First, you'll need to download the data. Click on the following button, then select "Save" and choose a spot on the hard drive that you'll remember (Desktop is a good one). The data are stored in a password-protected folder, so you'll need to remember the password to access it.

The identity of each variable is described in the following file:

The analyses we'll be looking at involve comparing ratings of the man to ratings of the woman, always on the same trait. For example, we'll be comparing ratings of how generous the man was to ratings of how generous the woman was.

We're especially interested in how this difference is influenced by our two independent variables: courtesy and payment. Perhaps the man appears more generous than the woman in some conditions, and less generous in others.

Once you have the data open in SPSS, select Analyze -> General Linear Model -> Repeated Measures...

We'll need to define the factor that is "repeated." In this case, the repeated factor is gender because the same people evaluate the male character and the female character, and we're interested in looking for a difference. So, call the within-subject factor "gender" and specify that it has two levels. Then click 'Add' and then 'Define'.

Next, you'll need to specify which measure we'll be comparing across gender. Let's look at how much participants thought they would consider each character as a possible friend. If you consult the Excel file, these items are coded as "ef_friend" and "em_friend" for the female and male characters, respectively. Select ef_friend from the long list of variables and press the black arrow to move it into the box labeled 'Within-Subjects Variables.' Do the same with em_friend:

Next, you will need to identify the "between-subjects" variables. These are the independent variables in our experiment: Courtesy and Payment. Scroll down until you see these variables and then move each one into the 'Between-Subjects' box:

Press OK to run the analysis. A new window will open called "Output1" and will contain the results of the analysis. You may need to look in the Start Bar at the bottom of your screen to see the new window. Scroll down until you see the following output:

These are the results for the within-subjects comparisons. The first line (beginning "GENDER") shows the test for a main effect of gender: did participants rate the male and female characters differently in terms of making a likely friend? In the far right column, you see the p-value: .001. Yes, there was a difference. To record the difference, use the following Excel file:

The excel file lists all the variables we need to test. The one we just tested was Ex_frnd (x to stand for the comparison of f with m). In each space, write in the p-value of the test from the chart above. That will allow us to quickly focus in on any effects that are significant or near-significant.

The other columns ("G*C", "G*P", and "G*C*P") correspond to the other effects tested in the output shown above. G*C is p = .043, and indicates a signicant Gender by Courtesy interaction. The last line in the output above should say Gender*Courtesy*Payment, but the *Payment got cut off. You would enter its result (.39) into the G*C*P column above.

Questions? Stop by or email me.