topic Profile analysis - validation of homogeneity of covariance in Statistical Procedures

Profile analysis - validation of homogeneity of covariance

Yuliya — Wed, 02 Mar 2016 19:08:56 GMT

Hi,

Could you suggest a solution to a Profile analysis when the assumption of homogeneity of covariance is violated.

I have three groups with different number of observations in every group (197, 188, 167) and nine dependent variables.

When I run the test for homogeneity of covariances, I get a significant Chi-square statistic.

Thank you in advance.

Regards,

Yuliya

Re: Profile analysis - validation of homogeneity of covariance

SteveDenham — Wed, 30 Mar 2016 12:38:53 GMT

Consider some sort of mixed modeling approach, where you can model heterogeneity of variance.

Steve Denham

Re: Profile analysis - validation of homogeneity of covariance

Yuliya — Fri, 01 Apr 2016 08:00:32 GMT

Hi Steve,

thank you very much for your reply.

It was my assignment at the University, so I have just described the problems I faced while analysing the data (violation of heterogeneity of variances) and possible decisions.

However, I have not considered mixed modelling approach, just because I had no idea about that. But thank you for your suggestion, I am quite curious so I will try to apply it.

Regards,

Yuliya