Dear CONN community,

I have a few covariates-of-interest with values that deviate from the GLM BLUE assumptions.

Although CONN employs a weight GLM, would it be more correct to log-transform these values and then add them to the model?

Thanks in advance,

Yuval.