linear mixed model error: "missing value where TRUE/FALSE needed"

[FraStra]

Dear Jamovi Experts,

I'm trying to perform a linear mixed model with the subject ID as a random factor and reaction times, type of object, and rating as fixed factors. When I try to perform the post-hoc test on the interaction "object * rating" I get this error message "missing value where TRUE/FALSE needed." I attached the model and the event file.

Thanks!
Francesca

[mcfanda@gmail.com]

If you do a `Contingency table' between `Familiarity_rating` and `Object_material`, you see that some cells are empty. Because of that, the model you planned cannot be estimated, and so are the post-hoc. The model needs to have data (several data points) in every cell defined by the interaction.
You may want to consider `Familiarity_rating` as a continuous variable. That will save many degrees of freedom and hopefully the model may work.

[FraStra]

Hi, thanks a lot for the reply! I tried to replace the missing values with some numbers, but I got the same error. I also tried to change the type of scale of the ratings from ordinal to continuous, as you suggested, but then this variable becomes a "covariate" instead of a "factor," and the post-hoc contrasts cannot be performed... Any other suggestion that I can try?

Thank you,
Francesca