SPSS Online Guide
The One-Way ANOVA with SPSS
The table below shows fictional data for a study of college GPA and class membership. Class refers to whether the individual reports being a first-year, sophomore, junior, or senior student. The analysis would begin by entering the data into the SPSS-Win spreadsheet as described previously. The variables can then be named and labeled as appropriate.
Once the data have been entered, click Statistics > Compare Means > One-way ANOVA.... A window will appear with a list of variables on the left and boxes labeled "Dependent List" and "Factor" on the right. Transfer the dependent variable (the variable for which means are to be computed) into the "Dependent List" box and the independent variable (the variable used as the grouping variable) into the "Factor" box. Next, click Define Range... and enter the minimum and maximum for the grouping variable. This may seem like an unnecessary step but it allows the researcher to exclude extreme values of the grouping variable. This could be desirable because too few participants were in one of the categories or a category had a label such as "Does Not Apply" or "Don't Know" in which the researcher was not interested. Once the range has been defined, click Continue. I also recommend clicking the Options... button and requesting descriptive statistics and a test for homogeneity of variance by clicking on the appropriate boxes. It may be interesting to take a look at what is available under the Contrasts... and Post Hoc... buttons which allow the researcher to select subsequent tests which may be needed to ascertain exactly where differences between and among the means may be found. Clicking Cancel takes one back to the main window without any subsequent tests being requested. Finally, click OK to cause the analysis to be computed.
Results are shown below. It begins with descriptive statistics and the results of the test for homogeneity of variance, ending with the familiar ANOVA summary table. The key information is the F-ratio and associated probability (F Prob.). In this example, the difference in GPAs among the four classes was not statistically significant as shown by the probability which is considerably more than .05. The Levene test tests the assumption that the group variances are homogenous. When the results of this test are significant, that is, the "2-tail Sig." is less than .05, the assumption has been violated. Procedures for dealing with this situation are discussed in most advanced statistics books. The solution involves adjusting the degrees of freedom for finding the critical value of F.
Test of Homogeneity of Variances
[This nonsignificant result is good because it shows that the homogeneity of variance assumption was not violated. A "Sig." value below .05 would be a cause for concern.]
[The value for "Sig." is greater than .05, therefore the result is NOT significant.]