currently, I am trying to calculate the powersize via MLPowSim. In terms of the Model setup, users should indicate a type for the predictor x1 involving the options binary (=1), continuous (=2) and all MVN (=3). Let me please ask two questions:
- I would like to know, which option (out of the three stated above) to indicate in case of a categorical predictor like job satisfaction (measured by a 10-point Likert-scale) for example.
- In addition, I have a general question: Would you potentially consider transforming ordinal variables into continuous variables (e.g. z-score transformation) and subsequently choose the binary option (=1)?
Sabine
