Using SPSS to Understand Research and Data Analysis.

5.2 Creating New Variables using the Transform Procedure

In this section, we explain how the new variables just described were created. All of this was done using an important SPSS feature called Transform. This feature allow the researcher to modify existing raw data into more useful forms for analysis.

• 5.2a The Compute Procedure

Recall that the masc1-masc5 and fem1-fem5 variables consist of self-ratings of how descriptive employees thought individual traits (stereotyped masculine and feminine) were of themselves. While some research questions might require the use of individual items such as these, more frequently researchers combine scores on individual items in a questionnaire into one total score. Thus, we combined these individual items into a single summated rating. This is done easily enough using the Compute subroutine of the Transform procedure. To compute a new variable by summing scores on several variables, simply select Transform, Compute from the Data Editor menu (see Figure 5.4).

A Compute Variable dialog box will appear (see Figure 5.5). To name the first new variable to be created, we typed masctot in the Target Variable: box. To define the computation to be peformed, we typed masc1+masc2+masc3+masc4+masc5 in the Numeric Expression: box. After we clicked OK, SPSS added the five masculine trait scores for every employee and put the total in a new column labeled masctot (our new variable).

As mentioned, SPSS tacks this new variable in the last column of the data file. Note that even though we used the masc1-masc5 variables to generate the new masctot variable, the original variables are still retained in our file. This is convenient should we want to analyze the individual traits at some point. But more importantly, we have also gained a new variable which will be much easier to work with.

To view a video of this process, click Play below.

.We followed this same procedure to compute femtot. We also used the Compute procedure to create the new variables, nach, naff and ndom. This time, however, we typed in the formula for computing the average of the five individual scores on the original ach1-ach5, aff1-aff5 and dom1-dom5 work motive variables. See Figure 5.6 for an example of this procedure used to compute nach.

Next we turn to another Transform procedure subroutine, called Recode. This procedure allows us to create a new variable by recoding scores on another variable. This is the procedure we used to convert the continuous variables, masctot and femtot into the categorical variables, masc and fem. We also did this to create the new variables, soc and task.