Quick Psych Question

[irish80122]

---

Members don't see this ad.

This is a small thing but it is something I want to get right.

Say you are measuring a change in depression so you do a pre and post test for depression. It makes sense that the post-test would be a dependent measure, but what is the pre-test? Would the pre-test be an independent measure? That makes sense to me but I just am not sure. Thanks

 

[psychanon]

This is a small thing but it is something I want to get right.

Say you are measuring a change in depression so you do a pre and post test for depression. It makes sense that the post-test would be a dependent measure, but what is the pre-test? Would the pre-test be an independent measure? That makes sense to me but I just am not sure. Thanks

What is the goal of this analysis? Are you comparing groups (e.g., treatment group vs. control)? If so, assigned group would be the independent variable, and you'd probably want to do an ANCOVA with pre-test as a covariate and post-test as dependent (or you could use difference scores, but ANCOVA is statistically preferable).

 

[irish80122]

All I am going for is to try to demonstrate that an activity increased negative affect as shown by the PANAS. I unfortunately do not believe I will have enough participants to have a control group so I am just trying to be as economical as I can. Therefore, we would have the pre-test, the manipulation, and then the post-test to see if positive and negative affect had changed as a result of the manipulation. I know it would be far stronger with a control group, I just don't think I will be able to field enough participants given that it is an honors thesis. It is also just part of my theory as well, I should just need to show that there is a negative shift.

I am not sure if that makes sense or not but advice is always welcome as sometimes you miss obvious things when you get bogged down planning a study.

 

[Neuro-Dr]

You are still going to need the SEM and test-retest reliability to calculate a RCI otherwise this will likely be error in a small sample.

 

[amy203]

I'm pretty sure you need to use a paired sample t-test. Without a control group, it's the only way of knowing if the difference between the scores on the two test is significant. See link:
http://www.georgetown.edu/departments/psychology/researchmethods/statistics/inferential/ttestdep.htm