3
$\begingroup$

I am in the final stages of a meta-regression on disease prevalence data and potentially associated covariates. I've entered the data as raw proportions and am using the logit transformed event rate as my effect size. I have standardized coefficients, standard error, confidence intervals, z-values, and p-values for the intercept and covariates for each of my models. I know that I need to inverse the logit transformation for my intercept in order to be able to interpret it as a raw proportion, but my question is: do I need to conduct any type of inverse logit transformation on the regression coefficients for the covariates?

Thanks to anyone willing to offer feedback on this question--I greatly appreciate it!

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

You can exponentiate them to give you odds ratios. If you form the confidence interval on the logit scale you can then exponentiate its limits to get the limits of the CI for the odds ratio. If odds ratio is not a familiar term you might want to look for material on logistic regression on this site. You do not do exactly the same as you do for the intercept where you also convert further from odds to proportion (as you say in your question but I thought it worth confirming it).

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.