I am looking for an approach to evalute the effectiveness of a rehabilitation program on back pain. Group allocation and time (two measurement points) serve as independent variables, while I also want to evaluate the specific treatment effect of a stress reduction program by including stress as additional dependent variable.
I hope this became clear and someone could help me in that matter. I already used regular MANOVA using the differences between the measurement points in back pain and stress as dependent variables, but I was hoping for another approach to analyze the data.
Kind regards,
Jahan
You say you also want to evaluate "the specific treatment effect of a stress reduction program" (on stress scores, presumably). But you've said nothing about how subjects were allocated to stress reduction treatment & control groups. Were there treatment & control groups for the stress reduction program? If not, how are you going to evaluate its effectiveness? I think you need to clarify the design before anyone can really help with the analysis. HTH.
Do I understand this correctly:
1. DV back pain
2. DV stress
doubly multivarite because you have two measurement points? Additionally, there is a between group factor?
If this is correct, you should have a look at Tabachnick & Fidell (2014) Using Multivariate Statistics. They give a whole SPSS example for such a doubly multivariate analysis. There are two ways in SPSS: either with the GLM command (which is also available via the GUI) or the MANOVA command.
Bruce you are right, there is one control group receiving TAU (control group), while the other group engages in TAU + stress reduction (intervention group).
Rainer, you understood correctly by stating the DV's and the between group factor. I will have a look at the chapter you recommended.
Thanks you a lot for your answers, both were pretty helpful!
I'm still confused about your design, Jahan. What is TAU? Is it the rehab program you mentioned? And is there another group that gets neither TAU nor the stress reduction (SR) treatment?
If there are two distinct treatments (TAU & SR), I think the best design would allocate Ss to treatment conditions using a 2x2 approach: 1 group gets neither treatment, one group gets TAU only, one group gets SR only, and one group gets both treatments. This would allow you to examine the main effects of each treatment and the interaction between them. (But I suspect the statistical horses have already left the stable--i.e., you already have the data.) (https://en.wiktionary.org/wiki/close_the_stable_door_after_the_horse_has_bolted)
Also, note that Googling on turns up things like this:
- https://www.ibm.com/support/knowledgecenter/en/SSLVMB_sub/spss/tutorials/glmr_diet_intro.html
HTH.
- Badges
- Science topic