"t value" associated with nlme/lme4 Posted by Rob Casson, at stats.stackexchange.com, 06 Dec 2015 I understand the problem of determining the degrees of freedom in multi-level models; hence, the decision by Doug Bates et al.