Select by clicking on "Conditional Probabilities". Conditional probabilities of observing adjacent categories. These are a series of Rasch dichotomous ogives. The intercepts with 0.5 probability are the Rasch-Andrich thresholds. They can be disordered relative to the latent variable.
You are looking at the conditional probability plot. Please ignore this plot unless you are concerned about this form of the Rasch model:
log (Pnij / Pni(j-1)) = Bn - Di- Fj
If you are concerned about this form of the model, then the lines on the conditional probability plot are the parallel dichotomous ogives for the relationship:
Plotted line = Pnij / ( Pnij + Pni(j-1) )
Buttons are described in Graph window.
