Mediation with an ordinal mediator

1

Mediation with an ordinal mediator( low, normal and high) and continuous predictor and outcome, any suggested method will be appreciated.

1 Like
2

Is the mediator ordinal in essence, or just superficially so because of the way it was recorded? If the latter, you might posit a latent continuous mediator, and treat the discretized ordinal value as an interval-censored measurement of that underlying continuous quantity.

1 Like
3

The mediator was continuous, but we are interested in its ordinal case, low, normal and high.
How can I do it using R?
Thanks

4

This is starting to sound more like the sort of thing people do in SAS.

5

What if RMSE= 0.463 and P-value (Chi-square) is zero, meanning that the model does not fit in a Structiral Equation Modeling? How we address this issue?

Thanks