SPSS Tutorial: Chi-Square Test
The aim of this tutorial, and its accompanying video, is to show you how to do a chi-square test of independence using the SPSS statistical package. The tutorial assumes you are able to import data into the software, but apart from that it is suitable for a SPSS beginner.
Part 1: The Calculation - Step by Step Instructions
The chi-square test is used to determine whether two (or more) categorical variables are associated with each other - that is, whether they are independent or dependent. As noted above, this tutorial makes the assumption that you have already imported data into the program, which means you should be looking at a screen that looks something like this (see screenshot below), though, of course, you might have data in more than two columns.
Step 1: Click Analyze on the menu bar, and then Descriptive Statistics -> Crosstabs. The crosstabs dialog will appear. You'll see the variables in your dataset in the list box on the left of the dialog. In our example, we only have two variables, but if you have more than two, you need to identify which two you want to test for independence.
Step 2: You've got to get one variable into the Row box (on the right), and the other into the Column box (also on the right). It doesn't matter which of the two variables goes into which box. Click on the first variable, and drag it to the Row box (or use the blue arrow), and then click on the second variable and drag it to the Column box.
Step 3: Set up the chi-square test by hitting the Statistics button, and selecting the Chi-square option within the dialog that appears.
Step 4: If you also want a measure of effect size, select Phi and Cramer's V in the same dialog box, and then press Continue, otherwise just press Continue.
Step 5: The next (optional) stage enables you to get expected counts as well as observed counts in the crosstabs table. Click on the Cells button on the right of the Crosstabs dialog, and then tick Expected in the Counts group on the left. Now click Continue.
Step 6: Press OK, to generate the chi-square statistic and crosstabs table.