3
$\begingroup$

I am wondering how interpret tables. I used a mixed ANOVA for analyzing: My within-subjects factor is time (pretest, posttest, follow-up) my between-subjects factor consists of conditions (experimental group and control group).

  • What does it mean when time and time*group are significant but Tests of Between-Subjects Effects aren’t significant?
  • What does “Intercept” mean in Tests of Between-Subjects Effects table?
  • How can I have Pairwise Comparisons table for every group separately?
$\endgroup$
  • $\begingroup$ Normally it isn't a good idea to follow an interaction with all,pairwise comparisons. The interaction tells you that differences among pre, post, and follow up differ as a function of group. What else are you interested in? For example, are you interested in whether the difference between post and follow up differ as a function of groups? $\endgroup$ – David Lane Mar 14 '17 at 0:18
0
$\begingroup$

1) Time. There's a main effect of time. There are one or more pairs among the levels that are significantly different from one another (although you could have a main effect and none of the pairs are different).

Time * group is the test of interaction. One simple way to describe is that the differences between the experimental and control groups are unequal at each level of the time variable. So the difference between experimental and control is different in the pre-test than the post-test. That's just one outcome though. There may be others.

The between-subjects test result means that when you compare the mean for the experimental to the mean for controls, hey are not different.

Typically you should interpret the interaction over the main effects especially when they are disordinal. That is, one level is higher across some or one levels of the other variable and equal or lower across the other(s).

2) If I recall correctly, you can usually ignore the Intercept, which tests whether the grand mean is = zero.

3) I don't know what software you're using so I have no idea.in SPSS you could set up contrasts.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.