MetaSEM with odds ratios

Posted on Mon, 09/25/2023 - 07:18

brendan Joined: 09/21/2023

Forums

I have a series of studies that have reported their results as odds ratios. I can calculate the effect size and sampling variance and perform a series of meta-analyses for each predictor-outcome relationship. Now I would like to fit a path model to the data.

I assume that to do this, my next step is to convert the odds ratios to r values?
Can someone point me in the direction of what I need to do next to generate the required correlation matrix?

Thankyou in advance

Replied on Mon, 09/25/2023 - 20:38

Mike Cheung Joined: 10/08/2009

MASEM uses correlation

MASEM uses correlation matrices as the inputs. I am unaware of any justifications for fitting MASEM on the correlations transformed from the odds ratios.

Can you directly fit a model on the odds ratios?

Log in or register to post comments

Replied on Tue, 09/26/2023 - 02:20

brendan Joined: 09/21/2023

I am not sure how I would fit

I am not sure how I would fit a model directly on the odds ratios because I would like to test possible mediation models. The odds ratios are presented across a series of studies and I can conduct meta-analyses of the odds (for example odds of A-B, B-C, and A-C).

I haven't seen anyone try to fit MASEM to the correlations transformed from odds ratios, but I was wondering if it was at least theoretically possible to do (e.g. create a correlation matrix for each study from the transformed odds ratios)