step fails when random effects are eliminated by reduce.random

[bbolker]

When all random effects are removed at the reduce.random stage, the resulting object is of class lm rather than merMod, and can't be converted to an lmerModLmerTest object. At a quick glance it looks like older versions may have had a fix for this case, i.e. passing the resulting object on to stats::step ...

https://stackoverflow.com/questions/55638476/how-to-do-stepwise-model-with-random-effect-lme4-lmertest

set.seed(101)
test <- data.frame(TM = factor(rep(rep(c("org","min"),each=3),3)),
                   dep = runif(18,0,20),
                   ind = runif(18,0,7),
                   dorp = factor(rep(1:3,each=6)))
library(lmerTest)
full.model <- lmer(dep ~ TM + ind + (1 | dorp),  data=test)
step(full.model)

Error in as_lmerModLmerTest(model) : model not of class 'lmerMod': cannot coerce to class 'lmerModLmerTest step.model<- lmerTest::step(full.model, direction="both",k=log(16))

[runehaubo]

It only took us 1.5 years + 10 min to fix this issue :-) Thanks for reporting!

Cheers Rune