I am performing a metaanalysis where I am trying to find predictors for an ordinal response variable. Additionally, I want to perform pair-wise correlations on some of the variables. I have a confounding categorical variable that I want to remove.

In my data, there are 4 levels of variable RV, and 40 levels of the confounder CON1. The data looks like this, except I have thousands of cases.

VARNAME RV      CON1        IV1         IV2         IV3        IV4 
VARTYPE Ordinal Categorical Categorical Categorical Continuous Continuous
CASE1   1       1           M           R           49         476.23
CASE2   1       2           F           S           31         465.11
CASE3   2       2           M           R           37         411.20
CASE4   3       7           M           X           41         407.33

I don't think I can adjust means because the response variable is ordinal, so no linear regression.

Proportional odds logistic regression might be a good choice, but I don't think the output will produce "standardized" variables, that I can perform the pairwise correlations on.

How do I remove the effect of the confounder variable?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.