How to Read Post-Hoc Results in SPSS

Techwalla may earn compensation through affiliate links in this story. Learn more about our affiliate and product review process here.
It is difficult to interpret post hoc results in SPSS.

Social scientists from fields such as psychology, political science and sociology use SPSS (Statistical Package for the Social Sciences) to analyze data they collect in their research. They frequently use an ANOVA (Analysis of Variance) to analyze data. If the data has three or more means, researchers use Post Hoc tests to determine which groups are significantly different from each other. For example, if you have three age groups and find a significant result with the ANOVA, you want to know which age groups are different.

Advertisement

ANOVA

Video of the Day

Step 1

Enter your data into the SPSS program or pull up data you have already entered.

Video of the Day

Step 2

Click on "Statistics" at the top of the page and select "Compare Means" then "One-Way ANOVA" from the dialog box.

Advertisement

Step 3

Select your dependent variable (for example, evaluation) from the list and push the arrow button to put it in the dependent variable list. Click on "OK."

Post Hoc Tests

Step 1

Click on "Post Hoc" at the bottom of the dialog box that appears with the output from your ANOVA.

Advertisement

Step 2

Click on one of the Post Hoc tests listed under "Equal Variances Assumed," such as Tukey, Duncan or Scheffe, if you assume there are equal variances. Select more than one test if you want to compare results. Click on "Continue."

Step 3

Click on one of the Post Hoc tests (such as Tamhane's T2 or Dunnett's T3) under "Equal variances Not Assumed," if that is the case with your data. Click "Continue."

Advertisement

Post Hoc Output

Step 1

View the Post Hoc output. The box to the left will list each of the Post Hoc tests you selected.

Advertisement

Step 2

Look at the first test, say the Tukey. Observe that each level of the independent variable is compared with each of the other levels. Say your independent variable is age groups, with three levels: 20s, 30s, 40s. Look at the column where 20s is listed. The next column will have 30s and 40s. Follow the row across to the column labeled "Sig." Sig, or significance, indicates the probability the differences are due to chance. This column will give you the significance level of the comparison of 20s and 30s, and 20s and 40s. A significance level of less than .05 indicates there are fewer than five possibilities in 100 the results are due to chance.

Advertisement

Step 3

Look at the column where 30s is listed. The next column will have 20s and 40s. Follow the row across to the column labeled "Sig." This column will give you the significance level of the comparison of 30s and 20s, and 30s and 40s.

Advertisement

Advertisement

references