At least I've never seen textbooks/papers that are explicit about this (if yes, please). Interesting that partial model misspecification still keeps a local statistical test accurate but this reminds me of this Barr's paper on mixed-effects model 3/n http
@wgi_pr31ea @jennycz Following up on this, @dalejbarr has a helpful new article discussing slopes for interactions. http://t.co/ez3m3cfopg
@le_feufollet Barr, Levy, Scheepers and Tily (2013): http://t.co/Q5YboePYXf and Barr (2013) http://t.co/FLNm7JwzUy
Random effects structure for testing interactions in linear mixed-effects models: Dale J. Barr http://t.co/QbfkrKMhzF #psychology