I am using the
twang package to create propensity scores, which are used as weights in a binomial glm using
survey::svyglm. The code looks something like this:
pscore <- ps(ppci ~ var1+var2+.........., data=dt....)
dt$w <- get.weights(pscore, stop.method="es.mean")
design.ps <- svydesign(ids=~1, weights=~w, data=dt,)
glm1 <- svyglm(m30 ~ ppci, design=design.ps,family=binomial)
This produces the following warning:
In eval(expr, envir, enclos) : non-integer #successes in a binomial glm!
Does anyone know what I could be doing wrong ?
I wasn't sure if this message would be better on stats.SE, but on balance I thought I would try here first.
There's nothing wrong,
glm is just picky when it comes to specifying binomial (and Poisson) models. It warns if it detects that the no. of trials or successes is non-integral, but it goes ahead and fits the model anyway. If you want to suppress the warning (and you're sure it's not a problem), use
There is nothing wrong, computationally, but statistically you may not be doing something that makes much sense. In such a case, it is probably better to use a robust regression method, which is generally a good idea for proportional response data if your data include units with exactly 1 or exactly 0.