Using this feedback from course "Quantitative methods" First thing to mention, there's a couple of typos (1m instead of lm and - instead of ~). I think you should be including the interaction between frequency and word length. And I think you should have some random effects: intercepts for both word and participant for sure. You could include random by-participant slopes for frequency and word_length as well -- probably it makes sense to do so since it's reasonable to expect those effects to vary by person, you can use (1+freq+word_length|participant) Please make changes to this line of code: model <- 1m(RT - freq + word_length, data= your_data_frame)