Non-adopters are currently coded as "one plus the last date of diffusion" so they're included in exposure calculations but removed from the regressions. This is easy to misread since summary will show 100% prevalence even when it isn't.
We should remove the last time point from the classic datasets since it isn't needed for modern analyses.
Non-adopters are currently coded as "one plus the last date of diffusion" so they're included in exposure calculations but removed from the regressions. This is easy to misread since
summarywill show 100% prevalence even when it isn't.We should remove the last time point from the classic datasets since it isn't needed for modern analyses.