Which analysis?
Message posted by Holly on October 25, 2000 at 12:00 AM (ET)

I am ready to begin analyzing data I have collected and am unsure of which analysis would be best. I have three naturally-occurring groups (i.e., no experimental manipulation of an IV) and am interested in whether these groups are significantly different on several DVs, some of which are continuous and some of which are dichotomous. Additionally, I have several measures of the same construct (i.e., two different measures of depression), so some of the DVs are correlated. Is a MANOVA acceptable? Or is this procedure only for truly experimental designs? If it is appropriate, how do I choose which DVs to enter into the MANOVA?
Second, I am interested in seeing whether any of my DVs predict group membership. Several of my DVs are correlated. I do not have enough subjects to run a factor analysis or subsequent structural equation modeling.
Any suggestions would be greatly appreciated!

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Which analysis?
Message posted by JG on October 26, 2000 at 12:00 AM (ET)