Is sesame modeling beta or log(beta/(1-beta))?
alanlamsiu opened this issue · comments
Hi @zwdzwd,
I remember seeing the details of how sesame models beta using covariates before but I couldn't find the related information recently. I am trying to understand whether sesame is doing beta ~ beta0 + beta1 x covariate1
or log(beta/(1-beta)) ~ beta0 + beta1 x covariate1
, assuming only covariate1 is the only covariate.
Thanks.
just beta, the default modeling doesn't use M -value. I think one could transform manually if needed.