MANCOVA?? ANCOVA?? and nested MANCOVA syntax issues

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Posts: 8
Joined: Tue Jan 06, 2009 10:56 am

MANCOVA?? ANCOVA?? and nested MANCOVA syntax issues

Postby impossible_princess » Tue Jan 06, 2009 12:35 pm


I'm looking for clarification that my analyses are the correct ones as well as help on syntax codes.

For my masters thesis I'm looking at the effects of an intervention on depression scores across three time points. In addition controlling for certain variables (i.e. covariates). Further the analysis needs to be nested within one of the groups (w1interv).

At the minute the variables looks like this:

Dependent variables (3):
- w1dassymp
- w2dasssymp_imp
- w4dassymp

Independent variables (2):
- w1interv(0, those in control schools; 1, those in intervention available schools)
- w2paqpace

Covariates (4):
- pgender
- married
- employment
- pbirth

And notably, the test of w2paqpace (those that engaged in the intervention) needs to be tested/nested within w1interv only (i.e. comparing effects of intervention engaged parents vs those who had the intervention available but chose to not attend, rather than comparing the entire sample).

Initially I was looking at conducting MANCOVA to test for significant effects...see the syntax below - I'm not sure if this is correct for the nested effects??

Secondly, would it be correct to use ANCOVA for specific analyses?? E.g. if I want to look at changes in family cohesion and family conflict - both as dependent variables.
In this case would it be best to run 2 separate analyses??
I.e. one with family cohesion as DV, then second one with family conflict as DV??
And importantly - What is the correct syntax for nested ANCOVA??

GLM w1pdasssym w2pdasssym w2pdasssym_imp BY w1interv w2paqpace WITH w1employed w1married pgender pbirth
/DESIGN=pbirth w1employed w1married pgender w1interv w2paqpace WITHIN w1interv.

Many thanks in advance.


Who is online

Users browsing this forum: No registered users and 1 guest