Page 1 of 1

The best way to show that a predictor is categorical?

Posted: Mon Aug 19, 2024 4:07 am
by whoopigoldberg
Hi there,

The power size is something I'm currently calculating with MLPowSim. Predictor x1 can be either a binary (1), a continuous (2), or an all MVN (3) type, and users need to indicate it in the model setup. Two things have come to mind:

Which of the three options given above is the best fit for a categorical predictor such as work satisfaction, which is measured on a 10-point Likert scalegeometry dash?

I was also wondering if you would think about taking ordinal variables and making them continuous (like a z-score) before choosing the binary choice (=1)? I would greatly appreciate your opinion on this matter.

Much obliged,
Whoopigoldberg

Re: The best way to show that a predictor is categorical?

Posted: Mon Aug 19, 2024 10:30 am
by ChrisCharlton
There is some discussion of this in the MLPowSim manual. Multiple category predictors for MLwiN are covered in section 5.2, and for R in 5.4. You may in particular want to look at section 5.4.2 for an explanation.