2.2 Exact Approaches in SPSS
2.2.1 Introduction
To perform an exact approach in SPSS, let us set an example for a research. In a preceding paragraph we suggested the question are yellow candies more often sticky than red candies? If we want to investigate this, we have two variables. The first variable is candy colour (yellow versus red) and the second variable is stickiness (sticky versus not sticky). Thus, these two variables are both categorical (with two categories).
We can create an exact probability distribution for the combination of colour and stickiness.
If SPSS offers an exact approach of the sampling distribution, the test dialog window contains an Exact button. You will find this in the dialog window for contingenxy table (Analyze > descriptive statistics > crosstabs) and in several legacy dialogs for non-parametric tests (Analyze > nonparametric tests > legacy diologs). In the Exact dialog, you check the Exact option and SPSS automatically sets an upper limit of five minutes to the execution of the command. You can leave this at five minutes.
As established, we have two categorical variables with two groups. Looking at the test selection table, we know that we use a Chi Square test to investigate these variables. Thus, within the dialog window contingency tables we select Chi-sqaure in the Statistics dialog and the option Exact in the Exact dialog.
Note a Fisher’s exact test is automatically run when there is a 2x2 contingency table, when we perform this test on larger contingency tables we need to select the Exact option, which is why we teach it to you here.