1
$\begingroup$

I'm having trouble regarding which test to run on my data in SPSS.

I have two groups (condition1 and condition2) which is also my independent variable (categorical).

Both groups were asked to rate their brand attitude (dependent variable 1) and purchase intention (dependent variable 2) at two points in time (before and after an intervention). So I have repeated measures at only two points in time.

Hypothesis: the increase in brand attitude and purchase intention is greater for people in group1 (condition1) than in group2 (condition2).

I already did a paired t test for each group to see the change in brand attitude and purchase intention, but now I want to compare the groups and see if the difference in means is significant.

Can you tell me which test to run?

Improve this question
$\endgroup$
1
  • $\begingroup$ Please tell us about the sample size and the type of scale used. Were there a priori sample size estimates? Can you inform us about the data distribution? $\endgroup$
    – doug.numbers
    Commented Jun 16, 2014 at 15:17

2 Answers 2

Reset to default
1
$\begingroup$

It seems like you want to run a between-within repeated measures analysis of variance. In this case, you have two categorical variables: group and time. If you indicate that the time variable is the repeated measures variables, and you put group as a factor, you will get the test you seek. You would interpret the group*time interaction to test whether the INCREASE/DECREASE (i.e. change) in brand attitude/purchase intention differs between groups. This is likely of most interest to you.

You can also examine wither time is associated with different means across both groups (this is the time variable) and between groups collapsed across time (this is the group variable). These could be of interest, but I expect you are more interested in seeing how group membership is associated with change in the dependent variable(s).

Improve this answer
$\endgroup$
2
  • $\begingroup$ Thank you! I found a significant main effect for time, but more importantly I found a significant interaction between group*time, but I'm struggling with how to interpret the interaction. I still don't get how I can prove that the change in brand attitudes (or purchase intentions) was significantly higher for people in condition1 than for the people in condition2. The change is clear in the profile plots. Thanks again! $\endgroup$
    – Lara
    Commented Jul 6, 2014 at 15:23
  • $\begingroup$ The interaction indicates that the change in groups is significantly different. You can know which change is greater by looking at a plot, as you did. $\endgroup$
    – Behacad
    Commented Jul 6, 2014 at 15:27
1
$\begingroup$

I'm not so familiar with SPSS, but it sounds to me that you are looking for the estimated marginal means (EMMEANS in SPSS?).

See this webpage for an example.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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