Mixed Models: Error! Help!

Everything related to the development of modules in jamovi
Post Reply
RCL
Posts: 16
Joined: Sat Apr 21, 2018 11:56 pm
Location: Edinburgh, United Kingdom

Mixed Models: Error! Help!

Post by RCL »

Hi all,

First of all, thanks to the developers of jamovi for such an awesome stats package! :D

I've been trying to use the GAMLj module in jamovi to run a mixed effects model. This is for a behavioural experiment where I want to see whether anxiety scores can account for sensitivity scores in a psychophysics task.

The dependent variable is emotion sensitivity (continuous). The fixed factors here are (2) exposure time (it has 7 levels of increasing durations, so ordinal), and anxiety score (continuous, from 0 to 100). The random effect is participant. So the model is something like this:

emotion_sensitivity ~ anxiety_score + exposure_time + anxiety_score*exposure_time + (1+ exposure_time|Participant)

How can I do this with GAMLj? I've tried many times, and every time it gives me an error message:
"number of observations (=224) <= number of random effects (=224) for term (0 + exposure_time|Participant); the random-effects parameters and the residual variance (or scale parameter) are probably unidentifiable."
Please, any help. Any ideas?
Thanks!!
Renzo
User avatar
mcfanda@gmail.com
Posts: 457
Joined: Thu Mar 23, 2017 9:24 pm

Re: Mixed Models: Error! Help!

Post by mcfanda@gmail.com »

Hi
it seems that there is a problem with model itself, because it cannot be estimated. How is your design (number of observations, number of participants, etc.)?
RCL
Posts: 16
Joined: Sat Apr 21, 2018 11:56 pm
Location: Edinburgh, United Kingdom

Re: Mixed Models: Error! Help!

Post by RCL »

Hey! Thanks for your reply!

I'm actually using means, because this is a repeated-measures for which the main analysis was run on a RM Anova. They told me that t was okay to run a mixed model with means instead of trials. Do you think that could be the problem?

The number of participants is 32, and the observations is 1120, but since I'm trying to run the model on means, it would be 7 observations per subject (the number of levels of the "exposure time" variable).

Thanks!
Post Reply