LMM - model fit is singular
I have a 2x2 repeated measures design, the dependent variable is memory performance and the fixed effects are condition (A, B) and recall time (immediate, delayed), the random effect is participant (n= 60). Each participant has therefore 4 memory performance scores. I wanted to analyse the data with JASPs linear mixed models module. When I entered the variables: dependent: memory performance, fixed effects: condition and time; random effect: participants, I get the message that the model is singular. A similar message in R with the code:
model <- lmer(memory ~ condition * time + (1 | participant), data = simu1, REML=TRUE)
led to no such message.
Can someone explain me what I'm doing wrong? Thank you very much!
Comments
By default JASP estimates all possible random effects (including those for interactions). You have to go under Model and remove (uncheck) random effects (in your case, you only have a random intercept, so only intercept should be checked):
Great, thank you very much! It worked, but only when I included the interaction effect...