This is the link that gave me the indication: https://stat.ethz.ch/pipermail/r-help/2007-March/127261.html Are there alternative ways to deal with a high count of zeros for count data with lmer?