Mixed Effects Modelling - Psycholinguistics Research

Discuss statistics related things

by ahmetakturk » Mon Jan 20, 2020 8:56 pm

Dear all,

I have conducted an experimental study in which participants reacted to certain words to classify them as existent or non-existent words in English by pressing the prespecified buttons on the keyboard. Basically, it is a lexical decision task.

I aim to see how the response times are affected by the participants' proficiency level and word frequency. There are a couple of independent variables specific to my study as well.

To this end, I have decided to use Mixed Effects Modelling as it has been suggested in psycholinguistic research and seems to have certain advantages over other analysis methods.

What I would like to ask is complicated as I have difficulty in interpreting Mixed Effects Modelling results. Could someone tell me what exactly in MEM analysis I need to "look at" to interpret the results (AIC, BIC, R-squared conditional, Random Effect LRT)? I used Participants and Words as random coefficients in the software, but I cannot quite express why I would do such a thing in my paper? I mean what advantages such process might have over other analysis methods such as regression?

Posts: 1
Joined: Mon Jan 20, 2020 8:22 pm

by mcfanda@gmail.com » Tue Jan 21, 2020 6:43 pm

You can use the GAMLj module to estimate the models you need. There you find also the AIC, BIC, R^2 and LRT you were referring to
User avatar
Posts: 179
Joined: Thu Mar 23, 2017 9:24 pm

Return to Statistics