Mixture Rasch model

[Igor]

Dear Seol,
Could you please help me with the following problem. I used a mixture Rasch model in a clinical study where I tried to look at heterogeneity of patients’ propensity to develop co-morbid conditions. Three items (comorbid conditions) were included into analysis, namely dyslipidemia (Item 1), non-alcohol fat liver disease (Item 2), arterial hypertension (Item 3), and insulin-resistance (Item 4). A two-class solution was determined, and the patients were allocated to two classes. As it comes from the results (output is provided), for the members of the Class 1, the Item 3 (arterial hypertension) seems to be the most difficult task to “accomplish” (i.e., their “ability” to have this condition was most problematic). However, when I performed a post-hoc comparison of the frequencies of the diagnosed hypertension in the members of the two classes, it came out that the frequency of arterial hypertension was significantly more common a finding in the patients allocated to the Class 1, just opposite to what has been expected looking at the difficulty of that item (see another output). Could you please help me explaining that paradox? Maybe it is not a paradox at all, but why did it happen then? Probably, something wrong with my interpretation of the findings? As well, is there any available user guide to the MRM in Jamovi that could help to perform that analysis more knowledgably?
Thank you for your help in advance.

[seol]

Dear Igor

I can help with module bugs, but I can't help with interpreting the results.

Kind Regards
Seol