I have already done subgroup analysis, and identified a couple significant categorical covariates. I have run a meta-regression with my continuous covariates, nothing was significant. But I was hoping to add the categorical variables in with the continuous variables to see if that model explains any of my heterogeneity.

I have already dummy coded the categorical variables into multiple indicator variables. But I can't see how adding nominal categorical variables into a meta-regression can work? And also when adding ordinal categorical variables, they become non-significant in the overall model!

Please help!