 variables.
flee beetle load (in beetles/gram of ragwort dry mass)
dry mass of ragwort weed
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
McEvoy, P. and Cox, C. (1991). Successful Biological Control of Ragwort, Senecio jacobaea, by introducing insects in Oregon, Ecological Applications 1(4): 430–442.
str(ex0817)str(ex0817)
The data are Democratic and Republican vote counts, by (a) absentee ballot and (b) voting machine, for 22 elections in Philadelphia's senatorial districts between 1982 and 1993.
ex0820ex0820
A data frame with 22 observations on the following 2 variables.
Year of election
a factor with levels "D1", "D2",
"D3", "D4", "D5", "D7", and "D8"
Number of absentee ballots indicating a vote for the Democratic candidate
Number of absentee ballots indicating a vote for the Republican candidate
Number of machine-counted ballots indicating a vote for the Democratic candidate
Number of machine-coutned ballots indicating a vote for the Republican candidate
Percentage of absentee ballots indicating a vote for the Democratic candidate
Percentage of machine-counted ballots indicating a vote for the Democratic candidate
a factor taking on the value "yes" for the
disputed election and "no" for all other elections
In a special election to fill a Pennsylvania State Senate seat in 1993, the Democrat, William Stinson, received 19,127 machine–counted votes and the Republican, Bruce Marks, received 19,691. In addition, there were 1,391 absentee ballots for Stinson and 366 absentee ballots for Marks, so that the total tally showed Stinson the winner by 461 votes. The large disparity between the machine–counted and absentee votes, and the resulting reversal of the outcome due to the absentee ballots caused some concern about possible illegal influence on the absentee votes. To see whether the discrepancy in absentee votes was larger than could be explained by chance, an econometrician considered the data given in this data frame (read from a graph in The New York Times, 11 April 1994).
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Ashenfelter, O (1994). Report on Expected Absentee Ballots. Department of Economics, Princeton University. See also Simon Jackman (2011). pscl: Classes and Methods for R Developed in the Political Science Computational Laboratory, Stanford University. Department of Political Science, Stanford University. Stanford, California. R package version 1.03.10. https://github.com/atahk/pscl/
str(ex0820)str(ex0820)
Data are the number of butterfly species in 16 islands of forest of various sizes in otherwise cleared areas in Brazil.
ex0822ex0822
A data frame with 16 observations on the following 2 variables.
area (ha) of forest patch
number of butterfly species
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Lovejoy, T.E., Rankin, J.M., Bierregaard, Jr., R.O., Brown, Jr., K.S., Emmons, L.H. and van der Voort, M. (1984). Ecosystem decay of Amazon forest remnants in Nitecki, M.H. (ed.) Extinctions, University of Chicago Press.
str(ex0822)str(ex0822)
The data are the average wine consumption rates (in liters per person per year) and number of ischemic heart disease deaths (per 1000 men aged 55 to 64 years) for 18 industrialized countries.
data(ex0823)data(ex0823)
A data frame with 18 observations on the following 3 variables.
a character vector indicating the country
consumption of wine (liters per person per year)
heart disease mortality rate (deaths per 1,000)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
St. Leger A.S., Cochrane, A.L. and Moore, F. (1979). Factors Associated with Cardiac Mortality in Developed Countries with Particular Reference to the Consumption of Wine, Lancet: 1017–1020.
str(ex0823)str(ex0823)
A high respiratory rate is a potential diagnostic indicator of respiratory infection in children. To judge whether a respiratory rate is “high” however, a physician must have a clear picture of the distribution of normal rates. To this end, Italian researchers measured the respiratory rates of 618 children between the ages of 15 days and 3 years.
ex0824ex0824
A data frame with 618 observations on the following 2 variables.
age in months of child
respiratory rate (breaths per minute)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Rusconi, F., Castagneto, M., Porta, N., Gagliardi, L., Leo, G., Pellegatta, A., Razon, S. and Braga, M. (1994). Reference Values for Respiratory Rate in the First 3 Years of Life, Pediatrics 94(3): 350–355.
str(ex0824)str(ex0824)
Data set shows the number of votes for Buchanan and Bush in all 67 counties in Florida during the U.S. presidential election of November 7, 2000.
ex0825ex0825
A data frame with 67 observations on the following 3 variables.
a character vector indicating the county
votes cast for P. Buchanan
votes cast for G.W. Bush
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex0825)str(ex0825)
The data are the average mass, metabolic rate, and lifespan for 95 species of mammals. Kleiber's law states that the metabolic rate of an animal species, on average, is proportional to its mass raised to the power of 3/4.
ex0826ex0826
A data frame with 95 observations on the following 5 variables.
the common name of the mammal species
the scientific name of the mammal species
the average body mass in kg
the average metabolic rate in kJ per day
the average lifespan in years
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex0826)str(ex0826)
These data are armed Forces Qualifying Test (AFQT) score percentiles, years of education, and annual income in 2005 for a subset of a random sample of 2,584 Americans selected in 1979 who were working in 2005 and re-interviewed in 2006.
ex0828ex0828
A data frame with 2,584 observations on the following 4 variables.
the subject identification number
percentile score on the AFQT test
years of education achieved by 2005
annual income in 2005
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
National Longitudinal Survey of Youth, U.S. Bureau of Labor Statistics, https://www.bls.gov/nls/home.htm (May 8, 2008).
ex0222, ex0330, ex0331,
ex0524, ex0525, ex0923,
ex1033, ex1223
str(ex0828)str(ex0828)
These data are the prevalence of autism per 10,000 ten-year old children in the United States in 1992, 1994, 1996, 1998, and 2000.
ex0829ex0829
A data frame with 5 observations on the following 2 variables.
year
the number of autism cases per 10,000 ten-year old children
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Newschaffer, C. J., Falb, M. D. and Gurney, J. G. (2005) National Autism Prevalence Trends From United States Special Education Data, Pediatrics 115: e277–e282.
str(ex0829)str(ex0829)
In four regions of the US (Northeast, Midwest, South and West), in three different sized metropolitan regions, researchers measured indicators of pace of life.
ex0914ex0914
A data frame with 36 observations on the following 4 variables.
bank clerk speed
pedestrian walking speed
postal clerk talking speed
age adjusted death rate due to heart disease
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Levine, R.V. (1990). The Pace of Life, American Scientist 78: 450–459.
str(ex0914)str(ex0914)
Data on corn yield and rainfall in six U.S. corn–producing states (Iowa, Nebraska, Illinois, Indiana, Missouri and Ohio), recorded for each year from 1890 to 1927.
ex0915ex0915
A data frame with 38 observations on the following 3 variables.
year of observation (1890–1927)
average corn yield for the six states (in bu/acre)
average rainfall in the six states (in in/year)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Ezekiel, M. and Fox, K.A. (1959). Methods of Correlation and Regression Analysis, John Wiley & Sons, New York.
str(ex0915)str(ex0915)
Researchers studied the development of a fly (Drosophila subobscura) that had been accidentally introduced from the Old World into North America around 1980.
ex0918ex0918
A data frame with 21 observations on the following 8 variables.
a factor with levels "NA" and
"EU"
latitude (degrees)
average wing size (10log
mm) of female flies on log scale
standard error of wing size (10log mm) of female flies on log scale
average wing size (10log
mm) of male flies on log scale
standard error of wing size (10log mm) of male flies on log scale
average basal length to wing size ratios of female flies
standard error of average basal length to wing size ratio of female flies
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Huey, R.B., Gilchrist, G.W., Carlson, M.L., Berrigan, D. and Serra, L. (2000). Rapid Evolution of a Geographic Cline in Size in an Introduced Fly, Science 287(5451): 308–309.
str(ex0918)str(ex0918)
Data set contains the year of the Kentucky Derby, the winning horse, the condition of the track and the average speed of the winner for years 1896–2011.
ex0920ex0920
A data frame with 116 observations on the following 8 variables.
year of Kentucky Derby
a character vector with the name of the winning horse
number of horses that started the race
the net winnings of the winner, in U.S. dollars
the winning time in seconds
the winning average speed, n miles per hour
a factor indicating track condition with levels
"Fast", "Good", "Dusty", "Slow", "Heavy",
"Muddy", and "Sloppy"
a factor with with 2 levels of track condition,
with levels "Fast" and "Slow"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Kentucky Derby: Kentucky Derby Racing Results.
str(ex0920)str(ex0920)
The data are the typical dry weight in mg, the typical ingestion rate (weight of food intake per day for one animal) in mg/day, and the percentage of the food that is composed of organic matter for 19 species of deposit feeders. The goal is to see whether the distribution of species' ingestion rates depends on the percentage of organic matter in the food, after accounting for the effect of species weight and to describe the association.
ex0921ex0921
A data frame with 19 observations on the following 4 variables.
a character variable with the name of the species
the dry weight of the species, in mg
ingestion rate in mg per day
percentage of organic matter in the food
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Cammen, L. M. (1980) Ingestion Rate: An Empirical Model for Aquatic Deposit Feeders and Detritivores, Oecologia 44: 303–310.
str(ex0921)str(ex0921)
These data are a subset of the National Longitudinal Study of Youth data, with annual incomes in 2005, intelligence test scores (AFQT) measured in 1981, and years of education completed by 2006 for 1,306 males and 1,278 females who were between the ages of 14 and 22 when selected for the survey in 1979, who were available for re-interview in 2006, and who had paying jobs in 2005. Is there any evidence that the mean salary for males exceeds the mean salary for females with the same years of education and AFQT scores? By how many dollars or by what percent is the male mean larger?
ex0923ex0923
A data frame with 2,584 observations on the following 5 variables.
the subject identification number
a factor with levels "female" and "male"
percentile score on the AFQT intelligence test
years of education achieved by 2005
annual income in 2005
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
National Longitudinal Survey of Youth, U.S. Bureau of Labor Statistics, https://www.bls.gov/nls/home.htm (May 8, 2008).
ex0222, ex0330, ex0331,
ex0524, ex0525, ex0828,
ex1033, ex1223
str(ex0923)str(ex0923)
Researchers randomly allocated 25 beakers containing minnow larvae to receive one of 25 treatment combinations of 5 levels of zinc and 5 levels of copper.
ex1014ex1014
A data frame with 25 observations on the following 3 variables.
amount of copper received (in ppm)
amount of zinc received (in ppm)
protein in minnow larvae exposed to copper and
zinc (g/larva)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Ryan, D.A., Hubert, J.J., Carter, E.M., Sprague, J.B. and Parrott, J. (1992). A Reduced-Rank Multivariate Regression Approach to Aquatic Joint Toxicity Experiments, Biometrics 48(1): 155–162.
str(ex1014)str(ex1014)
Depletion of the ozone layer allows the most damaging ultraviolet radiation to reach the Earth's surface. To measure the relationship, researchers sampled the ocean column at various depths at 17 locations around Antarctica during the austral spring of 1990.
ex1026ex1026
A data frame with 17 observations on the following 3 variables.
percent inhibition of primary phytoplankton production in water
UVB exposure
a factor with levels "Deep" and
"Surface"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Smith, R.C., Prézelin, B.B., Baker, K.S., Bidigare, R.R., Boucher, N.P., Coley, T., Karentz, D., MacIntyre, S., Matlick, H.A., Menzies, D., Ondrusek, M., Wan, Z. and Waters, K.J. (1992). Ozone Depletion: Ultraviolet Radiation and Phytoplankton Biology in Antarctic Waters, Science 255(5047): 952–959.
str(ex1026)str(ex1026)
Data are measurements on breeding pairs of land-bird species collected from 16 islands around Britain over the course of several decades. For each species, the data set contains an average time of extinction on those islands where it appeared, the average number of nesting pairs, the size of the species and the migratory status of the species.
ex1027ex1027
A data frame with 62 observations on the following 5 variables.
a character vector indicating the species
average extinction time in years
average number of nesting pairs
a factor with levels "L" and "S"
a factor with levels "M" and "R"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Pimm, S.L., Jones, H.L., and Diamond, J. (1988). On the Risk of Extinction, American Naturalist 132(6): 757–785.
str(ex1027)str(ex1027)
Data set with the numbers of Atlantic Basin tropical storms and hurricanes for each year from 1950–1997. The variable storm index is an index of overall intensity of hurricane season. Also listed are whether the year was a cold, warm or neutral El Nino year and a variable indicating whether West Africa was wet or dry that year.
ex1028ex1028
A data frame with 48 observations on the following 7 variables.
year
a factor with levels "cold",
"neutral" and "warm"
numeric variable with values -1 if
ElNino is "cold", 0 if "neutral" and 1 if
"warm"
numeric variable indicating whether West Africa was wet (1) or dry (0)
number of storms
number of hurricanes
index of overall intensity of hurricane season
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Data were gathered by William Gray of Colorado State University and reported on USA Today weather page: http://www.usatoday.com/weather/whurnum.htm
str(ex1028)str(ex1028)
Data set contains weekly wages in 1987 for a sample of 25,632 males between the age of 18 and 70 who worked full-time along with their years of education, years of experience, indicator variable for whether they were black, indicator variable for whether they worked in or near a city, and a code for the region in the US where they worked.
ex1029ex1029
A data frame with 25,437 observations on the following 6 variables.
a factor with levels "Midwest",
"Northeast", "South" and "West"
a factor with levels
"MetopolitanArea" and "NotMetropolitanArea"
experience in years
education in years
a factor with levels "Black" and
"NotBlack"
weekly wage in dollars
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Bierens, H.J. and Ginther, D.K. (2001). Integrated Conditional Moment Testing of Quantile Regression Models, Empirical Economics 26(1): 307–324, doi:10.1007/s001810000059.
str(ex1029)str(ex1029)
A data set with weekly earnings for 4,952 males between the age of 18 and 70 sampled in the March 2011 Current Population Survey (CPS). These males are a subset who had reported earnings and who responded as having race as either “Only White” or “Only Black.” Also recorded are the region of the country (with four categories: Northeast, Midwest, South, and West), the metropolitan status of the men's employment (with three categories: Metropolitan, Not Metropolitan, and Not Identified), age, education category (with 16 categories ranging from “Less than first grade” to “doctorate Degree”), and education code, which is a numerical value that corresponds roughly to increasing levels of education (and so may be useful for plotting). What evidence do the data provide that the distributions of weekly earnings differ in the populations of white and black workers after accounting for the other variables? By how many dollars or by what percent does the White population mean (or median) exceed the Black population mean (or median)?
ex1030ex1030
A data frame with 4,952 observations on the following 7 variables.
a factor with levels "Midwest",
"Northeast", "South" and "West"
a factor with levels
"Metopolitan", "Not Metropolitan " and
"Not Identified"
age in years
a factor with 16 levels:
"SomeCollegeButNoDegree", "AssocDegAcadem",
"NinthGrade", "BachelorsDegree", "TenthGrade",
"HighSchoolDiploma", "AssocDegOccupVocat", "DoctorateDegree", "TwelthButNoDiploma","LessThanFirstGrade", "EleventhGrade", "ProfSchoolDegree","FifthorSixthGrade","SeventhOrEighthGrade", "FirstSecondThirdOrFourthGrade"
a numerical variable indicating the approximate ordering of EducationCategory, with higher numbers indicating more education
a factor with levels "Black" and
"White"
weekly wage in dollars
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
U.S. Bureau of Labor Statistics and U.S. Bureau of the Census: Current Population Survey, March 2011 https://www.bls.gov/cps
str(ex1030)str(ex1030)
A data set with Clutch Volume (cubic milimeters) and adult Body Mass (kg) in six different groups of animals: modern maternal-care bird species (Mat), modern paternal-care bird species (Pat), modern biparental-care bird species (BiPI), modern maternal-care crocodiles (Croc), non-avian maniraptoran dinosaurs thought to be ancestors of modern birds (Mani), and other non-avian dinosaurs (Othr). The question of interest was which group of modern creatures most closely matches the relationship in the maniraptoran dinosaurs.
ex1031ex1031
A data frame with 443 observations on the following 6 variables.
the common name of the species
species genus
species name
a factor with 6 levels corresponding to the 6 groups of
animals: "BiP", "Croc", "Mani", "Mat",
"Othr", and "Pat"
the average body mass of individuals in the species (kg)
the total volume of all eggs in a clutch (average value for the species)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Varricchio, D. J., Moore, J.r., Erickson, G.M., Norell, M.A., Jackon, F.D. and Borkowski, J.J. (2008) Avian Paternal Care Had Dinosaur Origin Science 322: 1826–1828
str(ex1031)str(ex1031)
This is a subset of the National Longitudinal Study of Youth (NLSY79) data, with annual incomes in 2005 (in U.S. dollars, as Recorded in a 2006 interview); scores on the Word Knowledge, Paragraph Comprehension, Arithmetic Reasoning, and Mathematics Knowledge portions of the Armed Forces Vocational Aptitude Battery (ASVAB) of tests taken in 1981; and the percentile score of the Armed Forces Qualifying Test (AFQT), which is a linear combination of the four component tests mentioned above (but note that AFQT reported here is the percentile, which is not a linear combination of the four component scores). Which of the five test scores seem to be the most important predictors of 2005 income? Is the AFQT sufficient by itself?
ex1033ex1033
A data frame with 2,584 observations on the following 7 variables.
the subject identification number
score on the Arithmetic Reasoning test in 1981
score on the Word Knowledge Test in 1981
score on the Paragraph Comprehension test in 1981
score on the Mathematics Knowledge test in 1981
percentile score on the AFQT intelligence test in 1981
annual income in 2005
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
National Longitudinal Survey of Youth, U.S. Bureau of Labor Statistics, https://www.bls.gov/nls/home.htm (May 8, 2008).
ex0222, ex0330, ex0331,
ex0524, ex0525, ex0828,
ex0923, ex1223
str(ex1033)str(ex1033)
The data are are the cesium concentrations (in Bq/kg) in soil and in mushrooms at 17 wooded locations in Umbria, Central Italy, from August 1986 to November 1989. Researchers wished to investigate the cesium transfer from contaminated soil to plants after the Chernobyl nuclear power plant accident in April 1986 by describing the distribution of the mushroom concentration as a function of soil concentration.
ex1111ex1111
A data frame with 17 observations on the following 2 variables.
the cesium concentration in mushrooms, Bq/kg
the cesium concentration in soil, Bq/kg
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1111)str(ex1111)
Data are the isotopic composition of structural bone carbonate
() and the isotopic composition of the coexisting calcite
cements () in 18 bone samples from a specimen of the dinosaur
Tyrannosaurus rex. Evidence that the mean of is positively
associated with was used in an argument that the metabolic
rate of this dinosaur resembled warm-blooded more than cold-blooded
animals.
ex1120ex1120
A data frame with 18 observations on the following 2 variables.
isotopic composition of bone carbonate
isotopic composition of calcite cements
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Barrick, R.E. and Showers, W.J. (1994). Thermophysiology of Tyrannosaurus rex: Evidence from Oxygen Isotopes, Science 265(5169): 222–224.
str(ex1120)str(ex1120)
It has been theorized that developing countries cut down their forests to pay off foreign debt. Data are debt, deforestation, and population from 11 Latin American nations.
ex1122ex1122
A data frame with 11 observations on the following 4 variables.
a character vector indicating the country
debt (millions of dollars)
deforestation (thousands of ha)
population (thousands of people)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Gullison, R.R. and Losos, E.C. (1992). The Role of Foreign Debt in Deforestation in Latin America, Conservation Biology 7(1): 140–7.
str(ex1122)str(ex1122)
Does pollution kill people? Data in one early study designed to explore this issue from 5 Standard Metropolitan Statistical Areas in the U.S between 1959–1961.
ex1123ex1123
A data frame with 60 observations on the following 7 variables.
a character vector indicating the city
total age-adjusted mortality from all causes
mean annual precipitation (inches)
median number of school years completed for persons 25 years or older
percentage of population that is nonwhite
relative pollution potential of oxides of nitrogen
relative pollution potential of sulfur dioxide
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
McDonald, G.C. and Ayers, J.A. (1978). Some Applications of the “Chernoff Faces”: A Technique for Graphically Representing Multivariate Data in Wang, P.C.C. (ed.) Graphical Representation of Multivariate Data, Academic Press.
str(ex1123)str(ex1123)
An assessment of the factors affecting dispersal distances is important for understanding population spread, recolonization and gene flow which are central issues for conservation of many vertebrate species. Researchers gathered data on body weight, diet type and maximum natal dispersal distance for various animals.
ex1124ex1124
A data frame with 64 observations on the following 4 variables.
a character vector indicating the species
bodymass (kg)
maximum dispersal distance (km)
a factor with levels "C", "H" and
"O" indicating carnivore, herbivore, or omnivore
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Sutherland, G.D., Harestad, A.S., Price, K. and Lertzman, K.P. (2000). Scaling of Natal Dispersal Distances in Terrestrial Birds and Mammals, Conservation Ecology 4(1): 16.
str(ex1124)str(ex1124)
The data are the typical dry weight in mg, the typical ingestion rate (weight of food intake per day for one animal) in mg/day, and the percentage of the food that is composed of organic matter for 22 species of deposit feeders. The goal is to see whether the distribution of species' ingestion rates depends on the percentage of organic matter in the food, after accounting for the effect of species weight and to describe the association. The last three species happen to be Bivalves, so may behave differently from the others.
ex1125ex1125
A data frame with 22 observations on the following 5 variables.
a character variable with the name of the species
the dry weight of the species, in mg
ingestion rate in mg per day
percentage of organic matter in the food
a factor with levels "no" and "yes" to
indicate whether a species is a bivalve
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Cammen, L. M. (1980) Ingestion Rate: An Empirical Model for Aquatic Deposit Feeders and Detritivores, Oecologia 44: 303–310.
str(ex1125)str(ex1125)
Complete data set for problem introduced in ex1123.
Data from early study designed to explore the relationship between air
pollution and mortality.
ex1217ex1217
A data frame with 60 observations on the following 17 variables.
a character vector indicating the city
total age-adjusted mortality from all causes
mean annual precipitation (inches)
percent relative humidity (annual average at 1:00pm)
mean January temperature (degrees F)
mean July temperature (degrees F)
percentage of the population aged 65 years or over
population per household
median number of school years completed for persons 25 years or older
percentage of the housing that is sound with all facilities
population density (in persons per square mile of urbanized area)
percentage of population that is nonwhite
percentage of employment in white collar occupations
percentage of households with annual income under
3,000 in 1960
relative pollution potential of hydrocarbons
relative pollution potential of oxides of nitrogen
relative pollution potential of sulfur dioxide
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
McDonald, G.C. and Ayers, J.A. (1978). Some Applications of the “Chernoff Faces”: A Technique for Graphically Representing Multivariate Data in Wang, P.C.C. (ed.) Graphical Representation of Multivariate Data, Academic Press.
str(ex1217)str(ex1217)
The number of species on an island is known to be related to the island's area. Of interest is what other variables are also related to the number of species, after island area is accounted for, and whether the answer differs for native and non native species.
ex1220ex1220
A data frame with 30 observations on the following 8 variables.
a character vector indicating the island
total number of observed species
number of native species
area (km)
elevation (m)
distance from nearest island (km)
distance from Santa Cruz (km)
area of nearest island (km)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Johnson, M.P. and Raven, P.H. (1973). Species Number and Endemism: The Galapagos Archipelago Revisited, Science 179(4076): 893–895.
str(ex1220)str(ex1220)
These data are monthly rainfalls from September to March and the subjectively rated quality of the following spring wildflower display for each of a number of years at each of four desert locations in the southwestern United States (Upland Sonoran Desert near Tucson, the lower Colorado River Valley section of the Sonoran Desert, the Baja California region of the Sonoran Desert, and the Mojave Desert). The quality of the display was judged subjectively with ordered rating categories of poor, fair, good, great, and spectacular. The variable Score is numerical variable corresponding to these ordered categories. A goal is to find an equation for predicting quality of wildflower blooms from the rainfall variables.
ex1221ex1221
A data frame with 122 observations on the following 12 variables.
year of observed wildflower season
a factor variable with 4 levels: "baja",
"colorado", "mojave", and "upland"
the September rainfall, in inches
the October rainfall, in inches
the November rainfall, in inches
the December rainfall, in inches
the January rainfall, in inches
the February rainfall, in inches
the March rainfall, in inches
the total rainfall from September through March, in inches
a factor with a subjective assessment of the quality of
wildflower bloom with levels "FAIR", "GOOD", "GREAT",
"POOR", and "SPECTACULAR"
a numerical variable corresponding to the order of rating categories, with Poor=0, Fair=1, Good=2, Great=3, and Spectacular=4
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Arizona-Sonora Desert Museum, “Wildflower Flourishes and Flops: a 50–Year History,” https://www.desertmuseum.org/programs/flw_wildflwrbloom.php (July 25, 2011).
str(ex1221)str(ex1221)
This data set contains the vote counts by county in Florida for Buchanan and for four other presidential candidates in 2000, along with the total vote counts in 2000, the presidential vote counts for three presidential candidates in 1996, the vote count for Buchanan in his only other campaign in Florida—the 1996 Republican primary, the registration in Buchanan's Reform party and the total political party registration in the county.
ex1222ex1222
A data frame with 67 observations on the following 13 variables.
a character vector indicating the county
votes cast for Buchanan in 2000 presidential election
votes cast for Gore in 2000 presidential election
votes cast for Bush in 2000 presidential election
votes cast for Nader in 2000 presidential election
votes cast for Browne in 2000 presidential election
total vostes cast in 2000 presidential election
votes cast for Clinton in 1996 presidential election
votes cast for Dole in 1996 presidential election
votes cast for Perot in 1996 presidential election
votes cast for Buchanan in 1996 Republican primary
the registration in Buchanan's Reform party
the total political party registration
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1222)str(ex1222)
This is a subset of 2,584 individuals from the 1979 National Longitudinal Study of Youth (NLSY79) survey who were re-interviewed in 2006, who had paying jobs in 2005, and who had complete values for the variables listed below. A goal is to see whether intelligence (as measured by the ASVAB intelligence test score, AFQT, and its Components, Word, Parag, Math, and Arith) is a better predictor of 2005 income than education and socioeconomic status.
ex1223ex1223
A data frame with 2,584 observations on the following 32 variables.
the subject identification number
a variable taking on the value 1 if anyone in the respondent's household regularly read magazines in 1979, otherwise 0
a variable taking on the value 1 if anyone in the respondent's household regularly read newspapers in 1979, otherwise 0
a variable taking on the value 1 if anyone in the respondent's household had a library card in 1979, otherwise 0
mother's years of education
father's years of education
family's total net income in 1978
1 = Hispanic, 2 = Black, 3 = Not Hispanic or Black
a factor with levels "female" and "male"
years of education completed by 2006
score on the General Science test in 1981
score on the Arithmetic Reasoning test in 1981
score on the Word Knowledge Test in 1981
score on the Paragraph Comprehension test in 1981
score on the Numerical Operations test in 1981
score on the Coding Speed test in 1981
score on the Automotive and Shop test in 1981
score on the Mathematics Knowledge test in 1981
score on the Electronics Information test in 1981
score on the Paragraph Comprehension test in 1981
percentile score on the AFQT intelligence test in 1981
total annual income in 2005
self reported answer to 1st self esteem question, 2005
self reported answer to 2md self esteem question, 2005
self reported answer to 3rd self esteem question, 2005
self reported answer to 4th self esteem question, 2005
self reported answer to 5th self esteem question, 2005
self reported answer to 6th self esteem question, 2005
self reported answer to 7th self esteem question, 2005
self reported answer to 8th self esteem question, 2005
self reported answer to 9th self esteem question, 2005
self reported answer to 10th self esteem question, 2005
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
National Longitudinal Survey of Youth, U.S. Bureau of Labor Statistics, https://www.bls.gov/nls/home.htm (May 8, 2008).
ex0222, ex0330, ex0331,
ex0524, ex0525, ex0828,
ex0923, ex1033
str(ex1223)str(ex1223)
These data are weekly earnings for 9,835 Americans surveyed in the March 2011 Current Population Survey (CPS). What evidence is there from these data that males tend to receive higher earnings than females with the same values of the other variables? By how many dollars or by what percent does the male distribution exceed the female distribution?
ex1225ex1225
A data frame with 9,835 observations on the following 9 variables.
a factor with levels "Midwest", "Northeast",
"South", and "West"
a a factor with levels "Metropolitan",
"Not Identified", and "Not Metropolitan"
age in years
a factor with levels "Female" and "Male"
a factor with levels "Married" and
"NotMarried"
a numerical variable representing educational attainment, with higher numbers corresponding to higher educational categories
a factor with 16 levels of educational category
a a factor with levels "FedGov",
"LocalGov", "Private",and "StateGov"
weekly wages in U.S. dollars
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
U.S. Bureau of Labor Statistics and U.S. Bureau of the Census: Current Population Survey, March 2011 https://www.bls.gov/cps/data.htm July 25, 2011.
str(ex1225)str(ex1225)
About 65 million years ago, the dinosaurs suffered a mass extinction virtually overnight (in geologic time). Among many clues, one that all scientists regard as crucial is a layer of iridium-rich dust that was deposited over much of the earth at that time. The theory is that an event like a volcanic eruption or meteor impact caused a massive dust cloud that blanketed the earth for years killing off animals and their food sources. Dataset has Iridium depths by type of deposit.
ex1317ex1317
A data frame with 28 observations on the following 3 variables.
Iridium in samples (ppt)
a factor with levels "Limestone" and
"Shale"
a factor with six levels: "1", "2",
..., "6"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Alvarez, W. and Asaro, F. (1990). What Caused the Mass Extinction? An Extraterrestrial Impact, Scientific American 263(4): 76–84.
Courtillot, E. (1990). What Caused the Mass Extinction? A Volcanic Eruption. Scientific American 263(4): 85–92.
str(ex1317)str(ex1317)
A 1989 study investigated the effect of heredity and environment on intelligence. Data are the IQ scores for adopted children whose biological and adoptive parents were categorized either in the highest or the lowest socioeconomic status category.
ex1319ex1319
A data frame with 38 observations on the following 3 variables.
IQ scores of adopted children
a factor with levels "High" and
"Low"; the socioeconomic status of the adoptive parents
a factor with levels "High" and
"Low"; the socioeconomic status of the biological parents
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Capron, C. and Duyme, M. (1991). Children's IQ's and SES of Biological and Adoptive Parents in a Balanced Cross-fostering Study, European Bulletin of Cognitive Psychology 11(3): 323–348.
str(ex1319)str(ex1319)
Data set on 861 ACT Assessment Mathematics Usage Test scores from 1987. The test was given to a sample of high school seniors who met one of three profiles of high school mathematics course work: (a) Algebra I only; (b) two Algebra courses and Geometry; and (c) two Algebra courses, Geometry, Trigonometry, Advanced Mathematics and Beginning Calculus.
These data were generated from summary statistics for one particular form of the test as reported by Doolittle (1989).
ex1320ex1320
A data frame with 861 observations on the following 3 variables.
a factor with levels "female" and
"male"
a factor with levels "a", "b"
and "c"
ACT mathematics test score
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Doolittle, A.E. (1989). Gender Differences in Performance on Mathematics Achievement Items, Applied Measurement in Education 2(2): 161–177.
str(ex1320)str(ex1320)
A data set simulated to match the summary statistics and conclusions from Rosenthal and Jacobson's Pygmalion study on elementary school students. The researchers assigned students at random to a pygmalion or control treatment group. They supplied information to the teachers of those in the pygmalion group with the false information that an intelligence test had indicated that the student was likely to excel. The researchers wished to see if the change in intelligence test scores for the students tended to be larger for those students labeled as likely to excel.
ex1321ex1321
A data frame with 320 observations on the following 5 variables.
a student identification number
the student's grade, 1 through 6
a factor with 17 levels "1a", "1b", and
so on, to indicate the 17 distinct teacher/classrooms.
a factor with levels "pygmalion" and
"control" corresponding to whether the researchers had told the
teacher that the student was “likely to succeed” or not
the intelligence test score taken at the end of the school year minus the intelligence test score taken at the begging of the school year
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Rosenthal, R. and Jacobson, L. 1968, Pygmalion in the Classroom: Teacher Expectation and Pupil's Intellectual Development, Holt, Rinehart, and Winston, Inc.
str(ex1321)str(ex1321)
Researchers designed an experiment to investigate how delivery of brain cancer antibody is influenced by tumor size, antibody molecular weight, blood-brain barrier disruption, and delivery route.
ex1416ex1416
A data frame with 36 observations on the following 6 variables.
a factor with levels "AIB", "DEX7"
and "MTX"
a factor with levels "BD" and
"NS"
a factor with levels "IA" and
"IV"
days after inoculation
concentration of antibody in the part of the brain around the tumor
concentration of antibody in the unaffected part of the brain
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Barnett, P.A., Roman-Goldstain, S., Ramsey, F., McCormick, C.I., Sexton, G., Szumowski, J. and Neuwelt, E.A. (1995). Differential Permeability and Quantitative MR Imaging of a Human Lung Carcinoma Brain Xenograft in the Nude Rat, American Journal of Pathology 146(2): 436–449.
str(ex1416)str(ex1416)
Researchers designed an experiment to investigate how delivery of
brain cancer antibody is influenced by tumor size, antibody molecular
weight, blood-brain barrier disruption, and delivery route. The data
for the first replicate of this study is in ex1416.
This is the second replicate for the study.
ex1417ex1417
A data frame with 36 observations on the following 6 variables.
a factor with levels "AIB", "DEX70"
and "MTX"
a factor with levels "BD" and
"NS"
a factor with levels "IA" and
"IV"
days after inoculation
concentration of antibody in the part of the brain around the tumor
concentration of antibody in the unaffected part of the brain
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Barnett, P.A., Roman-Goldstain, S., Ramsey, F., McCormick, C.I., Sexton, G., Szumowski, J. and Neuwelt, E.A. (1995). Differential Permeability and Quantitative MR Imaging of a Human Lung Carcinoma Brain Xenograft in the Nude Rat, American Journal of Pathology 146(2): 436–449.
str(ex1417)str(ex1417)
These data were simulated to match the summary statistics and conclusions of Rosenthal and Fode's Clever Hans experiment. Each of 12 students trained rats to run a maze. The data set contains their number of successful runs out of 50 on each of 5 days. It also shows two summarizing statistics for each student: the overall success rate on all 5 days and the slope in the least squares regression of daily success rate (number of successes in a day divided by 50) on day. Also included are the student's response to the prior expectation of success question and the student's response to a post- experiment question about how relaxed they felt handling their rats (with higher values corresponding to more relaxed). The treatment variable shows whether or not the students were supplied with the fictitious information about whether their rats were bright or not.
ex1419ex1419
A data frame with 12 observations on the following 12 variables.
a student identification number
the student's prior expectation of rat-training success, on a scale from -10 to 10
a numerical variable for pairs of students grouped according to their values of PriorExp
a factor with levels "bright" and "dull"
corresponding to whether students were told (falsely) that their rats were
bright or not
the number of successful rat mazed runs on day 1
the number of successful rat mazed runs on day 2
the number of successful rat mazed runs on day 3
the number of successful rat mazed runs on day 4
the number of successful rat mazed runs on day 5
degree of relaxation students felt in handling their rats, on a scale from 0 to 10
the total proportion of successful maze runs in 5 days
the slope in the least squares regression of mean daily success as a function of day, estimated for each student individually
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Rosenthal, R. and Fode, K.L. (1963) The Effect of Experimenter Bias on the Performance of the Albino Rat Behavioral Science 8:3: 183–189.
str(ex1419)str(ex1419)
These data are the weight losses of subjects randomly assigned to one of three diets, and these additional covariates sex, initial age, and body mass index. Is there any evidence from these data that the mean weight loss differs for the different diets, after accounting for the effect of the covariates? How big are the difference?
ex1420ex1420
A data frame with 272 observations on the following 6 variables.
a subject identification number
a factor with levels "Low-Carbohydrate",
"Low-Fat"and "Mediterranean"
a factor with levels "F" and "M"
subject's age in years
body mass index in kg/squared meter
the amount of weight lost in 24 months on the diet, in kg
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1420)str(ex1420)
The data are the temperatures (in degrees Celsius) averaged for the southern hemisphere over a full year, for years 1850 to 2010. The 161-year average temperature has been subtracted, so each observation is the temperature difference from the series average.
ex1507ex1507
A data frame with 161 observations on the following 2 variables.
year in which yearly average temperature was computed, from 1850 to 2010
southern hemisphere temperature minus the 161-year average (degrees Celsius)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Jones, P.D., D. E. Parker, T. J. Osborn, and K. R. Briffa, (2011) Global and Hemispheric Temperature Anomalies and and Marine Instrumental Records, CDIAC, http://cdiac.ornl.gov/trends/temp/jonescru/jones.html, Aug 4, 2011
str(ex1507)str(ex1507)
The data are the annual sunspot counts in each year from 1700 to 2010.
ex1507ex1507
A data frame with 311 observations on the following 2 variables.
year in which sunspots were counted, from 1700 to 2010
the number of sunspots observed in a year
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
SIDC-Solar Influences Dta Center, http://sidc.oma.be/sunspot-data/ (July 15, 2011).
str(ex1509)str(ex1509)
Several factors suggest that the incidence of melanoma is related to solar radiation. These data are the age-adjusted melanoma incidence among males in the Connecticut Tumor Registry and the sunspot activity, 1936–1972 .
ex1514ex1514
A data frame with 37 observations on the following 3 variables.
year
male melanoma incidence in number of cases per 100,000 population
sunspot relative number
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Houghton, A., Munster, E.W. and Viola, M.V. (1978). Increased Incidence of Malignant Melanoma After Peaks of Sunspot Activity, Lancet: 759–760.
str(ex1514)str(ex1514)
Data on the annual numbers of lynx trapped in the Mackenzie River district of northwest Canada from 1821–1934.
ex1515ex1515
A data frame with 114 observations on the following 3 variables.
year
number of lynx trapped
number of sunspots
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Elston, C. and Nicholson, M. (1942). The Ten Year Cycle in Numbers of the Lynx in Canada, Journal of Animal Ecology 11(2): 215–244.
str(ex1515)str(ex1515)
Data shows the number of deaths due to firearms and the number of deaths due to motor vehicle accidents in the United States between 1968 and 1993.
ex1516ex1516
A data frame with 26 observations on the following 3 variables.
year
deaths due to firearms (in thousands per year)
deaths due to motor vehicles (in thousands per year)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Data read from a Centers for Disease Control and Prevention graph reported in The Oregonian, June 17, 1997.
str(ex1516)str(ex1516)
The Standard and Poor's 500 stock index (S&P 500) is a benchmark of stock market performance, based on 400 industrial firms, 40 financial stocks, 40 utilities, and 20 transportation stocks. These data include the value of a $1 investment in 1871 at the end of each year from 1871 to 1999, according to the S&P 500, assuming all dividends are reinvested. Describe the distribution of the S&P value as a function of year.
ex1517ex1517
A data frame with 129 observations on the following 2 variables.
year
Value of Stock at the end of the year
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1517)str(ex1517)
The data are the number of measles cases reported in the United States for each year from 1950 to 2008. A goal is to explore the effect of the introduction of the measles vaccine in 1963 on the series mean.
ex1518ex1518
A data frame with 59 observations on the following 3 variables.
year
number of measles cases
a factor with levels "no" and "yes"
indicating whether the measles vaccine had been licensed or not (yes for
every year starting with 1963)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Center for Disease Control,
https://www.cdc.gov/vaccines/pubs/pinkbook/downloads/appendices/G/cases-deaths.pdf
retrieved on July 23, 2009
str(ex1518)str(ex1518)
The data are the Sea Surface Temperatures (SST) and Southern Oscillation Index (SOI) measurements from 1950 to 2010.
ex1519ex1519
A data frame with 732 observations on the following 4 variables.
year
a numerical variable for month, with 1 = January
the Southern Oscillation Index)
the Sea Surfact Temperature)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
U.S. National Oceanographic and Atmospheric Administration (http://www.cpc.ncep.noaa.gov/data/indices/).
str(ex1519)str(ex1519)
Data are a subset from an observational, longitudinal, study on adopted children. Is child's intelligence related to intelligence of the biological mother and the intelligence of the adoptive mother?
ex1605ex1605
A data frame with 62 observations on the following 6 variables.
adoptive (foster) mother's years of education
biological mother's score on IQ test
IQ of child at age 2
IQ of child at age 4
IQ of child at age 8
IQ of child at age 13
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Skodak, M. and Skeels, H.M. (1949). A Final Follow-up Study of One Hundred Adopted Children, Journal of Genetic Psychology 75: 85–125.
str(ex1605)str(ex1605)
Adam Smith, in Wealth of Nations, observed that even religious monopolies become weak when they are not challenged by competition. Data to illustrate this point is from 21 countries in which the percentages of Catholics in the populations varied from a low 1.2% to a high 97.6%.
ex1611ex1611
A data frame with 21 observations on the following 4 variables.
a character vector indicating the country
percent Catholics in the population
priest to parishioner ratio
percent clergy indigenous
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Gill, A.J. (1994). Rendering unto Caesar? Religious Competition and Catholic Political Strategy in Latin America, 1962–79, American Journal of Political Science 38(2): 403–425.
str(ex1611)str(ex1611)
Samples of effluent were divided and sent to two laboratories for testing. Data are measurements of biochemical oxygen demand and suspended solid measurements obtained for 2 sample splits from the two laboratories.
ex1612ex1612
A data frame with 11 observations on the following 4 variables.
biochemical oxygen demand measurements from commercial laboratory
suspended solids measurements from commercial laboratory
biochemical oxygen demand measurements from state laboratory
suspended solids measurements from state laboratory
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Johnson, R.A. and Wichern, D.W. (1988). Applied Multivariate Statistical Analysis, Prentice-Hall.
str(ex1612)str(ex1612)
Data are the measurements from two very similar species of flea beetle.
ex1613ex1613
A data frame with 36 observations on the following 4 variables.
specimen identification number
measurement of first joint in micrometers
measurement of second joint in micrometers
a factor with levels "conc" and
"heik"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Lubischew, A.A. (1962). On the Use of Discriminant Functions in Taxonomy, Biometrics 18: 455–477.
str(ex1613)str(ex1613)
Recent studies in the field of psychoimmunology suggest a link exists between behavioral events and the functioning of one's immune system. Data shows the results of a study on 12 subjects who were monitored during three distinct activities. The first activity consisted of neutral activity such as reporting tasks. During the second activity, subjects listened to audiotape exercises relating to images of heaviness, warmth in the body, relaxation, suggestions to remember happy events, etc. The third activity included a nonaudio tape follow up stimulus consisting of continued relaxation as in activity 2 and a verbal discussion of the positive aspects of the audiotape.
ex1614ex1614
A data frame with 12 observations on the following 3 variables.
subject identification number
Interleukin-1 levels (counts per minute) from blood samples taken during activity A
Interleukin-1 levels (counts per minute) from blood samples taken during activity B
Interleukin-1 levels (counts per minute) from blood samples taken during activity C
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Keppel, W. (1993). Effects of Behavioral Stimuli on Plasma Interleukin-1 Activity in Humans at Rest, Journal of Clinical Psychology 49(6): 777–785.
str(ex1614)str(ex1614)
Data shows a partial listing of a data set with ratios of average math to average verbal SAT scores in the United States and the District of Columbia for 1989 and 1996–1999.
ex1615ex1615
A data frame with 51 observations on the following 6 variables.
a character vector indicating the state
average MATH SAT scores divided by average VERBAL SAT score in 1989
average MATH SAT scores divided by average VERBAL SAT score in 1996
average MATH SAT scores divided by average VERBAL SAT score in 1997
average MATH SAT scores divided by average VERBAL SAT score in 1998
average MATH SAT scores divided by average VERBAL SAT score in 1999
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1615)str(ex1615)
In an experiment to identify genes of the plant Arabidopsis that react to a particular pathogen, researchers used RNA sequencing to produce gene profiles for a number of plants not subjected to the pathogen and several plants subjected to the pathogen. Tests comparing the distribution of gene expression in the two groups were performed for each gene individually. The data are the p-values from all these tests. The goal is to use a identify a set of genes that differentially express in the two groups, subject to some specified value for expected false discovery rate, such as 5%.
ex1620ex1620
A data frame with 20,245 observations on the following 3 variables.
an identification number for genes
a character variable with the name of the gene
the p-value from a test that the mean expression level for the gene differs in the two groups (ignoring multiple testing)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Chang, J, Department of Botany and Plant Pathology, Oregon State University, personnal communication.
str(ex1620)str(ex1620)
Actual pig fat and measurements of pig fat from magnetic resonance images at 13 locations for 12 pigs.
ex1708ex1708
A data frame with 12 observations on the following 14 variables.
actual pig fat (in percent)
magnetic resonance image at location 1
magnetic resonance image at location 2
magnetic resonance image at location 3
magnetic resonance image at location 4
magnetic resonance image at location 5
magnetic resonance image at location 6
magnetic resonance image at location 7
magnetic resonance image at location 8
magnetic resonance image at location 9
magnetic resonance image at location 10
magnetic resonance image at location 11
magnetic resonance image at location 12
magnetic resonance image at location 13
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Glasbey, C.A and Flowler, P.A. (1992). Regression Models Fitted Using Conditional Independence to Estimate Pig Fatness from Magnetic Resonance Images, The Statistician 41(2): 179–184.
str(ex1708)str(ex1708)
Data show measures that differ among denominations of American Protestant and Catholic churches.
ex1715ex1715
A data frame with 18 observations on the following 6 variables.
a character vector indicating the church denomination
distinctiveness (strictness of discipline on a seven point scale)
average percentage of weeks that individuals attended a church meeting (% weekly)
average number of secular organisations to which members belong
average percentage of members that describe themselves as being strong church members (%)
average income of members (US)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Iannaccone, L.R. (1994). Why Strict Churches Are Strong, American Journal of Sociology 99(5): 1180–1211.
str(ex1715)str(ex1715)
In the 1970's the U.S. Commission on Civil Rights investigated charges that insurance companies were attempting to redefine Chicago “neighborhoods” in order to cancel existing homeowner insurance policies or refuse to issue new ones. Dataset has data on homeowner and residential fire insurance policy issuances from 47 zip codes in the Chicago area.
ex1716ex1716
A data frame with 47 observations on the following 8 variables.
last 2 digits of zip code
fires per 1000 housing units
thefts per 1000 population
percentage of housing units built prior to 1940
median family income
percentage minority
number of new policies per 100 housing units
number of FAIR plan policies and renewals per 100 housing units
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Andrews, D.F. and Herzberg, A.M. (1985). Data: A Collection of Problems from many Fields for the Student and Research Worker, Springer-Verlag.
str(ex1716)str(ex1716)
Survival times for two groups of lymphoma patients.
ex1914ex1914
A data frame with 34 observations on the following 4 variables.
months after diagnosis
a factor with levels "no" and
"radiation"
number of patients known to survive beyond this month
number of patients known to die after this many months
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Neuwelt, E.A., Goldman, D.L., Dahlborg, S.A., Crossen, J., Ramsey, F., Roman-Goldstein, S., Braziel, R. and Dana, B. (1991). Primary CNS Lymphoma Treated with Osmotic Blood-brain Barrier Disruption: Prolonged Survival and Preservation of Cognitive Function, Journal of Clinical Oncology 9(9): 1580–1590.
str(ex1914)str(ex1914)
Fictitious data set based on results of an experiment where subjects were randomly divided into two groups and given a placebo or vitamin c to take during the cold season. At the end of the cold season, the subjects were interviewed by a physician who determined whether they had or had not suffered a cold during the period. Skeptics interviewed the 800 subjects to determine who knew and who did not know to which group they had been assigned. Vitamin C has a bitter taste and those familiar with it could recognize whether their pills contained it.
ex1916ex1916
A data frame with 4 observations on the following 4 variables.
a factor with levels "no" and "yes"
a factor with levels "placebo" and
"vitC"
number of people who got a cold
number of people who did not get a cold
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1916)str(ex1916)
Dataset from a study which investigated the added risk of breast cancer due to alcohol consumption. A sample of confirmed breast cancer patients were compared with a sample of cancer free women who were close in age and from the same neighborhood as the cases. Data was collected on the alcohol consumption and body mass of both sets of women.
ex1917ex1917
A data frame with 6 observations on the following 4 variables.
a factor with levels "high",
"low" and "medium"
a factor with levels "high" and
"low"
number of women with breast cancer
number of women without breast cancer
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Rosenberg, L., Palmer, J.R., Miller, D.R., Clarke, E.A. and Shapiro, S. (1990). A Case-Control Study of Alcoholic Beverage Consumption and Breast Cancer, American Journal of Epidemiology 131(1): 6–14.
str(ex1917)str(ex1917)
In 1846 the Donner party became stranded while crossing the Sierra Nevada Mountains near Lake Tahoe. The data frame has the counts for male and female survivors for six age groups.
ex1918ex1918
A data frame with 12 observations on the following 4 variables.
a numerical code corresponding to six age categories, with 1 = "15-19",
2 = "20-29", 3 = "30-39", 4 = "40-49", 5 = "50-59" and
6 = "60-69"
a factor with levels "female" and
"male"
number that lived
number that died
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Grayson, D.K. (1990). Donner Party Deaths: A Demographic Assessment, Journal of Anthropological Research 46: 223–242.
str(ex1918)str(ex1918)
Data shows the numbers of compact sports utility vehicles involved in fatal accidents in the U.S. between 1995 and 1999, categorized according to travel speed, make of car (Ford or other), and cause of accident (tire-related or other).
ex1919ex1919
A data frame with 8 observations on the following 4 variables.
a numerical code corresponding to 4 categories of speed (in miles per hour), with 1 = "0-40",
2 = "41-55", 3 = "56-65" and 4 = ">65"
a factor with levels "Ford" and
"Other"
cause of accident was other than tire-related
cause of accident was tire-related
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1919)str(ex1919)
In the study of different diets for losing weight
(ex0623, ex1420 and ex1922),
there appear to have been many more experimental subjects that dropped out
from the low carbohydrate diet group than from the other two diet groups. This
data set contains the numbers who did and didn't drop out in each diet group.
Is there any evidence that the drop out rate differs in the three groups?
ex1921ex1921
A data frame with 3 observations on the following 4 variables.
a factor with levels "LowCarb",
"LowFat", and "Medit"
the number of subjects who dropped out of the study
the number of subjects who completed the study
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1921)str(ex1921)
For the study of different diets for losing weight
(ex0623, ex1420 and
ex1921), it is desired to see whether women were more or
less likely to
drop out of the study than men (after accounting for the apparent differential
drop out rates associated with diet). This data set includes the numbers that
dropped out and completed the study for each combination of Sex and Diet.
ex1922ex1922
A data frame with 6 observations on the following 4 variables.
a factor with levels "LowCarb",
"LowFat", and "Medit"
a factor with levels "Men" and "Women"
the number of subjects who dropped out of the study
the number of subjects who completed the study
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1922)str(ex1922)
One issue concerning the validity of the clutch volume and parental care study
of ex1031 is the selection of the bird species in the
set of currently living animals. Was the selection just as good as a random
sample of species from each of the groups? One way to study this for birds, at
least, is to compare the numbers of species from each of the 29 orders of birds
in the study with the known total number of species in each of the orders. If
the selection of birds had been at random, the expected proportion of species
in the study from one particular order, n, is the proportion of all species in
that order (N=9,866) times the total number of species in the sample (414).
That is, the expected number in each sample, if random sampling were used, is
(N/9,866)x 414. Calculate the expected numbers and compare the observed numbers
with them using Pearson's chi-square statistic.
ex1923ex1923
A data frame with 29 observations on the following 3 variables.
a character variable with the name of the order
the known number of species in the order
the number of sampled species from the order
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex1923)str(ex1923)
This data frame contains the launch temperatures (degrees Fahrenheit) and an indicator of O-ring failures for 24 space shuttle launches prior to the space shuttle Challenger disaster of January 28, 1986.
ex2011ex2011
A data frame with 24 observations on the following 2 variables.
Launch temperature (in degrees Fahrenheit)
Indicator of O-ring failure
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex2011)str(ex2011)
Duchenne Muscular Dystrophy (DMD) is a genetically transmitted disease, passed from a mother to her children. Boys with the disease usually die at a young age; but affected girls usually do not suffer symptoms, may unknowingly carry the disease and may pass it to their offspring. It is believed that about 1 in 3,300 women are DMD carriers. A woman might suspect she is a carrier when a related male child develops the disease. Doctors must rely on some kind of test to detect the presence of the disease. This data frame contains data on two enzymes in the blood, creatine kinase (CK) and hemopexin (H) for 38 known DMD carriers and 82 women who are not carriers. It is desired to use these data to obtain an equation for indicating whether a women is a likely carrier.
ex2012ex2012
A data frame with 120 observations on the following 3 variables.
Indicator whether the woman has DMD ("Case") or
not ("Control")
Creatine kinase reading
Hemopexin reading
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Andrews, D.F. and Herzberg, A.M. (1985). Data: A Collection of Problems From Many Fields For The Student And Research Worker, Springer-Verlag, New York.
str(ex2012)str(ex2012)
A study examined the association between nesting locations of the Northern Spotted Owl and availability of mature forests. Wildlife biologists identified 30 nest sites. The researchers selected 30 other sites at random coordinates in the same forest. On the basis of aerial photographs, the percentage of mature forest (older than 80 years) was measured in various rings around each of the 60 sites.
ex2015ex2015
A data frame with 60 observations on the following 8 variables.
Site, a factor with levels "Random" and
"Nest"
Percentage of mature forest in ring with outer radius 0.91 km
Percentage of mature forest in ring with outer radius 1.18 km
Percentage of mature forest in ring with outer radius 1.40 km
Percentage of mature forest in ring with outer radius 1.60 km
Percentage of mature forest in ring with outer radius 1.77 km
Percentage of mature forest in ring with outer radius 2.41 km
Percentage of mature forest in ring with outer radius 3.38 km
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Ripple W.J., Johnson, D.H., Thershey, K.T. and Meslow E.C. (1991). Old–growth and Mature Forests Near Spotted Owl Nests in Western Oregon, Journal of Wildlife Management 55(2): 316–318.
str(ex2015)str(ex2015)
Hermon Bumpus analysed various characteristics of some house sparrows that were found on the ground after a severe winter storm in 1898. Some of the sparrows survived and some perished. This data set contains the survival status, age, the length from tip of beak to tip of tail (in mm), the alar extent (length from tip to tip of the extended wings, in mm), the weight in grams, the length of the head in mm, the length of the humerus (arm bone, in inches), the length of the femur (thigh bones, in inches), the length of the tibio–tarsus (leg bone, in inches), the breadth of the skull in inches and the length of the sternum in inches.
ex2016ex2016
A data frame with 87 observations on the following 11 variables.
Survival status, factor with levels
"Perished" and "Survived"
a numerical code corresponding to two categories of age, with 1 = "adult" and
2 = "juvenile"
total length (in mm)
alar extent (in mm)
weight (in grams)
length of beak and head (in mm)
length of humerus (in inches)
length of femur (in inches)
length of tibio–tarsus (in inches)
width of skull (in inches)
length of keel of sternum (in inches)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex2016)str(ex2016)
The Catholic church has explicitly opposed authoritarian rule in some (but not all) Latin American countries. Although such action could be explained as a desire to counter repression or to increase the quality of life of its parishioners, A.J. Gill supplies evidence that the underlying reason may be competition from evangelical Protestant denominations. He compiled the data given in this data frame.
ex2017ex2017
A data frame with 12 observations on the following 5 variables.
Catholic church stance, factor with levels "Pro"
and "Anti"
Latin American country
Physical Quality of Life Index in the mid-1970s; Average of live expectancy at age 1, infant mortality and literacy at age 15+.
Average civil rights score for the period of authoritarian rule until 1979
Percentage increase of competitive religious groups during the period 1900–1970
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Gill, A.J. (1994). Rendering unto Caesar? Religious Competition and Catholic Strategy in Latin America, 1962–1979, American Journal of Political Science 38(2): 403–425.
str(ex2017)str(ex2017)
This data frame contains data on 1995 and later model compact sports utility vehicles involved in fatal accidents in the United States between 1995 and 1999, excluding those that were struck by another car and excluding accidents that, according to police reports, involved alcohol.
ex2018ex2018
A data frame with 2,321 observations on the following 4 variables.
Type of sports utility vehicle, factor with levels
"Other" and "Ford"
Vehicle age (in years); surrogate for age of tires
Number of passengers
Cause of fatal accident, factor with levels
"NotTire" and "Tire"
The Ford Explorer is a popular sports utility vehicle made in the United States and sold throughout the world. Early in its production concern arose over a potential accident risk associated with tires of the prescribed size when the vehicle was carrying heavy loads, but the risk was thought to be acceptable if a low tire pressure was recommended. The problem was apparently exacerbated by a particular type of Firestone tire that was overly prone to separation, especially in warm temperatures. This type of tire was a common one used on Explorers in model years 1995 and later. By the end of 1999 more than 30 lawsuits had been filed over accidents that were thought to be associated with this problem. U.S. federal data on fatal car accidents were analysed at that time, showing that the odds of a fatal accident being associated with tire failure were three times as great for Explorers as for other sports utility vehicles.
Additional data from 1999 and additional variables may be used to further explore the odds ratio. It is of interest to see whether the odds that a fatal accident is tire-related depend on whether the vehicle is a Ford, after accounting for age of the car and number of passengers. Since the Ford tire problem may be due to the load carried, there is some interest in seeing whether the odds associated with a Ford depend on the number of passengers.
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex2018)str(ex2018)
Following a successful test of an anti-ballistic missile (ABM) in July 1999, many prominent U.S. politicians called for the early deployment of a full ABM system. The scientific community was less enthusiastic about the efficacy of such a system. This data set contains the success or failure of ABM tests between March 1983 and May 1995. Do these data suggest any improvement in ABM test success probability over time?
ex2019ex2019
A data frame with 17 observations on the following 3 variables.
date of an ABM test
number of months after March 1983
a factor with levels "Failure" and
"Success"
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Lewis, G. N., Postol, T. A. and Pike, J. (1999) Why National Missile Defense Won't Work, Scientific American 281(2): 36–41.
str(ex2019)str(ex2019)
These data are from a randomized experiment to asses the effect of large doses of vitamin C on the incidence of colds.
ex2113ex2113
A data frame with 4 observations on the following 4 variables.
the daily dose of vitamin C, in g
the number of subjects given that dose of vitamin C
the number of subjects who did not become ill
the proportion of subjects who did not become ill
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Anderson, T.W., Suranyi,G., and Beaton, G.H. (1974) The Effect on Winter Illness of Large Doses of Vitamin C Canadian Medical Association Journal 111 31–36.
str(ex2113)str(ex2113)
The study the effect of context questions prior to target questions, researchers conducted a poll involving 1,054 subjects selected randomly from the Chicago phone directory. To include possibly unlisted phones, selected numbers were randomly altered in the last position. This data frame contains the responses to one of the questions asked concerning continuing U.S. aid to the Nicaraguan Contra rebels. Eight different versions of the interview were given, representing all possible combinations of three factors at each of two levels. The experimental factors were Context, Mode and Level.
Context refers to the type of context questions preceding the
question about Nicaraguan aid. Some subjects received a context
question about Vietnam, designed to elicit reticence about having the
U.S. become involved in another foreign war in a third–world country.
The other context question was about Cuba, designed to elicit
anti–communist sentiments.
Mode refers to whether the target question immediately followed
the context question or whether there were other questions scattered
in between.
Level refers to two versions of the context question. In the
"high" level the question was worded to elicit a higher level of
agreement than in the "low" level wording.
ex2115ex2115
A data frame with 8 observations on the following 7 variables.
Factor referring to the context of the question preceding the target question about U.S. aid to the Nicaraguan Contra rebels
Factor with levels "not" and
"scattered", "scattered" is used if the target question
was not asked directly after the context question
Factor with levels "low" and
"high", refers to the wording of the question
Number of people interviewed
Number of people in favor of Contra Aid
Number of people not in favor of Contra Aid
Percentage in favour of Contra aid
Increasingly, politicians look to public opinion surveys to shape their public stances. Does this represent the ultimate in democracy? Or are seemingly scientific polls being rigged by the manner of questioning? Psychologists believe that opinions—expressed as answers to questions—are usually generated at the time the question is asked. Answers are based on a quick sampling of relevant beliefs held by the subject, rather than a systematic canvas of all such beliefs. Furthermore, this sampling of beliefs tends to overrepresent whatever beliefs happen to be most accessible at the time the question is asked. This aspect of delivering opinions can be abused by the pollster. Here, for example, is one sequence of questions:
“Do you believe the Bill of Rights protects personal freedom?”
“Are you in favor of a ban on handguns?”
Here is another:
“Do you think something should be done to reduce violent crime?”
“Are you in favor of a ban on handguns?”
The proportion of yes answers to question 2 may be quite different depending on which question 1 is asked first.
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Tourangeau, R., Rasinski, K.A., Bradburn, N. and D'Andrade, R. (1989). Belief Accessibility and Context Effects in Attitude Measurement, Journal of Experimental Social Psychology 25: 401–421.
str(ex2115)str(ex2115)
An experiment at the Marine/Freshwater Biomedical Sciences Center at Oregon State University investigated the carcinogenic effects of aflatoxicol, a metabolite of Aflatoxin B1, which is a toxic by-product produced by a mold that infects cottonseed meal, peanuts and grains. Twenty tanks of rainbow trout embryos were exposed to one of five doses of Aflatoxicol for one hour. The data represent the numbers of fish in each tank and the numbers of these that had liver tumours after one year.
ex2116ex2116
A data frame with 20 observations on the following 3 variables.
Dose (in ppm)
Number of trout with liver tumours
Number of trout in tank
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex2116)str(ex2116)
The probability of a male birth in humans is about .51. It has previously been noticed that lower proportions of male births are observed when offspring is conceived at times of exposure to smog, floods or earthquakes. Danish researchers hypothesised that sources of stress associated with severe life events may also have some bearing on the sex ratio. To investigate this theory they obtained the sexes of all 3,072 children who were born in Denmark between 1 January 1980 and 31 December 1992 to women who experienced the following kind of severe life events in the year of the birth or the year prior to the birth: death or admission to hospital for cancer or heart attack of their partner or of their other children. They also obtained sexes on a sample of 20,337 births to mothers who did not experience these life stress episodes. This data frame contains the data that were collected. Noticed that for one group the exposure is listed as taking place during the first trimester of pregnancy. The rationale for this is that the stress associated with the cancer or heart attack of a family member may well have started before the recorded time of death or hospital admission.
ex2117ex2117
A data frame with 5 observations on the following 4 variables.
Indicator for groups to which mothers belong
Indicator for time at which severe life event occurred
Number of births
Percentage of boys born
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Hansen, D., Møller, H. and Olsen, J. (1999). Severe Periconceptional Life Events and the Sex Ratio in Offspring: Follow Up Study based on Five National Registers, British Medical Journal 319(7209): 548–549.
str(ex2117)str(ex2117)
Researchers in Kenya identified a cohort of more that 1,000 prostitutes who were known to be a major reservoir of sexually transmitted diseases in 1985. It was determined that more than 85% of them were infected with human immunodeficiency virus (HIV) in February, 1986. The researchers identified men who acquired a sexually-transmitted disease from this group of women after the men sought treatment at a free clinic. The data frame contains data on the subset of those men who did not test positive for HIV on their first visit and who agreed to participate in the study. The men are categorised according to whether they later tested positive for HIV during the study period, whether they had one or multiple sexual contacts with the prostitutes and whether they were circumcised.
ex2118ex2118
A data frame with 4 observations on the following 5 variables.
Whether men had single or multiple contact with prostitutes
Whether the men are circumcised, factor with
levels "No" and "Yes"
Number of men that tested positive for HIV
Number of men
Number of men that did not test positive for HIV (should be Number-HIV)
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Cameron, D.W., D'Costa, L.J., Maitha, G.M., Cheang, M., Piot, P., Simonsen, J.N., Ronald, A.R., Gakinya, M.N., Ndinya-Achola, J.O., Brunham, R.C. and Plummer, F. A. (1989). Female to Male Transmission of Human Immunodeficiency Virus Type I: Risk Factors for Seroconversion in Men, The Lancet 334(8660): 403–407.
str(ex2118)str(ex2118)
This data frame gives the results of 10 separate case–control studies on the association of breast cancer and whether a woman had breast–fed children.
ex2119ex2119
A data frame with 20 observations on the following 4 variables.
Factor indicating the study from which data was taken
Whether women had breast–fed children (lactated)
Number of women with breast cancer
Number of women without breast cancer
Meta–analysis refers to the analysis of analyses. When the main
results of studies can be cast into 22 tables of counts,
it is natural to combine individual odds ratios with a logistic
regression model that includes a factor to account for different odds
from the different studies. In addition, the odds ratio itself might
differ slightly among studies because of different effects on
different populations or different research techniques. One approach
for dealing with this is to suppose an underlying common odds ratio
and to model between–study variability as extra–binomial variation.
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Data gathered from various sources by Karolyn Kolassa as part of a Master's project, Oregon State University.
str(ex2119)str(ex2119)
These data were simulated to match the summary statistics and conclusions of Rosenthal and Fode's Clever Hans experiment. Each of 12 students trained rats to run a maze. The data set contains their number of successful runs out of 50 on each of 5 days, the student's prior expectation of success (on a scale from -10 to 10), and a variable indicating treatment–whether or not the students were supplied with the fictitious information that their rights were bright.
21202120
A data frame with 60 observations on the following 5 variables.
a student identification number
the student's prior expectation of rat-training success, on a scale from -10 to 10
a factor with levels "bright" and "dull"
corresponding to whether students were told (falsely) that their rats were
bright or not
day of the study, ranging from 1 to 5
the number of successful maze runs on a day, out of 50
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Rosenthal, R. and Fode, K.L. (1963) The Effect of Experimenter Bias on the Performance of the Albino Rat Behavioral Science 8:3: 183–189.
str(ex2120)str(ex2120)
Some sociologist suspect that highly publicised suicides may trigger additional suicides. In one investigation of this hypothesis, D.P. Phillips collected information about 17 airplane crashes that were known (because of notes left behind) to be murder–suicides. For each of these crashes, Phillips reported an index of the news coverage (circulation of nine newspapers devoting space to the crash multiplied by length of coverage) and the number of multiple-fatality plane crashes during the week following the publicised crash. This data frame contains the collected data.
ex2216ex2216
A data frame with 17 observations on the following 2 variables.
Index for the amount of newspaper coverage given the murder–suicide
Multiple-fatality crashes in the week following a murder–suicide by plane crash
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Phillips, D.P. (1978). Airplane Accident Fatalities Increase Just After Newspaper Stories About Murder and Suicide, Science 201: 748–750.
str(ex2216)str(ex2216)
The data are the number of cancer deaths among survivors of the atomic bombs dropped on Japan during World War II, categorized by time (years) after the bomb that death occurred and the amount of radiation exposure that the survivors received from the blast. Also listed in each cell is the person- years at risk, in 100's. This is the sum total of all years spent by all persons in the category. The data can be analyzed by supposing the number of cancer deaths in each cell is Poisson with mean = risk x rate, where risk is the person-years at risk and rate is the rate of cancer deaths per person per year. How does the rate depend on the radiation exposure, after accounting for years after exposure?
ex2220ex2220
A data frame with 42 observations on the following 4 variables.
radiation exposure, in rads
years after the exposure
number of survivors in the group
number of survivors in the group who died of Cancer
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Pierce, D.A., personal communication
str(ex2220)str(ex2220)
When she died in 1817, the English novelist Jane Austen had not yet finished the novel Sanditon, but she did leave notes on how she intended to conclude the book. The novel was completed by a ghost writer, who attempted to emulate Austen's style. In 1978, a researcher reported counts of some words found in chapters of books written by Austen and in chapters written by the emulator. These data are given in this data frame.
ex2222ex2222
A data frame with 24 observations on the following 3 variables.
Number of occurrences of a word in various chapters of books written by Jane Austen and the ghost writer
Title of books used
Words used
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Morton, A.Q. (1978). Literary Detection: How to Prove Authorship and Fraud in Literature and Documents, Charles Scribner's Sons, New York.
str(ex2222)str(ex2222)
On January 27, 1986, the night before the space shuttle Challenger exploded, an engineer recommended to the National Aeronautics and Space Administration (NASA) that the shuttle not be launched in the cold weather. The forecasted temperature for the Challenger launch was 31 degrees Fahrenheit—the coldest launch ever. After an intense 3-hour telephone conference, officials decided to proceed with the launch. This data frame contains the launch temperatures and the number of O-ring problems in 24 shuttle launches prior to the Challenger.
ex2223ex2223
A data frame with 24 observations on the following 2 variables.
Launch temperatures (in degrees Fahrenheit)
Numbers of O-ring incidents
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
str(ex2223)str(ex2223)
This data frame contains data on characteristics and numbers of failures observed in valve types from one pressurised water reactor.
ex2224ex2224
A data frame with 90 observations on the following 7 variables.
a numerical code corresponding to 5 categories of system (1 = containment, 2 = nuclear, 3 = power conversion, 4 = safety, 5 = process auxiliary)
a numerical code corresponding to 4 different operator types (1 = air, 2 = solenoid, 3 = motor=driven, 4 = manual)
a numerical code corresponding to 6 different valve types (1 = ball, 2 = butterfly, 3 = diaphram, 4 = gate, 5 = globe, 6 = directional control)
a numerical code corresponding to 3 head size categories (1 = less than 2 inches, 2 = 2–10 inches, 3 = 10–30 inches)
a numerical code corresponding to two categories of operation mode (1 = normally closed, 2 = normally open)
Number of failures observed
Lengths of observation time
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Moore, L.M. and Beckman, R.J. (1988). Appropriate One-Sided Tolerance Bounds on the Number of Failures using Poisson Regression, Technometrics 30: 283–290.
str(ex2224)str(ex2224)
As an example of field observation in evidence of theories of sexual selection, S.J. Arnold and M.J. Wade presented the following data set on size and number of mates observed in 38 bullfrogs.
ex2225ex2225
A data frame with 38 observations on the following 2 variables.
Body size (in mm)
Number of mates
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Arnold, S.J. and Wade, M.J. (1984). On the Measurement of Natural and Sexual Selection: Aplications, Evolution 38: 720–734.
str(ex2225)str(ex2225)
Apparently, larger planets have more moons, but is it the volume (as indicated by diameter) or mass that are more relevant, or is it both? These data include the diameter, mass, distance from the sun, and number of moons for 13 planets, gas giants, and dwarf planets in our solar system. Which size variable best explains mean number of moons (possible after accounting for distance from sun). (Consider negative binomial regression.)
ex2226ex2226
A data frame with 13 observations on the following 5 variables.
a character variable with the name of the planet, gas giant, or dwarf planet)
distance from sun, relative to earth's
diameter of the planet, relative to earth's
mass, relative to earth's
number of moons
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Wikipedia: https://en.wikipedia.org/wiki/Planet August 10, 2011
str(ex2226)str(ex2226)
Data frame contains the percentage of unsuccessful hatching from enclosures containing 150 eggs each in a study to investigate whether UV-B is responsible for low hatch rates.
ex2414ex2414
A data frame with 71 observations on the following 4 variables.
percentage of frog eggs failing to hatch
factor variable with levels "NoFilter",
"UV-BTransmitting" and "UV-BBlocking"
factor variable with levels
"ThreeCreeks", "SparksLake", "SmallLake" and
"LostLake"
Photolyase activity
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
Blaustein, A.R., Hoffman, P.D., Hokit, D.G., Kiesecker, J.M., Walls, S.C. and Hays, J.B. (1994). UV Repair and Resistance to Solar UV-B in Amphibian Eggs: A Link to Population Declines? Proceedings of the National Academy of Science, USA 91: 1791–1795.
str(ex2414)str(ex2414)
If the option “pdfviewer” is set, this command will display the PDF version of the help pages.
Sleuth3Manual()Sleuth3Manual()
Berwin A Turlach [email protected]
Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cengage Learning.
## Not run: Sleuth3Manual()## Not run: Sleuth3Manual()