[Sleuth2-commits] r37 - in pkg/Sleuth3: . inst/doc man
noreply at r-forge.r-project.org
noreply at r-forge.r-project.org
Fri Aug 3 18:26:35 CEST 2012
Author: berwin
Date: 2012-08-03 18:26:35 +0200 (Fri, 03 Aug 2012)
New Revision: 37
Added:
pkg/Sleuth3/man/case0601.Rd
pkg/Sleuth3/man/case0602.Rd
pkg/Sleuth3/man/case0701.Rd
pkg/Sleuth3/man/case0702.Rd
pkg/Sleuth3/man/case0801.Rd
pkg/Sleuth3/man/case0802.Rd
pkg/Sleuth3/man/case0901.Rd
pkg/Sleuth3/man/case0902.Rd
pkg/Sleuth3/man/case1001.Rd
pkg/Sleuth3/man/case1002.Rd
pkg/Sleuth3/man/case1101.Rd
pkg/Sleuth3/man/case1102.Rd
pkg/Sleuth3/man/case1201.Rd
pkg/Sleuth3/man/case1202.Rd
pkg/Sleuth3/man/case1301.Rd
pkg/Sleuth3/man/case1302.Rd
pkg/Sleuth3/man/case1401.Rd
pkg/Sleuth3/man/case1402.Rd
Modified:
pkg/Sleuth3/DESCRIPTION
pkg/Sleuth3/inst/doc/Sleuth3-manual.pdf
pkg/Sleuth3/man/Sleuth3-package.Rd
pkg/Sleuth3/man/Sleuth3Manual.Rd
pkg/Sleuth3/man/case0101.Rd
pkg/Sleuth3/man/case0102.Rd
pkg/Sleuth3/man/case0201.Rd
pkg/Sleuth3/man/case0202.Rd
pkg/Sleuth3/man/case0301.Rd
pkg/Sleuth3/man/case0302.Rd
pkg/Sleuth3/man/case0401.Rd
pkg/Sleuth3/man/case0402.Rd
pkg/Sleuth3/man/case0501.Rd
pkg/Sleuth3/man/case0502.Rd
Log:
Sleuth3:
Replaced .Rd files with new versions provided by Dan.
Added new .Rd files provided by Dan.
Updated publication date in DESCRIPTION file.
Bumped version number, hence also updated inst/doc/Sleuth3-manual.pdf
Modified: pkg/Sleuth3/DESCRIPTION
===================================================================
--- pkg/Sleuth3/DESCRIPTION 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/DESCRIPTION 2012-08-03 16:26:35 UTC (rev 37)
@@ -1,15 +1,15 @@
Package: Sleuth3
Title: Data sets from Ramsey and Schafer's "Statistical Sleuth (3rd ed)"
-Version: 0.0-4
-Date: 2012-06-24
+Version: 0.0-5
+Date: 2012-08-03
Author: Original by F.L. Ramsey and D.W. Schafer,
modifications by Daniel W. Schafer, Jeannie Sifneos and Berwin A. Turlach
-Description: Data sets from Ramsey, F.L. and Schafer, D.W. (2012), "The
+Description: Data sets from Ramsey, F.L. and Schafer, D.W. (2013), "The
Statistical Sleuth: A Course in Methods of Data Analysis (3rd
- ed)", Duxbury.
+ ed)", Cengage Learning.
Maintainer: Berwin A Turlach <Berwin.Turlach at gmail.com>
LazyData: yes
Depends: R (>= 2.12.0)
-Suggests: lattice, multcomp
+Suggests: lattice, multcomp, car, leaps
License: GPL (>= 2)
URL: http://r-forge.r-project.org/projects/sleuth2/
Modified: pkg/Sleuth3/inst/doc/Sleuth3-manual.pdf
===================================================================
(Binary files differ)
Modified: pkg/Sleuth3/man/Sleuth3-package.Rd
===================================================================
--- pkg/Sleuth3/man/Sleuth3-package.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/Sleuth3-package.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -6,7 +6,7 @@
The R Sleuth3 package
}
\description{
-Data sets from Ramsey and Schafer's "Statistical Sleuth (2nd ed)"
+Data sets from Ramsey and Schafer's "Statistical Sleuth (3rd ed)"
}
\details{This package contains a variety of datasets. For a complete
list, use \code{library(help="Sleuth3")} or
Modified: pkg/Sleuth3/man/Sleuth3Manual.Rd
===================================================================
--- pkg/Sleuth3/man/Sleuth3Manual.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/Sleuth3Manual.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -11,8 +11,8 @@
Sleuth3Manual()
}
\references{
- Ramsey, F.L. and Schafer, D.W. (2002). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\author{
Berwin A Turlach \email{Berwin.Turlach at gmail.com}
Modified: pkg/Sleuth3/man/case0101.Rd
===================================================================
--- pkg/Sleuth3/man/case0101.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0101.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -18,8 +18,8 @@
}
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
Amabile, T. (1985). Motivation and Creativity: Effects of Motivational
@@ -29,9 +29,9 @@
\examples{
attach(case0101)
str(case0101)
-boxplot(Score ~ Treatment)
+boxplot(Score ~ Treatment) # Basic boxplots for each level of Treatment
-boxplot(Score ~ 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")
Modified: pkg/Sleuth3/man/case0102.Rd
===================================================================
--- pkg/Sleuth3/man/case0102.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0102.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -17,8 +17,8 @@
}
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
Roberts, H.V. (1979). Harris Trust and Savings Bank: An Analysis of
@@ -31,7 +31,7 @@
}
\examples{
attach(case0102)
-str(case0102)
+str(case0102)
boxplot(Salary ~ Sex,
ylab= "Starting Salary (U.S. Dollars)",
Modified: pkg/Sleuth3/man/case0201.Rd
===================================================================
--- pkg/Sleuth3/man/case0201.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0201.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -26,12 +26,12 @@
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
- Grant, P. (1986). \bold{Ecology and Evolution of Darwin's Finches},
+ Grant, P. (1986). \emph{Ecology and Evolution of Darwin's Finches},
Princeton University Press, Princeton, N.J.
}
@@ -45,9 +45,11 @@
mean(Depth[Year==1978]) - mean(Depth[Year==1976])
-yearFactor <- factor(Year)
-t.test(Depth ~ yearFactor, var.equal=TRUE)
-t.test(Depth ~ yearFactor, var.equal=TRUE, alternative = "less")
+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)",
@@ -56,20 +58,10 @@
## 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)
+ 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)
}
Modified: pkg/Sleuth3/man/case0202.Rd
===================================================================
--- pkg/Sleuth3/man/case0202.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0202.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -8,7 +8,7 @@
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$^3$) of several regions and
+ imaging to measure the volumes (in cm\eqn{^3}) of several regions and
subregions of the twins' brains.
}
\usage{case0202}
@@ -22,8 +22,8 @@
}
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
Suddath, R.L., Christison, G.W., Torrey, E.F., Casanova, M.F. and
@@ -38,30 +38,22 @@
diff <- Unaffected-Affected
summary(diff)
t.test(diff) # Paired t-test is a one-sample t-test on differnces
-t.test(Unaffected,Affected,pair=TRUE) # same thing
+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")
+ 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)
+ 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)
Modified: pkg/Sleuth3/man/case0301.Rd
===================================================================
--- pkg/Sleuth3/man/case0301.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0301.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -20,8 +20,8 @@
}
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
Simpson, J., Olsen, A., and Eden, J. (1975). A Bayesian Analysis of a
@@ -30,15 +30,17 @@
}
\examples{
attach(case0301)
-str(case0301)
+str(case0301) #Seeded is level 1 of Treatment (it's first alphabetically)
boxplot(Rainfall ~ Treatment)
-boxplot(log(Rainfall) ~ Treatment)
+boxplot(log(Rainfall) ~ Treatment) # Boxplots of natural logs of Rainfall
-t.test(log(Rainfall) ~ Treatment, alternative="greater", var.equal=TRUE)
+t.test(log(Rainfall) ~ Treatment, var.equal=TRUE,
+ alternative="greater") # 1-sided t-test; alternative: level 1 mean is greater
-# Use 2-sided alternative to get a 2-sided confidence interval:
-myTest <- t.test(log(Rainfall) ~ Treatment, alternative="two.sided", var.equal=TRUE)
+
+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
@@ -48,25 +50,17 @@
main="Distributions of Rainfalls from Cloud Seeding Experiment")
## POLISHED BOXPLOTS FOR PRESENTATION:
-par(mar=c(4,4,4,4))
-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) # Create some tick marks for original scale
+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)
}
Modified: pkg/Sleuth3/man/case0302.Rd
===================================================================
--- pkg/Sleuth3/man/case0302.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0302.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -19,8 +19,8 @@
}
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
Centers for Disease Control Veterans Health Studies: Serum
@@ -30,27 +30,35 @@
}
\examples{
attach(case0302)
-str(case0302)
+str(case0302) # Note: Level 1 of Veteran is "Other" (first alphabeticall)
+
boxplot(Dioxin ~ Veteran)
-t.test(Dioxin ~ Veteran, alternative="less", var.equal=TRUE)
-t.test(Dioxin ~ Veteran, alternative="less", var.equal=TRUE, subset=(Dioxin < 40))
-t.test(Dioxin ~ Veteran, alternative="less", var.equal=TRUE, subset=(Dioxin < 20))
-t.test(Dioxin ~ Veteran, var.equal=TRUE)
+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
-par(mfrow=c(2,1), mar=c(3,3,1,1))
+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 pct CI: 3.72 to 4.65 ppt)",
- pos=4, cex=.75)
+ 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 pct CI: 4.06 to 4.64 ppt)",
- pos=4, cex=.75)
-text(13,145,"[Estimated Difference in Means: 0.07 ppt (95 pct CI: -0.63 to 0.48 ppt])",
- pos=4, cex=.75)
+ 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)
}
Modified: pkg/Sleuth3/man/case0401.Rd
===================================================================
--- pkg/Sleuth3/man/case0401.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0401.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -18,8 +18,8 @@
}
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
Feynman, R.P. (1988). \emph{What do You Care What Other People
@@ -35,29 +35,26 @@
mCool <- mean(Incidents[Launch=="Cool"])
mWarm <- mean(Incidents[Launch=="Warm"])
mDiff <- mCool - mWarm
-c(mCool,mWarm,mDiff)
+c(mCool,mWarm,mDiff) # Show the values of these variables
-## PERMUTATION TEST BASED ON DIFFERENCE IN AVERAGES, VIA REPEATED RANDOM RE-GROUPING (ADVANCED)
-numRep <- 50 # Number of random re-groupings.
- # NOTE: this should be changed to something like 50,000 for useful results.
-rDiff <- rep(0,numRep) # A place to store the differences in averages from the re-groupings
-for (rep in 1:numRep) { # Repeat the commands following numRep times
- randomGroup <- rep("rWarm",24) # Initially set this to 24 values "rWarm"
- randomGroup[sample(1:24,4)] <- "rCool" # Replace 4 and random with "rCool"
- mW <- mean(Incidents[randomGroup=="rWarm"])
- mC <- mean(Incidents[randomGroup=="rCool"])
- rDiff[rep] <- mC-mW # Store the difference in averages
- }
-hist(rDiff, # Draw a histogram of the difference in averages from numRep random re-groupings
+## 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) # Make a vertical line at the actually observed difference
-pValue <- sum(rDiff >= 1.3)/numRep # 1-sided p-value = proportion of re-groupings with diff >= observed
+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)
+text(mDiff,75000, paste(" -->",round(pValue,4)), adj=-0.1)
detach(case0401)
}
Modified: pkg/Sleuth3/man/case0402.Rd
===================================================================
--- pkg/Sleuth3/man/case0402.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0402.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -22,8 +22,8 @@
}
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
Sweller, J., Chandler, P., Tierney, P. and Cooper, M. (1990).
@@ -32,24 +32,27 @@
176--192.
}
\examples{
-str(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")
-wilcox.test(Time ~ Treatment, exact=FALSE, correct=TRUE, alternative="two.sided", conf.int=TRUE)
+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)
+xTreatment <- ifelse(Treatment=="Conventional",1,2) # Make numerical values
myPointCode <- ifelse(Censored==0,21,24)
-plot(Time ~ jitter(xTreatment,.2),
- ylab="Completion Time (Sec.)", xlab="Training Method (jittered for visibility)",
+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)
-axis(1, tick=FALSE, at=c(1,2), labels=c("Conventional (n=14 Students)","Modified (n=14 Students)") )
+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")
Modified: pkg/Sleuth3/man/case0501.Rd
===================================================================
--- pkg/Sleuth3/man/case0501.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0501.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -33,8 +33,8 @@
}
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
Weindruch, R., Walford, R.L., Fligiel, S. and Guthrie D. (1986). The
@@ -46,15 +46,19 @@
str(case0501)
attach(case0501)
-myDiet <- factor(Diet, levels=c("NP","N/N85","N/R50","R/R50","lopro","N/R40") ) # Re-order levels for better boxplot ordering
-myNames <- c("NP(49)","N/N85(57)","N/R50(71)","R/R50(56)","lopro(56)","N/R40(60)")
-boxplot(Lifetime ~ myDiet, ylab= "Lifetime (months)", names=myNames, xlab="Treatment (and sample size)")
-myAov1 <- aov(Lifetime ~ Diet)
-plot(myAov1, which=1)
+# 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")
+pairwise.t.test(Lifetime,Diet, pool.SD=TRUE, p.adj="none") # All t-tests
-## p-VALUES AND CONFIDENCE INTERVALS FOR SPECIFIED COMPARISONS OF MEANS, USING GLHT IN MULTCOMP PACKAGE
+## p-VALUES AND CONFIDENCE INTERVALS FOR SPECIFIED COMPARISONS OF MEANS
if(require(multcomp)){
diet <- factor(Diet,labels=c("lopro", "NN85", "NR40", "NR50", "NP", "RR50"))
myAov2 <- aov(Lifetime ~ diet - 1)
@@ -64,16 +68,16 @@
"dietRR50 - dietNR50 = 0",
"dietlopro - dietNR50 = 0",
"dietNN85 - dietNP = 0") )
- summary(myComparisons,test=adjusted("none")) # No multiple comparison adjustment (for these planned comparisons)
- confint(myComparisons, calpha = univariate_calpha()) # No multiple comparison adjustment
-}
+ 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)
+ 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)
Modified: pkg/Sleuth3/man/case0502.Rd
===================================================================
--- pkg/Sleuth3/man/case0502.Rd 2012-07-24 16:20:24 UTC (rev 36)
+++ pkg/Sleuth3/man/case0502.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -24,13 +24,13 @@
\describe{
\item{\code{Percent}}{is the percent of women on the venire's of the Spock
trial judge and 6 other Boston area judges}
- \item{\code{Judge}}{a factor with levels \code{"Spock's"}, \code{"A"},
+ \item{\code{Judge}}{is a factor with levels \code{"Spock's"}, \code{"A"},
\code{"B"}, \code{"C"}, \code{"D"}, \code{"E"} and \code{"F"}}
}
}
\source{
- Ramsey, F.L. and Schafer, D.W. (2012). \emph{The Statistical Sleuth: A
- Course in Methods of Data Analysis (3rd ed)}, Duxbury.
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
}
\references{
Zeisel, H. and Kalven, H. Jr. (1972). Parking Tickets and Missing
@@ -41,27 +41,29 @@
str(case0502)
attach(case0502)
-myNames <- c("A (5)", "B (6)", "C (9)", "D (2)", "E (6)", "F (9)", "Spock's (9)" ) # Factor level names (with sample sizes) for boxplots
+# 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")
+ 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--is there evidence that any means differ from any others? (Yes)
+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")) # Indicator for Spock's judge
+SpockOrOther <- factor(ifelse(Judge=="Spock's","Spock","Other"))
aovFull <- aov(Percent ~ Judge)
aovReduced <- aov(Percent ~ SpockOrOther)
-anova(aovReduced,aovFull) # Is there evidence that the 7 mean model fits better than the 2 mean model? (No)
-t.test(Percent ~ SpockOrOther, var.equal=TRUE) # Is there evidence in the 2 mean model that the means differ? (Yes)
+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 (NOT ASSUMING NON-SPOCK JUDGES HAVE A COMMON MEAN)
-if(require(multcomp)){
+## 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))
+ 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")) # Is there evidence that Spock mean differs from average of other 6? (Yes)
+ summary(myComparison, test=adjusted("none"))
confint(myComparison)
}
Added: pkg/Sleuth3/man/case0601.Rd
===================================================================
--- pkg/Sleuth3/man/case0601.Rd (rev 0)
+++ pkg/Sleuth3/man/case0601.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -0,0 +1,87 @@
+\name{case0601}
+\alias{case0601}
+\docType{data}
+\title{Discrimination Against the Handicapped}
+\description{
+ 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.
+}
+\usage{case0601}
+\format{
+ A data frame with 70 observations on the following 2 variables.
+ \describe{
+ \item{\code{Score}}{is the score each student gave to the applicant}
+ \item{\code{Handicap}}{is a factor variable with 5
+ levels---\code{"None"}, \code{"Amputee"}, \code{"Crutches"},
+ \code{"Hearing"} and \code{"Wheelchair"}}
+ }
+}
+\source{
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
+}
+\references{
+ Cesare, S.J., Tannenbaum, R.J. and Dalessio, A. (1990). Interviewers'
+ Decisions Related to Applicant Handicap Type and Rater Empathy, \emph{Human
+ Performance} \bold{3}(3): 157--171.
+}
+\examples{
+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)
+}
+\keyword{datasets}
Property changes on: pkg/Sleuth3/man/case0601.Rd
___________________________________________________________________
Added: svn:mime-type
+ text/plain
Added: svn:eol-style
+ native
Added: pkg/Sleuth3/man/case0602.Rd
===================================================================
--- pkg/Sleuth3/man/case0602.Rd (rev 0)
+++ pkg/Sleuth3/man/case0602.Rd 2012-08-03 16:26:35 UTC (rev 37)
@@ -0,0 +1,79 @@
+\name{case0602}
+\alias{case0602}
+\docType{data}
+\title{Mate Preference of Platyfish}
+\description{
+ 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.
+}
+\usage{case0602}
+\format{
+ A data frame with 84 observations on the following 3 variables.
+ \describe{
+ \item{\code{Proportion}}{The proportion of courtship time spent by 84
+ females with the yellow-sword males}
+ \item{\code{Pair}}{Factor variable with 6 levels---\code{"Pair1"},
+ \code{"Pair2"}, \code{"Pair3"}, \code{"Pair4"}, \code{"Pair5"} and
+ \code{"Pair6"}}
+ \item{\code{Length}}{Body size of the males}
+ }
+}
+\source{
+ Ramsey, F.L. and Schafer, D.W. (2013). \emph{The Statistical Sleuth: A
+ Course in Methods of Data Analysis (3rd ed)}, Cengage Learning.
+}
+\references{
+ Basolo, A.L. (1990). Female Preference Predates the Evolution of the
+ Sword in Swordtail Fish, \emph{Science} \bold{250}: 808--810.
+}
+\examples{
+str(case0602)
+attach(case0602)
+
+## EXPLORATION
+plot(Proportion ~ Length,
+ xlab="Length of the Two Males",
+ ylab="Proportion of Time Female Spent with Yellow-Sword Male",
+ main="Proportion of Time Spent with Yellow Rather than Transparent Sword Male")
[TRUNCATED]
To get the complete diff run:
svnlook diff /svnroot/sleuth2 -r 37
More information about the Sleuth2-commits
mailing list