I don't think this is an Instance for roc. Were treatments randomized? RM ANOVA should work I would think.
Hi Guys. First post in this forum (actually, in any statistics forum). I am not from an English speaking country, so you may find some grammatical errors.
Here’s my problem. I’ve tried to search in a lot of places and cant find a specific answer for this question:
I’m a physician and I work with treatments for spinal cord injury (a lesion caused by spinal trauma). I work with an animal model (rats).
- I have 5 groups of rats. Each group has 8 rats (I have done de sample size calculation).
- Each group receives a slightly different treatment (and one is the control group)
- The rats are monitored by 28 days. At specific days (1,3,7,14,21,28) the rats are evaluated by a scale that goes from 0 to 21 and represent their motor functionality (0 is paraplegic, 21 is normal).
Now I need to answer this question: Witch treatment is better? What test should I use in SPSS?
Here is what I’ve made to this point.
1) ANOVA (with post roc test) between the five groups in each day of evaluation (1,3,7,14,21,28)
2) Repeated measures ANOVA (with post roc test), which tells me that there is a significant difference between the groups, however, the post hoc test don’t show me what I want to see! (It shows me the comparison between the days but I cant see what treatment is better or worse in each day)
Is there a better test to be done (like a test that compares the curves between the treatments), or should I do the test mentioned above, but in a different (more correct way)?
Thank you very much!
I hope I was not too confusing.
Lucas
I don't think this is an Instance for roc. Were treatments randomized? RM ANOVA should work I would think.
Stop cowardice, ban guns!
LucasConzatti (05-03-2015)
I suppose that allocation to groups was at random.- I have 5 groups of rats. Each group has 8 rats (I have done de sample size calculation).
It depends by what you mean by "which treatment is better". You2) Repeated measures ANOVA (with post roc test), which tells me that there is a significant difference between the groups, however, the post hoc test don’t show me what I want to see! (It shows me the comparison between the days but I cant see what treatment is better or worse in each day)
could just define a time point where you compare groups and
conclude from that which treatment is better (oneway ANOVA).
You could define some cutoff for remission or recovery and then
analyse whether (and when) groups reach this cutoff, using
Kaplan-Meier-survival analysis. Or you perorm a "mixed" ANOVA
with the repeated-measures factor "time" and the between-subjects
factor "group". The interaction of these two factors tells you whether
changes over time are different between groups.
With kind regards
K.
LucasConzatti (05-04-2015)
Tweet |