305.week6 Working with real data
William Revelle
2022-05-02
First some prelimaries
The ‘psych’ and ‘psychTools’ packages have been updated to versions 2.2.5. You need to install them and then restart RStudio.
After doing so,
library(psych)
library(psychTools)
sessionInfo() #to make sure you have the most recent release## R version 4.2.0 (2022-04-22)
## Platform: aarch64-apple-darwin20 (64-bit)
## Running under: macOS Monterey 12.3.1
##
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/4.2-arm64/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/4.2-arm64/Resources/lib/libRlapack.dylib
##
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] psychTools_2.2.5 psych_2.2.5
##
## loaded via a namespace (and not attached):
## [1] lattice_0.20-45 digest_0.6.28 grid_4.2.0 R6_2.5.1 nlme_3.1-157 jsonlite_1.7.2
## [7] magrittr_2.0.1 evaluate_0.14 rlang_0.4.12 stringi_1.7.5 jquerylib_0.1.4 bslib_0.3.1
## [13] rmarkdown_2.11 tools_4.2.0 foreign_0.8-82 stringr_1.4.0 parallel_4.2.0 xfun_0.28
## [19] yaml_2.3.5 fastmap_1.1.0 compiler_4.2.0 mnormt_2.0.2 tmvnsim_1.0-2 htmltools_0.5.2
## [25] knitr_1.36 sass_0.4.0Alice Eagly and I have been working on a manuscript examining the effect of data aggregation on the validity of scales. The paper (https://personality-project.org/revelle/publications/eagly.revelle.21.pdf ) Understanding the Magnitude of Psychological Differences Between Women and Men Requires Seeing the Forest and the Trees (in press)) uses data from an article by Ursala Athenstaedt (2003). Here we show the steps necessary to convert a raw data set into the final, publication ready graphics that we have in our paper.
In the process of final proof reading, we discovered some errors in the way we had processed the data. This is a nice example of the to document your scripts and to double check (and then check again) your analyses.
Getting the original data
The data are in an SPSS sav file that I have put into the course files. We need to get the name of that file, and then read it.
fn <- "https://personality-project.org/courses/350/3studien.sav"
ursala <- read.file(fn)## Data from the SPSS sav file https://personality-project.org/courses/350/3studien.sav has been loaded.dim(ursala) #how many subjects (rows) and variables (columns)## [1] 677 114colnames(ursala) ## [1] "STUDIE" "GESCHL" "V1" "V2" "V3" "V4" "V5" "V6" "V7"
## [10] "V8" "V9" "V10" "V11" "V12" "V13" "V14" "V15" "V16"
## [19] "V17" "V18" "V19" "V20" "V21" "V22" "V23" "V24" "V25"
## [28] "V26" "V27" "V28" "V29" "V30" "V31" "V32" "V33" "V34"
## [37] "V35" "V36" "V37" "V38" "V39" "V40" "V41" "V42" "V43"
## [46] "V44" "V45" "V46" "V47" "V48" "V49" "V50" "V51" "V52"
## [55] "V53" "V54" "V55" "V56" "V57" "V58" "V59" "V60" "V61"
## [64] "V62" "V63" "V64" "V65" "V66" "V67" "V68" "V69" "V70"
## [73] "V71" "V72" "V73" "V74" "MMINUS1" "FMINUS1" "MPLUS1" "FPLUS1" "MMINUS2"
## [82] "MPLUS2" "MMINUS3" "FPLUS2" "FPLUS3" "FMINUS2" "FPLUS4" "FMINUS3" "MMINUS4" "FPLUS5"
## [91] "MMINUS5" "FMINUS4" "FPLUS6" "MPLUS3" "FMINUS5" "MPLUS4" "MMINUS6" "MPLUS5" "MMINUS7"
## [100] "MPLUS6" "MMINUS8" "FPLUS7" "FPLUS8" "FMINUS6" "FMINUS7" "MPLUS7" "MMINUS9" "NGRO"
## [109] "FMINUS" "FPLUS" "MMINUS" "MPLUS" "FBEHAV" "MBEHAV"describe(ursala) #always describe the data!## vars n mean sd median trimmed mad min max range skew kurtosis se
## STUDIE 1 677 1.90 0.69 2 1.87 0.00 1 3 2 0.13 -0.88 0.03
## GESCHL 2 665 1.47 0.50 1 1.46 0.00 1 2 1 0.12 -1.99 0.02
## V1 3 677 5.35 1.63 6 5.55 1.48 1 9 8 -0.89 0.38 0.06
## V2 4 677 3.39 2.33 3 3.22 2.97 1 9 8 0.40 -1.32 0.09
## V3 5 677 4.03 2.10 4 4.01 2.97 1 9 8 0.04 -1.17 0.08
## V4 6 677 3.99 2.26 4 3.97 2.97 1 9 8 0.01 -1.41 0.09
## V5 7 677 5.42 1.86 6 5.69 1.48 1 9 8 -0.91 -0.28 0.07
## V6 8 677 5.43 1.72 6 5.69 1.48 1 9 8 -0.99 0.22 0.07
## V7 9 677 5.42 1.83 6 5.69 1.48 1 9 8 -0.96 -0.09 0.07
## V8 10 677 2.81 2.04 2 2.53 1.48 1 9 8 0.84 -0.45 0.08
## V9 11 677 2.46 2.07 1 2.07 0.00 1 9 8 1.25 0.27 0.08
## V10 12 677 3.89 2.09 4 3.83 2.97 1 9 8 0.19 -1.04 0.08
## V11 13 677 4.40 2.09 5 4.48 2.97 1 9 8 -0.16 -1.11 0.08
## V12 14 677 2.77 1.81 2 2.54 1.48 1 9 8 0.97 0.42 0.07
## V13 15 677 5.86 1.20 6 6.00 1.48 2 9 7 -0.78 1.05 0.05
## V14 16 677 5.60 1.66 6 5.83 1.48 1 9 8 -0.96 0.35 0.06
## V15 17 677 3.70 2.25 3 3.59 2.97 1 9 8 0.31 -1.22 0.09
## V16 18 677 5.33 1.70 6 5.51 1.48 1 9 8 -0.69 -0.23 0.07
## V17 19 677 4.66 2.05 5 4.79 1.48 1 9 8 -0.42 -0.86 0.08
## V18 20 677 5.93 1.50 7 6.18 1.48 1 9 8 -1.25 1.26 0.06
## V19 21 677 2.29 1.76 1 1.96 0.00 1 9 8 1.36 0.92 0.07
## V20 22 677 3.85 2.17 4 3.81 2.97 1 9 8 0.10 -1.38 0.08
## V21 23 677 2.80 1.97 2 2.54 1.48 1 9 8 0.87 -0.40 0.08
## V22 24 677 3.06 1.94 3 2.86 2.97 1 9 8 0.62 -0.68 0.07
## V23 25 677 5.11 1.97 6 5.32 1.48 1 9 8 -0.66 -0.86 0.08
## V24 26 677 2.63 2.16 2 2.28 1.48 1 9 8 1.10 -0.20 0.08
## V25 27 677 4.82 1.52 5 4.89 1.48 1 9 8 -0.41 -0.28 0.06
## V26 28 677 4.99 1.82 5 5.17 1.48 1 9 8 -0.64 -0.48 0.07
## V27 29 677 3.83 1.82 4 3.81 2.97 1 9 8 0.11 -1.01 0.07
## V28 30 677 3.68 2.30 3 3.58 2.97 1 9 8 0.25 -1.40 0.09
## V29 31 677 4.48 2.38 5 4.58 2.97 1 9 8 -0.29 -1.46 0.09
## V30 32 677 4.48 2.20 5 4.59 2.97 1 9 8 -0.33 -1.27 0.08
## V31 33 677 5.73 1.26 6 5.86 1.48 1 9 8 -0.83 0.56 0.05
## V32 34 677 4.83 1.95 5 4.96 2.97 1 9 8 -0.41 -0.97 0.08
## V33 35 677 5.93 1.59 7 6.25 0.00 1 9 8 -1.46 1.37 0.06
## V34 36 677 4.13 2.27 4 4.15 2.97 1 9 8 -0.07 -1.44 0.09
## V35 37 677 2.45 2.12 1 2.06 0.00 1 9 8 1.27 0.19 0.08
## V36 38 677 4.19 2.24 4 4.22 2.97 1 9 8 -0.12 -1.40 0.09
## V37 39 677 2.94 2.17 2 2.65 1.48 1 13 12 0.92 0.10 0.08
## V38 40 677 4.42 2.06 4 4.51 2.97 1 9 8 -0.18 -1.18 0.08
## V39 41 677 4.73 1.93 5 4.88 1.48 1 9 8 -0.45 -0.89 0.07
## V40 42 677 3.87 2.19 4 3.83 2.97 1 9 8 0.11 -1.39 0.08
## V41 43 677 5.38 1.65 6 5.59 1.48 1 9 8 -0.81 -0.15 0.06
## V42 44 677 2.43 1.85 2 2.11 1.48 1 9 8 1.17 0.30 0.07
## V43 45 677 4.17 2.13 4 4.21 2.97 1 9 8 -0.09 -1.35 0.08
## V44 46 677 2.48 2.10 1 2.10 0.00 1 9 8 1.16 -0.10 0.08
## V45 47 677 5.06 2.03 6 5.29 1.48 1 7 6 -0.67 -0.93 0.08
## V46 48 677 4.16 2.27 4 4.20 2.97 1 7 6 -0.09 -1.51 0.09
## V47 49 677 2.87 2.22 2 2.59 1.48 1 7 6 0.73 -1.03 0.09
## V48 50 677 3.77 2.12 4 3.71 2.97 1 7 6 0.18 -1.35 0.08
## V49 51 677 5.73 1.40 6 5.95 1.48 1 7 6 -1.04 0.29 0.05
## V50 52 677 4.69 1.80 5 4.82 1.48 1 7 6 -0.48 -0.78 0.07
## V51 53 677 4.44 1.86 5 4.53 1.48 1 7 6 -0.35 -1.02 0.07
## V52 54 677 4.33 2.19 5 4.41 2.97 1 7 6 -0.21 -1.42 0.08
## V53 55 677 2.95 2.21 2 2.69 1.48 1 9 8 0.71 -0.95 0.08
## V54 56 677 4.16 2.21 4 4.20 2.97 1 9 8 -0.11 -1.43 0.08
## V55 57 677 5.07 1.81 6 5.25 1.48 1 9 8 -0.58 -0.75 0.07
## V56 58 677 2.49 1.83 2 2.20 1.48 1 9 8 1.02 -0.14 0.07
## V57 59 677 3.79 2.19 4 3.74 2.97 1 9 8 0.13 -1.40 0.08
## V58 60 677 4.27 1.76 4 4.30 1.48 1 9 8 -0.17 -0.87 0.07
## V59 61 677 3.82 1.99 4 3.77 2.97 1 9 8 0.13 -1.20 0.08
## V60 62 677 3.77 1.93 4 3.71 2.97 1 9 8 0.12 -1.09 0.07
## V61 63 677 2.05 1.88 1 1.61 0.00 1 9 8 1.68 1.42 0.07
## V62 64 677 4.95 1.64 5 5.08 1.48 1 9 8 -0.52 -0.39 0.06
## V63 65 677 3.29 2.24 3 3.10 2.97 1 9 8 0.43 -1.29 0.09
## V64 66 677 4.56 1.78 5 4.66 1.48 1 7 6 -0.35 -0.79 0.07
## V65 67 677 5.19 1.77 6 5.42 1.48 1 7 6 -0.82 -0.31 0.07
## V66 68 677 2.80 1.87 2 2.57 1.48 1 9 8 0.86 -0.44 0.07
## V67 69 677 4.15 2.09 4 4.16 2.97 1 9 8 -0.09 -1.09 0.08
## V68 70 677 3.96 2.68 4 3.86 2.97 1 44 43 5.07 72.33 0.10
## V69 71 677 4.95 2.00 5 5.16 2.97 1 7 6 -0.64 -0.87 0.08
## V70 72 677 3.72 2.18 4 3.65 2.97 1 7 6 0.14 -1.40 0.08
## V71 73 677 3.55 2.08 3 3.44 2.97 1 7 6 0.32 -1.21 0.08
## V72 74 677 3.60 2.38 3 3.49 2.97 1 9 8 0.29 -1.52 0.09
## V73 75 677 4.28 2.40 5 4.34 2.97 1 9 8 -0.18 -1.57 0.09
## V74 76 677 3.17 2.50 2 2.96 1.48 1 9 8 0.55 -1.44 0.10
## MMINUS1 77 677 2.73 1.67 2 2.50 1.48 1 9 8 1.59 3.41 0.06
## FMINUS1 78 677 2.71 1.79 2 2.46 1.48 1 9 8 1.36 2.07 0.07
## MPLUS1 79 677 4.45 1.49 5 4.49 1.48 1 9 8 0.19 1.15 0.06
## FPLUS1 80 677 4.70 1.48 5 4.76 1.48 1 9 8 0.07 1.05 0.06
## MMINUS2 81 677 2.33 1.52 2 2.09 1.48 1 9 8 1.76 4.40 0.06
## MPLUS2 82 677 4.57 1.27 5 4.62 1.48 1 9 8 0.06 1.55 0.05
## MMINUS3 83 677 2.36 1.43 2 2.17 1.48 1 9 8 1.66 4.52 0.05
## FPLUS2 84 677 5.09 1.25 5 5.28 1.48 1 9 8 -0.92 2.23 0.05
## FPLUS3 85 677 4.38 1.24 4 4.36 1.48 1 9 8 0.40 2.36 0.05
## FMINUS2 86 677 3.48 1.44 3 3.39 1.48 1 9 8 0.93 2.10 0.06
## FPLUS4 87 677 4.99 1.02 5 5.04 1.48 1 9 8 -0.01 3.08 0.04
## FMINUS3 88 677 3.23 1.38 3 3.11 1.48 1 9 8 1.31 3.64 0.05
## MMINUS4 89 677 2.68 1.41 2 2.56 1.48 1 9 8 0.69 0.55 0.05
## FPLUS5 90 677 4.96 0.95 5 5.04 1.48 1 9 8 -0.62 1.55 0.04
## MMINUS5 91 677 2.63 1.66 2 2.40 1.48 1 9 8 1.58 3.53 0.06
## FMINUS4 92 677 3.58 1.31 4 3.55 1.48 1 9 8 0.27 0.22 0.05
## FPLUS6 93 677 4.69 1.15 5 4.80 1.48 1 9 8 -0.66 1.20 0.04
## MPLUS3 94 677 3.52 1.58 3 3.51 1.48 1 9 8 0.05 -0.82 0.06
## FMINUS5 95 677 3.17 1.47 3 3.13 1.48 1 9 8 0.24 -0.69 0.06
## MPLUS4 96 677 4.63 1.33 5 4.78 1.48 1 9 8 -0.73 0.43 0.05
## MMINUS6 97 677 3.51 1.58 4 3.51 1.48 1 9 8 -0.06 -0.80 0.06
## MPLUS5 98 677 4.21 1.23 4 4.28 1.48 1 9 8 -0.45 0.32 0.05
## MMINUS7 99 677 2.38 1.21 2 2.24 1.48 1 9 8 1.05 1.28 0.05
## MPLUS6 100 677 3.89 1.09 4 3.90 1.48 1 9 8 0.39 3.10 0.04
## MMINUS8 101 677 2.10 1.24 2 1.92 1.48 1 9 8 1.32 2.34 0.05
## FPLUS7 102 677 4.92 0.98 5 5.01 1.48 1 9 8 -0.61 2.20 0.04
## FPLUS8 103 677 4.46 1.31 5 4.58 1.48 1 9 8 -0.52 0.38 0.05
## FMINUS6 104 677 2.28 1.16 2 2.16 1.48 1 9 8 0.93 1.28 0.04
## FMINUS7 105 677 2.58 1.40 2 2.43 1.48 1 9 8 0.88 0.89 0.05
## MPLUS7 106 677 4.52 1.15 5 4.62 1.48 1 9 8 -0.76 0.67 0.04
## MMINUS9 107 677 2.51 1.23 2 2.41 1.48 1 9 8 0.73 0.49 0.05
## NGRO 108 657 90.12 27.05 86 88.49 28.17 32 186 154 0.56 -0.02 1.06
## FMINUS 109 677 21.02 5.74 20 20.69 5.93 8 42 34 0.57 0.32 0.22
## FPLUS 110 677 38.19 5.87 39 38.38 5.93 13 57 44 -0.34 1.10 0.23
## MMINUS 111 677 20.84 6.74 20 20.51 5.93 8 59 51 0.69 1.49 0.26
## MPLUS 112 677 29.79 6.07 30 30.01 5.93 8 48 40 -0.35 0.10 0.23
## FBEHAV 113 677 122.92 30.16 124 122.95 35.58 50 192 142 -0.03 -0.88 1.16
## MBEHAV 114 677 95.33 24.76 96 95.38 28.17 36 159 123 -0.01 -0.68 0.95Cleaning up the data
We see several things from the description of the file. Although the items are supposed to range from 1 to 7, many of them include 9, one even includes 43. The items are in columns 3 - 76, with some scale information after that.
We will ‘scrub’ the data to convert all missing values to NA and then describe the data again. We do this by specifying that the maximum value we want is 7.
We want to select ‘complete.cases’ (no missing values). We do this by finding which subjects are complete, and then making up the new data set with just complete cases.
cleaned <- scrub(ursala,where= 3:107, max=7 )
complete <-complete.cases(cleaned)
cleaned <- cleaned[complete,] #just those with complete cases
dim(cleaned)## [1] 570 114table(cleaned[,2]) #how many cases?##
## 1 2
## 305 265We seem to have few subjects than Athenstaedt reported (310 females, 266 males). Perhaps we should clean somewhat fewer variables?
(After several attempts, the right value seems to be columns 1-92).
cleaned <- scrub(ursala,where= 3:92, max=7 )
complete <-complete.cases(cleaned)
cleaned <- cleaned[complete,] #just those with complete cases
dim(cleaned)## [1] 576 114table(cleaned[,2]) #how many cases?##
## 1 2
## 310 266ursala.clean <- cleanedMaking a dictionary
Rather than refer to the data as merely items V1 .. V76, it would be nice to know the items. We create a csv file in our friendly text editor, and the read this in as a dictionary.
fn <- "https://personality-project.org/courses/350/Athenstaedt.dictionary.csv"
ursala.dictionary <- read.file(fn) #automatically reads csv## Data from the .csv file https://personality-project.org/courses/350/Athenstaedt.dictionary.csv has been loaded.rownames(ursala.dictionary) <- ursala.dictionary[,1]
headTail(ursala.dictionary, top=8,bottom=8)## Item Content
## Sex Sex Gender
## V1 V1 To pay attention to ones appearance in the office
## V2 V2 Offer fire to somebody
## V3 V3 Paint an Apartment
## V4 V4 Mow the Lawn
## V5 V5 Make the Bed
## V6 V6 Hold the Door Open for your Partner
## V7 V7 Do the Dishes
## ... <NA> <NA>
## V72 V72 Do the Ironing
## V73 V73 Do the Laundry
## V74 V74 Put on Make-up
## MR1 MR1 Femininity Scale
## MR2 MR2 Masculinity Scale
## Femininity Femininity Femininity Scale
## Masulinity Masulinity Masculinity Scale
## MF MF Pooled ScaleAlthough Athenstaedt gave 76 items, her subsequent analysis just used 53 of these. We select those items and then do some basic analysis.
One of those items “put on make up” shows a much bigger effect size than any of the others, and we drop it to make the data more reasonable.
select <- c("V1", "V2", "V10", "V12", "V15", "V17", "V18", "V19", "V20",
"V24", "V25", "V27", "V37", "V50", "V53", "V58", "V60", "V62",
"V66", "V67", "V68")
sel <- which(colnames(ursala.clean) %in% select)
ursala.small <- ursala.clean[-sel]
cd.ursala <- cohen.d(ursala.small[2:55],group= "GESCHL",dictionary=ursala.dictionary)
cd.ursala## Call: cohen.d(x = ursala.small[2:55], group = "GESCHL", dictionary = ursala.dictionary)
## Cohen d statistic of difference between two means
## lower effect upper Item Content
## V3 0.04 0.21 0.37 V3 Paint an Apartment
## V4 0.45 0.62 0.78 V4 Mow the Lawn
## V5 -0.65 -0.49 -0.32 V5 Make the Bed
## V6 0.34 0.51 0.67 V6 Hold the Door Open for your Partner
## V7 -0.84 -0.67 -0.51 V7 Do the Dishes
## V8 0.31 0.47 0.64 V8 Do Extreme Sports
## V9 0.78 0.95 1.13 V9 Tinker with the Car
## V11 0.29 0.45 0.62 V11 Assemble Prefabricated Furniture
## V13 -0.75 -0.59 -0.42 V13 Listen Attentively to Others
## V14 -0.91 -0.74 -0.57 V14 Tell your Partner about Problems at Work
## V16 -0.95 -0.78 -0.61 V16 Set the Table
## V21 -0.97 -0.80 -0.63 V21 Watch Soap Operas
## V22 -0.60 -0.44 -0.27 V22 Take a Friends Arm
## V23 -1.45 -1.27 -1.09 V23 Wrap Presents Beautifully
## V26 0.01 0.17 0.33 V26 Work Overtime
## V28 -0.51 -0.34 -0.18 V28 Babysit
## V29 1.03 1.21 1.39 V29 Change Fuses
## V30 0.39 0.55 0.72 V30 Clean a Drain
## V31 -0.69 -0.53 -0.36 V31 Take Care of Somebody
## V32 0.90 1.07 1.25 V32 Do Repair Work
## V33 0.43 0.59 0.76 V33 Change Light Bulbs
## V34 0.12 0.29 0.45 V34 Wash the Car
## V35 0.20 0.37 0.53 V35 Ride a Motorcycle
## V36 0.36 0.53 0.69 V36 Cook Meat on the Grill
## V38 -1.11 -0.93 -0.76 V38 Dust the Furniture
## V39 0.46 0.63 0.79 V39 Buy Electric Appliances
## V40 -0.72 -0.55 -0.38 V40 Go Dancing
## V41 -0.83 -0.66 -0.49 V41 Go for a Walk through Town
## V42 -0.74 -0.57 -0.41 V42 Go to the Ballet
## V43 -0.86 -0.69 -0.52 V43 Hug a Friend
## V44 -1.07 -0.90 -0.73 V44 Do Handiwork (e.g. Knitting)
## V45 -1.10 -0.92 -0.75 V45 Change Bed Sheets
## V46 -1.34 -1.16 -0.99 V46 Sew on a Button
## V47 -1.28 -1.10 -0.93 V47 Do Aerobics
## V48 0.56 0.73 0.89 V48 Watch Sports on Television
## V49 -0.79 -0.63 -0.46 V49 Talk about Problems
## V51 0.28 0.45 0.61 V51 Talk about Politics
## V52 -1.00 -0.83 -0.66 V52 Take Care of Flowers
## V54 0.64 0.81 0.98 V54 Shovel Snow
## V55 0.14 0.31 0.47 V55 Read non-Fiction Books
## V56 -0.07 0.09 0.26 V56 Organize Company Parties
## V57 0.58 0.75 0.92 V57 Do Home Improvement Jobs
## V59 -0.69 -0.52 -0.36 V59 Buy a Present for a Colleague
## V61 -0.88 -0.72 -0.55 V61 Make Jam
## V63 -0.96 -0.79 -0.62 V63 Decorate the Office with Flowers
## V64 0.44 0.61 0.78 V64 Pick up the Dinner Bill
## V65 -0.55 -0.38 -0.22 V65 Shop for the Family
## V69 -0.69 -0.53 -0.36 V69 Cook
## V70 0.83 1.00 1.17 V70 Help your Partner Put on His or Her Coat
## V71 -0.91 -0.74 -0.57 V71 Wash Windows
## V72 -1.42 -1.24 -1.06 V72 Do the Ironing
## V73 -1.27 -1.10 -0.92 V73 Do the Laundry
## V74 -2.52 -2.31 -2.10 V74 Put on Make-up
##
## Multivariate (Mahalanobis) distance between groups
## [1] 4
## r equivalent of difference between two means
## V3 V4 V5 V6 V7 V8 V9 V11 V13 V14 V16 V21 V22 V23 V26 V28
## 0.10 0.29 -0.24 0.24 -0.32 0.23 0.43 0.22 -0.28 -0.35 -0.36 -0.37 -0.21 -0.53 0.08 -0.17
## V29 V30 V31 V32 V33 V34 V35 V36 V38 V39 V40 V41 V42 V43 V44 V45
## 0.52 0.27 -0.25 0.47 0.28 0.14 0.18 0.25 -0.42 0.30 -0.26 -0.31 -0.28 -0.33 -0.41 -0.42
## V46 V47 V48 V49 V51 V52 V54 V55 V56 V57 V59 V61 V63 V64 V65 V69
## -0.50 -0.48 0.34 -0.30 0.22 -0.38 0.38 0.15 0.05 0.35 -0.25 -0.34 -0.36 0.29 -0.19 -0.25
## V70 V71 V72 V73 V74
## 0.45 -0.35 -0.53 -0.48 -0.76#note that we just want column 2
cd.ursala1 <- cohen.d(ursala.small[2:54],group= "GESCHL",dictionary=ursala.dictionary[,2, drop=FALSE], sort="descending")## Warning in xtfrm.data.frame(x): cannot xtfrm data framescd.ursala1## Call: cohen.d(x = ursala.small[2:54], group = "GESCHL", sort = "descending",
## dictionary = ursala.dictionary[, 2, drop = FALSE])
## Cohen d statistic of difference between two means
## lower effect upper Content
## V29 1.03 1.21 1.39 Change Fuses
## V32 0.90 1.07 1.25 Do Repair Work
## V70 0.83 1.00 1.17 Help your Partner Put on His or Her Coat
## V9 0.78 0.95 1.13 Tinker with the Car
## V54 0.64 0.81 0.98 Shovel Snow
## V57 0.58 0.75 0.92 Do Home Improvement Jobs
## V48 0.56 0.73 0.89 Watch Sports on Television
## V39 0.46 0.63 0.79 Buy Electric Appliances
## V4 0.45 0.62 0.78 Mow the Lawn
## V64 0.44 0.61 0.78 Pick up the Dinner Bill
## V33 0.43 0.59 0.76 Change Light Bulbs
## V30 0.39 0.55 0.72 Clean a Drain
## V36 0.36 0.53 0.69 Cook Meat on the Grill
## V6 0.34 0.51 0.67 Hold the Door Open for your Partner
## V8 0.31 0.47 0.64 Do Extreme Sports
## V11 0.29 0.45 0.62 Assemble Prefabricated Furniture
## V51 0.28 0.45 0.61 Talk about Politics
## V35 0.20 0.37 0.53 Ride a Motorcycle
## V55 0.14 0.31 0.47 Read non-Fiction Books
## V34 0.12 0.29 0.45 Wash the Car
## V3 0.04 0.21 0.37 Paint an Apartment
## V26 0.01 0.17 0.33 Work Overtime
## V56 -0.07 0.09 0.26 Organize Company Parties
## V28 -0.51 -0.34 -0.18 Babysit
## V65 -0.55 -0.38 -0.22 Shop for the Family
## V22 -0.60 -0.44 -0.27 Take a Friends Arm
## V5 -0.65 -0.49 -0.32 Make the Bed
## V59 -0.69 -0.52 -0.36 Buy a Present for a Colleague
## V31 -0.69 -0.53 -0.36 Take Care of Somebody
## V69 -0.69 -0.53 -0.36 Cook
## V40 -0.72 -0.55 -0.38 Go Dancing
## V42 -0.74 -0.57 -0.41 Go to the Ballet
## V13 -0.75 -0.59 -0.42 Listen Attentively to Others
## V49 -0.79 -0.63 -0.46 Talk about Problems
## V41 -0.83 -0.66 -0.49 Go for a Walk through Town
## V7 -0.84 -0.67 -0.51 Do the Dishes
## V43 -0.86 -0.69 -0.52 Hug a Friend
## V61 -0.88 -0.72 -0.55 Make Jam
## V14 -0.91 -0.74 -0.57 Tell your Partner about Problems at Work
## V71 -0.91 -0.74 -0.57 Wash Windows
## V16 -0.95 -0.78 -0.61 Set the Table
## V63 -0.96 -0.79 -0.62 Decorate the Office with Flowers
## V21 -0.97 -0.80 -0.63 Watch Soap Operas
## V52 -1.00 -0.83 -0.66 Take Care of Flowers
## V44 -1.07 -0.90 -0.73 Do Handiwork (e.g. Knitting)
## V45 -1.10 -0.92 -0.75 Change Bed Sheets
## V38 -1.11 -0.93 -0.76 Dust the Furniture
## V73 -1.27 -1.10 -0.92 Do the Laundry
## V47 -1.28 -1.10 -0.93 Do Aerobics
## V46 -1.34 -1.16 -0.99 Sew on a Button
## V72 -1.42 -1.24 -1.06 Do the Ironing
## V23 -1.45 -1.27 -1.09 Wrap Presents Beautifully
##
## Multivariate (Mahalanobis) distance between groups
## [1] 3.6
## r equivalent of difference between two means
## V3 V4 V5 V6 V7 V8 V9 V11 V13 V14 V16 V21 V22 V23 V26 V28
## 0.10 0.29 -0.24 0.24 -0.32 0.23 0.43 0.22 -0.28 -0.35 -0.36 -0.37 -0.21 -0.53 0.08 -0.17
## V29 V30 V31 V32 V33 V34 V35 V36 V38 V39 V40 V41 V42 V43 V44 V45
## 0.52 0.27 -0.25 0.47 0.28 0.14 0.18 0.25 -0.42 0.30 -0.26 -0.31 -0.28 -0.33 -0.41 -0.42
## V46 V47 V48 V49 V51 V52 V54 V55 V56 V57 V59 V61 V63 V64 V65 V69
## -0.50 -0.48 0.34 -0.30 0.22 -0.38 0.38 0.15 0.05 0.35 -0.25 -0.34 -0.36 0.29 -0.19 -0.25
## V70 V71 V72 V73
## 0.45 -0.35 -0.53 -0.48error.dots(cd.ursala1,head=26,tail=26,main="Athenstaedt's Behavioral Data - Effect sizes and standard errors")
abline(v=0)
Identify two dimensions of the scales
Do a factor analysis, show the items, then choose those items that have their highest loadings on each factor.
We need to hand fix these because one item is mis-classified.
f2 <- fa(ursala.small[3:54],2) #this drops the make up item## Loading required namespace: GPArotation fa.lookup(f2, dictionary=ursala.dictionary)## MR1 MR2 com h2 Item Content
## V46 0.72 0.01 1.00 0.52 V46 Sew on a Button
## V45 0.71 -0.02 1.00 0.51 V45 Change Bed Sheets
## V72 0.68 -0.08 1.03 0.48 V72 Do the Ironing
## V38 0.66 -0.01 1.00 0.44 V38 Dust the Furniture
## V71 0.65 0.15 1.10 0.43 V71 Wash Windows
## V16 0.65 0.07 1.02 0.41 V16 Set the Table
## V73 0.64 -0.07 1.02 0.42 V73 Do the Laundry
## V52 0.60 0.06 1.02 0.36 V52 Take Care of Flowers
## V7 0.56 -0.03 1.01 0.32 V7 Do the Dishes
## V23 0.53 -0.21 1.30 0.34 V23 Wrap Presents Beautifully
## V5 0.52 0.07 1.04 0.27 V5 Make the Bed
## V69 0.51 0.00 1.00 0.26 V69 Cook
## V61 0.49 0.10 1.08 0.24 V61 Make Jam
## V65 0.49 0.21 1.37 0.27 V65 Shop for the Family
## V31 0.49 0.11 1.10 0.24 V31 Take Care of Somebody
## V63 0.47 0.00 1.00 0.23 V63 Decorate the Office with Flowers
## V44 0.46 -0.10 1.09 0.23 V44 Do Handiwork (e.g. Knitting)
## V47 0.45 -0.19 1.36 0.25 V47 Do Aerobics
## V43 0.41 -0.08 1.08 0.18 V43 Hug a Friend
## V41 0.39 -0.12 1.19 0.17 V41 Go for a Walk through Town
## V14 0.39 -0.12 1.20 0.17 V14 Tell your Partner about Problems at Work
## V28 0.38 0.12 1.21 0.15 V28 Babysit
## V49 0.34 -0.13 1.28 0.14 V49 Talk about Problems
## V59 0.34 0.02 1.01 0.11 V59 Buy a Present for a Colleague
## V42 0.34 -0.04 1.03 0.12 V42 Go to the Ballet
## V13 0.33 -0.13 1.29 0.13 V13 Listen Attentively to Others
## V22 0.29 -0.04 1.04 0.09 V22 Take a Friends Arm
## V40 0.23 -0.07 1.19 0.06 V40 Go Dancing
## V32 -0.09 0.77 1.03 0.62 V32 Do Repair Work
## V29 -0.11 0.74 1.04 0.58 V29 Change Fuses
## V54 0.02 0.66 1.00 0.43 V54 Shovel Snow
## V57 0.00 0.65 1.00 0.42 V57 Do Home Improvement Jobs
## V30 0.10 0.63 1.05 0.40 V30 Clean a Drain
## V4 0.06 0.62 1.02 0.38 V4 Mow the Lawn
## V39 0.05 0.59 1.01 0.34 V39 Buy Electric Appliances
## V33 0.06 0.58 1.02 0.33 V33 Change Light Bulbs
## V9 -0.22 0.57 1.28 0.40 V9 Tinker with the Car
## V36 0.12 0.54 1.10 0.30 V36 Cook Meat on the Grill
## V11 0.03 0.54 1.01 0.29 V11 Assemble Prefabricated Furniture
## V70 -0.02 0.51 1.00 0.26 V70 Help your Partner Put on His or Her Coat
## V3 0.23 0.51 1.39 0.29 V3 Paint an Apartment
## V34 0.11 0.48 1.11 0.24 V34 Wash the Car
## V64 -0.05 0.38 1.03 0.15 V64 Pick up the Dinner Bill
## V6 0.08 0.36 1.10 0.14 V6 Hold the Door Open for your Partner
## V26 0.10 0.31 1.22 0.10 V26 Work Overtime
## V48 -0.12 0.31 1.30 0.11 V48 Watch Sports on Television
## V21 0.27 -0.30 1.98 0.17 V21 Watch Soap Operas
## V35 -0.09 0.28 1.22 0.09 V35 Ride a Motorcycle
## V56 0.15 0.27 1.53 0.09 V56 Organize Company Parties
## V8 -0.10 0.27 1.25 0.09 V8 Do Extreme Sports
## V55 0.05 0.27 1.06 0.07 V55 Read non-Fiction Books
## V51 0.02 0.26 1.01 0.07 V51 Talk about Politics summary(f2)##
## Factor analysis with Call: fa(r = ursala.small[3:54], nfactors = 2)
##
## Test of the hypothesis that 2 factors are sufficient.
## The degrees of freedom for the model is 1223 and the objective function was 7.23
## The number of observations was 576 with Chi Square = 4014.92 with prob < 1.9e-293
##
## The root mean square of the residuals (RMSA) is 0.06
## The df corrected root mean square of the residuals is 0.06
##
## Tucker Lewis Index of factoring reliability = 0.683
## RMSEA index = 0.063 and the 10 % confidence intervals are 0.061 0.065
## BIC = -3758.6
## With factor correlations of
## MR1 MR2
## MR1 1.00 -0.08
## MR2 -0.08 1.00 ursala.keys <- factor2cluster(f2)
ursala.keys <- cbind(ursala.keys,ursala.keys[,1] - ursala.keys[,2]) #add a total
ursala.keys["V21",] <- c(1,0,1) # a hard patch
urs.keys <- keys2list(ursala.keys)
names(urs.keys)[3] <- "F+M behavior"
urs.keys # show them ## $MR1
## [1] "V5" "V7" "V13" "V14" "V16" "V21" "V22" "V23" "V28" "V31" "V38" "V40" "V41" "V42" "V43" "V44"
## [17] "V45" "V46" "V47" "V49" "V52" "V59" "V61" "V63" "V65" "V69" "V71" "V72" "V73"
##
## $MR2
## [1] "V3" "V4" "V6" "V8" "V9" "V11" "V26" "V29" "V30" "V32" "V33" "V34" "V35" "V36" "V39" "V48"
## [17] "V51" "V54" "V55" "V56" "V57" "V64" "V70"
##
## $`F+M behavior`
## [1] "-V3" "-V4" "V5" "-V6" "V7" "-V8" "-V9" "-V11" "V13" "V14" "V16" "V21" "V22"
## [14] "V23" "-V26" "V28" "-V29" "-V30" "V31" "-V32" "-V33" "-V34" "-V35" "-V36" "V38" "-V39"
## [27] "V40" "V41" "V42" "V43" "V44" "V45" "V46" "V47" "-V48" "V49" "-V51" "V52" "-V54"
## [40] "-V55" "-V56" "-V57" "V59" "V61" "V63" "-V64" "V65" "V69" "-V70" "V71" "V72" "V73" urs.scales <- scoreItems(urs.keys,ursala.small)
#show the basic statistics
urs.scales## Call: scoreItems(keys = urs.keys, items = ursala.small)
##
## (Unstandardized) Alpha:
## MR1 MR2 F+M behavior
## alpha 0.9 0.88 0.88
##
## Standard errors of unstandardized Alpha:
## MR1 MR2 F+M behavior
## ASE 0.0081 0.0099 0.008
##
## Average item correlation:
## MR1 MR2 F+M behavior
## average.r 0.24 0.24 0.13
##
## Median item correlation:
## MR1 MR2 F+M behavior
## 0.22 0.21 0.11
##
## Guttman 6* reliability:
## MR1 MR2 F+M behavior
## Lambda.6 0.93 0.91 0.92
##
## Signal/Noise based upon av.r :
## MR1 MR2 F+M behavior
## Signal/Noise 9 7.2 7.6
##
## Scale intercorrelations corrected for attenuation
## raw correlations below the diagonal, alpha on the diagonal
## corrected correlations above the diagonal:
## MR1 MR2 F+M behavior
## MR1 0.900 -0.092 0.89
## MR2 -0.082 0.877 -0.77
## F+M behavior 0.789 -0.677 0.88
##
## In order to see the item by scale loadings and frequency counts of the data
## print with the short option = FALSE expand.ursala <- data.frame(ursala.small[2:54],urs.scales$scores)What are the Cohen.d values for our scales?
Use the ’cohen.d function for just those scale values.
cd.scales <- cohen.d(expand.ursala[c(1,54,55,56)],"GESCHL")
cd.scales## Call: cohen.d(x = expand.ursala[c(1, 54, 55, 56)], group = "GESCHL")
## Cohen d statistic of difference between two means
## lower effect upper
## MR1 -2.02 -1.82 -1.63
## MR2 1.06 1.24 1.42
## F.M.behavior -3.12 -2.89 -2.65
##
## Multivariate (Mahalanobis) distance between groups
## [1] 2.9
## r equivalent of difference between two means
## MR1 MR2 F.M.behavior
## -0.67 0.53 -0.82Some basic psychometrics
We want to find the reliabilities of our new scales. We will call the ‘reliability’ function which in turn calls multiple reliability functions (omega, splitHalf). Note that the ‘alpha’ values match what we found when we used ‘scoreItems’.
rel <- reliability(urs.keys,expand.ursala)
ursala.average.cd <- cd.validity(cd.ursala,urs.keys)
ursala.validity <- cor(expand.ursala[c(1,54:56)])[1,2:4]
ursala.item.validity <- item.validity(expand.ursala,expand.ursala[,"GESCHL"],urs.keys)
ursala.validity <- cor(expand.ursala[c(1,54:56)])[1,2:4]
rel## Measures of reliability
## reliability(keys = urs.keys, items = expand.ursala)
## omega_h alpha omega.tot Uni r.fit fa.fit max.split min.split mean.r med.r n.items
## MR1 0.57 0.90 0.91 0.74 0.81 0.91 0.94 0.78 0.24 0.22 29
## MR2 0.70 0.87 0.89 0.73 0.78 0.94 0.92 0.78 0.23 0.21 23
## F+M behavior 0.13 0.88 0.90 0.22 0.39 0.56 0.93 0.66 0.13 0.11 52 ursala.average.cd #in d units## [,1]
## MR1 -0.7453770
## MR2 0.5815288
## F+M behavior -0.6729057 ursala.item.validity # in r units## [,1]
## MR1 -0.3422164
## MR2 0.2713089
## F+M behavior -0.3108535Predicted validity
Eagly and Revelle (2022) suggest that the predicted validity of a scale is the average item validity/sqrt(average within scale correlation)
The ‘predicted.validity’ function does this.
predicted <- predicted.validity(expand.ursala,expand.ursala[,"GESCHL"],urs.keys)
predicted #asymptotic values are the most we can expect ## Call: predicted.validity(x = expand.ursala, criteria = expand.ursala[,
## "GESCHL"], keys = urs.keys)
##
## Predicted Asymptotic Scale Validity:
## [,1]
## MR1 -0.70
## MR2 0.56
## F+M behavior -0.88
##
## For predicted scale validities, average item validities, or scale reliabilities, print the separate objectspredicted$predicted # compare these with what we actually observed## [,1]
## MR1 -0.6675261
## MR2 0.5253957
## F+M behavior -0.8226188ursala.validity #the correlations with the criterion## MR1 MR2 F.M.behavior
## -0.6722222 0.5270454 -0.8213784Organize a data frame of results
This is the table to put into the manuscript. It is interesting to compare the average absolute effect size (.67) with what we get for the pooled values (2.89).
table.df <- data.frame(k=rel$result.df[,11],
rel$result.df[,c(1:3,9)], item.val=ursala.item.validity, scale=ursala.validity,
predicted = predicted$predicted, average.cd = ursala.average.cd, cd=cd.scales$cohen.d[,"effect"] )
round(table.df,2)## k omega_h alpha omega.tot mean.r item.val scale predicted average.cd cd
## MR1 29 0.57 0.90 0.91 0.24 -0.34 -0.67 -0.67 -0.75 -1.82
## MR2 23 0.70 0.87 0.89 0.23 0.27 0.53 0.53 0.58 1.24
## F+M behavior 52 0.13 0.88 0.90 0.13 -0.31 -0.82 -0.82 -0.67 -2.89mean(abs(cd.ursala1$cohen.d[,"effect"])) #how big is the average absolute effect size?## [1] 0.6729057Now do the final graphics
cd.items <- cohen.d(expand.ursala[-56] , group = "GESCHL",dictionary=ursala.dictionary[,2,drop=FALSE])
error.dots(cd.items,head=30,tail=30, main="Athenstaedt’s Behavioral Data \n Items and Scales",
select=2:53)
error.dots(cd.items,head=30,tail=30, main="Athenstaedt’s Behavioral Data \n Items and Scales",
select=c(1,54), color='red', fg="red", add=TRUE)
abline(v=0)
As well as the scatter histogram
scatterHist(MR1~ MR2 + GESCHL, data=expand.ursala[1:55],cex.point=.4,smooth=FALSE,
xlab="Masculine Scale",ylab="Feminine Scale",
correl=FALSE,d.arrow=TRUE,col=c("red","blue"), lwd=4,
cex.main=1.5,main="Scatter Plot and Density",cex.axis=2)