Using R to score personality scales*

William Revelle
Northwestern University

February 25, 2013

*Part of a set of tutorials for the psych package.

Contents

1 Overview for the impatient
2 An example
 2.1 Getting the data
  2.1.1 Copy the data from another program using the copy and paste commands of your operating system
  2.1.2 An example data set
3 Scoring scales: an example
 3.1 Long output
 3.2 Get the actual scores for analysis.
4 The example, continued
5 Even more analysis
 5.1 Exploring a real data set

Abstract

The psych package (Revelle, 2013) was developed to perform most basic psychometric functions using R (R Development Core Team, 2012) One frequently called upon is the need to take a set of items (e.g., a questionnaire) and score one or more scales on that questionnaire. Scores for subsequent analysis, reliabilities and intercorrelations are easily done using the score.items function.

Suppose you have given a questionnaire with some items (n) to some participants (N). You would like to create scale scores for each person on k different scales. This may be done using the psych package in R. The following assumes that you have installed R and downloaded the psych package.

1 Overview for the impatient

Remember, before using psych you must make it active:

  library(psych)

  1. Enter the data into a spreadsheet (Excel or Numbers) or a text file using a text editor (Word, Pages, BBEdit). The first line of the file should include names for the variables (e.g., Q1, Q2, ... Qn).
  2. Copy the data to the clipboard (using the normal copy command for your spreadsheet or word processor).
  3. Read the data into R using the read.clipboard command. (Depending upon your data file, this might need to be read.clipboard.csv (for comma separated data fields) or read.clipboard.tab (for tab separated data fields).
  4. Construct a set of scoring keys for the scales you want to score. This is simply the item numbers that go into each scale. A negative sign implies that the item will be reverse scored.
  5. Use the score.items function to score the scales.
  6. Use the output for score.items for further analysis.

2 An example

Suppose we have 12 items for 20 subjects. The items represent 4 different scales: Positive Energetic Arousal (EAp), Negative Energetic Arousal (EAn), Tense Arousal (TAp) and negative Tense Arousal (TAn, also known as being relaxed). These four scales can also be thought of a forming two higher order constructs, Energetic Arousal (EA) and Tense Arousal (TA). EA is just EAp - EAn, and similarly TA is just TAp - TAn.

2.1 Getting the data

There are of course many ways to enter data into R. Reading from a local file using read.table is perhaps the most preferred. You first need to find the file and then read it. This can be done with the file.choose and read.table functions:

  file.name <- file.choose()
  my.data <- read.table(file.name)

file.choose opens a search window on your system just like any open file command does. It doesn’t actually read the file, it just finds the file. The read command is also necessary.

2.1.1 Copy the data from another program using the copy and paste commands of your operating system

However, many users will enter their data in a text editor or spreadsheet program and then want to copy and paste into R. This may be done by using read.table and specifying the input file as “clipboard" (PCs) or “pipe(pbpaste)" (Macs). Alternatively, the read.clipboard set of functions are perhaps more user friendly:

read.clipboard
is the base function for reading data from the clipboard.
read.clipboard.csv
for reading text that is comma delimited.
read.clipboard.tab
for reading text that is tab delimited (e.g., copied directly from an Excel file).
read.clipboard.lower
for reading input of a lower triangular matrix with or without a diagonal. The resulting object is a square matrix.
read.clipboard.upper
for reading input of an upper triangular matrix.
read.clipboard.fwf
for reading in fixed width fields (some very old data sets)

For example, given a data set copied to the clipboard from a spreadsheet, just enter the command

  > my.data <- read.clipboard()

This will work if every data field has a value and even missing data are given some values (e.g., NA or -999). If the data were entered in a spreadsheet and the missing values were just empty cells, then the data should be read in as a tab delimited or by using the read.clipboard.tab function.

  > my.data <- read.clipboard(sep="\t")   #define the tab option, or
  > my.tab.data <- read.clipboard.tab()   #just use the alternative function

For the case of data in fixed width fields (some old data sets tend to have this format), copy to the clipboard and then specify the width of each field (in the example below, the first variable is 5 columns, the second is 2 columns, the next 5 are 1 column the last 4 are 3 columns).

  > my.data <- read.clipboard.fwf(widths=c(5,2,rep(1,5),rep(3,4))

2.1.2 An example data set

Consider the data in Table 1. Read them into the clipboard and go. (These data are the first 20 cases from the msq data set in the psych package).


Table 1: A sample data file with 12 items for 20 subjects.
    active alert aroused sleepy tired drowsy anxious jittery nervous calm relaxed at-ease
  1       1     1       1      1     1      1       1       1       1    1       1       1
  2       1     1       0      1     1      1       0       0       0    1       1       1
  3       1     0       0      0     1      0       0       0       0    1       2       2
  4       1     1       1      1     1      1       1       3       2    1       2       1
  5       2     1       2      1     1      1      NA       1       0    3       3       3
  6       2     1       1      2     2      2      NA       0       0    2       2       1
  7       0     1       0      2     3      3      NA       0       0    2       2       1
  8       0     0       0      1     2      1      NA       0       0    1       2       0
  9       1     0       1      2     0      2      NA       1       0    0       2       2
  10      0     2       0      2     2      2      NA       1       0    2       2       1
  11      0     0       0      3     2      2      NA       0       0    2       2       2
  12      1     1       0      1     1      1      NA       1       0    1       1       0
  13      0     0       0      3     3      2      NA       1       0    0       2       0
  14      2     1       1      1     0      0      NA       0       0    2       2       1
  15      0     2       0      0     2      1      NA       0       0    3       3       3
  16      0     0       0      3     3      3      NA       1       0    1       1       1
  17      0     1       1      1     1      1      NA       0       0    1       1       1
  18      3     2       0      2     2      3      NA       0       0    3       3       3
  19      0     0       0      3     3      2      NA       0       0    2       1       0
  20      0     1       0      1     2      1      NA       0       0    3       2       2
  

  library(psych)
  my.data <- read.clipboard.tab()  #tab delimited data from a spreadsheet or
  my.data<- read.clipboard()   #data from a text editor with spaces between the fields.
  describe(my.data)  # to make sure you got the right data in.
  

  > describe(my.data)  # to make sure you got the right data in.
          var  n mean   sd median trimmed  mad min max range  skew kurtosis   se
  active    1 20 0.75 0.91    0.5    0.62 0.74   0   3     3  0.87    -0.37 0.20
  alert     2 20 0.80 0.70    1.0    0.75 0.74   0   2     2  0.25    -1.06 0.16
  aroused   3 20 0.40 0.60    0.0    0.31 0.00   0   2     2  1.06    -0.01 0.13
  sleepy    4 20 1.55 0.94    1.0    1.56 1.48   0   3     3  0.22    -1.10 0.21
  tired     5 20 1.65 0.93    2.0    1.69 1.48   0   3     3 -0.05    -1.06 0.21
  drowsy    6 20 1.50 0.89    1.0    1.50 1.48   0   3     3  0.21    -0.89 0.20
  anxious   7  4 0.50 0.58    0.5    0.50 0.74   0   1     1  0.00    -2.44 0.29
  jittery   8 20 0.50 0.76    0.0    0.38 0.00   0   3     3  1.70     3.00 0.17
  nervous   9 20 0.15 0.49    0.0    0.00 0.00   0   2     2  2.94     7.68 0.11
  calm     10 20 1.60 0.94    1.5    1.62 0.74   0   3     3  0.09    -1.10 0.21
  relaxed  11 20 1.85 0.67    2.0    1.81 0.00   1   3     2  0.15    -0.93 0.15
  at.ease  12 20 1.30 0.98    1.0    1.25 1.48   0   3     3  0.38    -0.96 0.22
  

3 Scoring scales: an example

To score particular items on particular scales, we must create a set of scoring keys. These simply tell us which items go on which scales. Note that we can have scales with overlapping items.

Two things to note. The number of variables is the total number of variables (columns) in the data file. You do not need to include all of these items in the scoring keys, but you need to say how many there are. For the keys, items are scored either +1, -1 or 0 (not scored). Just specify the items to score and their direction.

  my.keys <- make.keys(nvars=12,list(EA=c(1:3,-4,-5,-6),TA=c(7:9,-10,-11,-12),
                      EAp =1:3,EAn=4:6,TAp =7:9,TAn=10:12))
  my.scales <- score.items(my.keys,my.data)
  my.scales   #show the output
  my.scores <- my.scales$scores

  Produces this output:
  > my.keys <- make.keys(nvars=12,list(EA=c(1:3,-4,-5,-6),TA=c(7:9,-10,-11,-12),
  +                     EAp =1:3,EAn=4:6,TAp =7:9,TAn=10:12))
  > my.scales <- score.items(my.keys,my.data)
  > my.scales   #show the output
  Call: score.items(keys = my.keys, items = my.data)
  
  (Unstandardized) Alpha:
          EA   TA  EAp  EAn  TAp  TAn
  alpha 0.77 0.73 0.57 0.86 0.78 0.82
  
  Average item correlation:
              EA   TA EAp  EAn  TAp TAn
  average.r 0.36 0.31 0.3 0.68 0.54 0.6
  
   Guttman 6* reliability:
             EA   TA  EAp  EAn  TAp TAn
  Lambda.6 0.92 0.91 0.78 0.94 0.89 0.9
  
  Scale intercorrelations corrected for attenuation
   raw correlations below the diagonal, alpha on the diagonal
   corrected correlations above the diagonal:
         EA     TA   EAp    EAn   TAp   TAn
  EA   0.77 -0.215  1.15 -1.102  0.21  0.38
  TA  -0.16  0.728 -0.47  0.032  0.84 -1.15
  EAp  0.76 -0.301  0.57 -0.569  0.29  0.73
  EAn -0.90  0.026 -0.40  0.863 -0.12 -0.11
  TAp  0.16  0.630  0.19 -0.097  0.78 -0.24
  TAn  0.30 -0.885  0.50 -0.091 -0.20  0.82
  
   In order to see the item by scale loadings and frequency counts of the data
   print with the short option = FALSE
  

Two things to notice about this output is a) the message about how to get more information (item by scale correlations and frequency counts) and b) that the correlation matrix between the six scales has the raw correlations below the diagonal, alpha reliabilities on the diagonal, and correlations adjusted for reliability above the diagonal. Because EAp and EAn are both part of EA, they correlate with the total more than would be expected given their reliability. Hence the impossible values greater than |0.0|.

3.1 Long output

To get the scale correlations corrected for item overlap and scale reliability, we print the object that we found, but ask for long output.

  print(my.scales,short=FALSE)

  
  Call: score.items(keys = my.keys, items = my.data)
  
  (Unstandardized) Alpha:
          EA   TA  EAp  EAn  TAp  TAn
  alpha 0.77 0.73 0.57 0.86 0.78 0.82
  
  Average item correlation:
              EA   TA EAp  EAn  TAp TAn
  average.r 0.36 0.31 0.3 0.68 0.54 0.6
  
   Guttman 6* reliability:
             EA   TA  EAp  EAn  TAp TAn
  Lambda.6 0.92 0.91 0.78 0.94 0.89 0.9
  
  Scale intercorrelations corrected for attenuation
   raw correlations below the diagonal, alpha on the diagonal
   corrected correlations above the diagonal:
         EA     TA   EAp    EAn   TAp   TAn
  EA   0.77 -0.215  1.15 -1.102  0.21  0.38
  TA  -0.16  0.728 -0.47  0.032  0.84 -1.15
  EAp  0.76 -0.301  0.57 -0.569  0.29  0.73
  EAn -0.90  0.026 -0.40  0.863 -0.12 -0.11
  TAp  0.16  0.630  0.19 -0.097  0.78 -0.24
  TAn  0.30 -0.885  0.50 -0.091 -0.20  0.82
  
  Item by scale correlations:
   corrected for item overlap and scale reliability
             EA    TA   EAp   EAn   TAp   TAn
  active   0.55 -0.29  0.75 -0.30  0.06  0.40
  alert    0.40 -0.42  0.58 -0.20  0.07  0.57
  aroused  0.57  0.06  0.60 -0.43  0.40  0.17
  sleepy  -0.79  0.20 -0.40  0.86 -0.03 -0.27
  tired   -0.85 -0.08 -0.60  0.82 -0.20 -0.02
  drowsy  -0.73 -0.05 -0.18  0.90 -0.05  0.04
  anxious  0.03  0.40  0.24  0.10  0.77 -0.05
  jittery  0.12  0.59  0.17 -0.06  0.86 -0.24
  nervous  0.27  0.55  0.23 -0.23  0.91 -0.16
  calm     0.17 -0.78  0.45  0.04 -0.30  0.81
  relaxed  0.26 -0.65  0.48 -0.06 -0.06  0.79
  at.ease  0.38 -0.74  0.53 -0.21 -0.14  0.85
  
  Non missing response frequency for each item
             0    1    2    3 miss
  active  0.50 0.30 0.15 0.05  0.0
  alert   0.35 0.50 0.15 0.00  0.0
  aroused 0.65 0.30 0.05 0.00  0.0
  sleepy  0.10 0.45 0.25 0.20  0.0
  tired   0.10 0.35 0.35 0.20  0.0
  drowsy  0.10 0.45 0.30 0.15  0.0
  anxious 0.50 0.50 0.00 0.00  0.8
  jittery 0.60 0.35 0.00 0.05  0.0
  nervous 0.90 0.05 0.05 0.00  0.0
  calm    0.10 0.40 0.30 0.20  0.0
  relaxed 0.00 0.30 0.55 0.15  0.0
  at.ease 0.20 0.45 0.20 0.15  0.0

3.2 Get the actual scores for analysis.

Although we would probably not look at the raw scores, we can if we want by asking for the scores object which is part of the my.scales output. For printing purposes, we round them to two decimal places for compactness.

  my.scores <- my.scales$scores
  round(my.scores,2)

  
  > round(my.scores,2)
       EA   TA  EAp  EAn  TAp  TAn
  1  1.50 1.50 1.00 1.00 1.00 1.00
  2  1.33 1.00 0.67 1.00 0.00 1.00
  3  1.50 0.67 0.33 0.33 0.00 1.67
  4  1.50 1.83 1.00 1.00 2.00 1.33
  5  1.83 0.25 1.67 1.00 0.50 3.00
  6  1.17 0.75 1.33 2.00 0.17 1.67
  7  0.33 0.75 0.33 2.67 0.17 1.67
  8  0.83 1.08 0.00 1.33 0.17 1.00
  9  1.17 1.08 0.67 1.33 0.50 1.33
  10 0.83 0.92 0.67 2.00 0.50 1.67
  11 0.33 0.58 0.00 2.33 0.17 2.00
  12 1.33 1.42 0.67 1.00 0.50 0.67
  13 0.17 1.42 0.00 2.67 0.50 0.67
  14 2.00 0.75 1.33 0.33 0.17 1.67
  15 1.33 0.08 0.67 1.00 0.17 3.00
  16 0.00 1.25 0.00 3.00 0.50 1.00
  17 1.33 1.08 0.67 1.00 0.17 1.00
  18 1.17 0.08 1.67 2.33 0.17 3.00
  19 0.17 1.08 0.00 2.67 0.17 1.00
  20 1.00 0.42 0.33 1.33 0.17 2.33
  

4 The example, continued

Once you have the results, you should probably want to describe them and also show a graphic of the scatterplot using the pairs.panels function (Figure 1).

  describe(my.scores)
  pairs.panels(my.scores}

  P> describe(my.scores)
      var  n mean   sd median trimmed  mad  min  max range  skew kurtosis   se
  EA    1 20 1.04 0.57   1.17    1.05 0.49 0.00 2.00  2.00 -0.38    -1.05 0.13
  TA    2 20 0.90 0.47   0.96    0.91 0.43 0.08 1.83  1.75 -0.07    -0.81 0.11
  EAp   3 20 0.65 0.55   0.67    0.60 0.49 0.00 1.67  1.67  0.42    -1.01 0.12
  EAn   4 20 1.57 0.82   1.33    1.56 0.74 0.33 3.00  2.67  0.27    -1.35 0.18
  TAp   5 20 0.38 0.45   0.17    0.29 0.12 0.00 2.00  2.00  2.34     5.62 0.10

PIC

Figure 1: A simple scatter plot matrix shows the histograms for the variables on the diagonal, the correlations above the diagonal, and the scatter plots below the diagonal. The best fitting loess regression is shown as the red line.

5 Even more analysis

Far more analyses could be done with these data, but the basic scale scoring techniques is a start. Download the vignette for using psych for even more guidance. http://cran.r-project.org/web/packages/psych/vignettes/overview.pdf. On a Mac, this is also available in the vignettes list in the help menu.

In addition, look at the examples in the help for score.items.

5.1 Exploring a real data set

.

The 12 mood items for 20 subjects were taken from the much larger data set, msq in the psych package. That data set has 92 variables for 3896 subjects. We can repeat our analysis of EA and TA on that data set.

First we get the data for the items that match our small example. Then we describe the data, and finally, find the 6 scales as we did before.

  select <- colnames(my.data)
  select[12] <- 'at-ease'
  small.msq <- msq[select]
  describe(small.msq)
  msq.scales <- score.items(my.keys,small.msq)
  msq.scales   #show the output
  

         var    n mean   sd median trimmed  mad min max range  skew kurtosis   se
  active    1 3890 1.03 0.93      1    0.95 1.48   0   3     3  0.47    -0.76 0.01
  alert     2 3885 1.15 0.91      1    1.09 1.48   0   3     3  0.33    -0.76 0.01
  aroused   3 3890 0.71 0.85      0    0.59 0.00   0   3     3  0.95    -0.04 0.01
  sleepy    4 3880 1.25 1.05      1    1.18 1.48   0   3     3  0.40    -1.04 0.02
  tired     5 3886 1.39 1.04      1    1.36 1.48   0   3     3  0.22    -1.10 0.02
  drowsy    6 3884 1.16 1.03      1    1.08 1.48   0   3     3  0.46    -0.93 0.02
  anxious   7 2047 0.67 0.86      0    0.54 0.00   0   3     3  1.09     0.26 0.02
  jittery   8 3890 0.59 0.80      0    0.45 0.00   0   3     3  1.24     0.81 0.01
  nervous   9 3879 0.35 0.65      0    0.22 0.00   0   3     3  1.93     3.47 0.01
  calm     10 3814 1.55 0.92      2    1.56 1.48   0   3     3 -0.01    -0.83 0.01
  relaxed  11 3889 1.68 0.88      2    1.72 1.48   0   3     3 -0.17    -0.68 0.01
  at-ease  12 3879 1.59 0.92      2    1.61 1.48   0   3     3 -0.09    -0.83 0.01
  
  Call: score.items(keys = my.keys, items = small.msq)
  
  (Unstandardized) Alpha:
          EA   TA  EAp  EAn  TAp TAn
  alpha 0.87 0.75 0.81 0.93 0.64 0.8
  
  Average item correlation:
              EA   TA  EAp  EAn  TAp  TAn
  average.r 0.54 0.34 0.58 0.81 0.37 0.57
  
   Guttman 6* reliability:
            EA   TA  EAp EAn  TAp  TAn
  Lambda.6 0.9 0.77 0.76 0.9 0.59 0.74
  
  Scale intercorrelations corrected for attenuation
   raw correlations below the diagonal, alpha on the diagonal
   corrected correlations above the diagonal:
          EA      TA    EAp    EAn    TAp    TAn
  EA   0.874 -0.0207  1.004 -1.006  0.218  0.168
  TA  -0.017  0.7515 -0.011  0.024  1.096 -1.140
  EAp  0.842 -0.0084  0.806 -0.618  0.360  0.246
  EAn -0.906  0.0197 -0.534  0.927 -0.067 -0.076
  TAp  0.163  0.7590  0.258 -0.052  0.638 -0.512
  TAn  0.141 -0.8837  0.198 -0.065 -0.366  0.800
  
  

References

   R Development Core Team (2012). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0.

   Revelle, W. (2013). psych: Procedures for Personality and Psychological Research. Northwestern University, Evanston, http://cran.r-project.org/web/packages/psych/. R package version 1.3.1.

Index

describe, 7

file.choose, 2

pairs.panels, 7
psych, 1, 2, 7

R function
    describe, 7
    file.choose, 2
    pairs.panels, 7
    psych package
        describe, 7
        pairs.panels, 7
        read.clipboard, 2
        read.clipboard.csv, 3
        read.clipboard.fwf, 3
        read.clipboard.lower, 3
        read.clipboard.tab, 3
        read.clipboard.upper, 3
        score.items, 1, 2, 7
    read.clipboard, 2
    read.clipboard.csv, 3
    read.clipboard.fwf, 3
    read.clipboard.lower, 3
    read.clipboard.tab, 3
    read.clipboard.upper, 3
    read.table, 2
    score.items, 1, 2, 7
R package
    psych, 1, 2, 7
read.clipboard, 2
read.clipboard.csv, 3
read.clipboard.fwf, 3
read.clipboard.lower, 3
read.clipboard.tab, 3
read.clipboard.upper, 3
read.table, 2

score.items, 1, 2, 7