Skip to content

meaning of the use.u argument in bootMer and simulateMerMod (Ben Bolker)

1 message · Simon Chamaillé-Jammes

#
I played around with Poisson-errors model, simulate() and the use.u
argument today, and when use.u is set to TRUE my confidence intervals
(here simply calculated as the interval between the 2.5 and 97.5
percentile of the bootstrapped statistics) sometimes did not include the
point estimate from the original data (this never happens with
use.u=F). 
So I would be reluctant to use this before understanding what's
happening.