I have to make some tests for independence between several composite variables, such as BPRS-Total, BPRS-Negative, BPRS-Positive etc., and a number of explanatory variables. The problem is that these composite variables have too many levels. My questions are, can I reduce the levels by taking the average of the items included, will it be correct? And is there a standard system of cut-off values defining the patient's condition based on BPRS (24) total?
Thank you in advance!
dear Stefan,
Check these modifications out:
Article Factor structure of Bech's version of the Brief Psychiatric ...
Article A Modification of the Relative Weightings of Symptoms Utiliz...
Check out paper by Khan et al on non-parametric item response analysis of the Positive and Negative Syndrome Scale: the PANSS was developed by using the 18-item Overall and Gorham BPRS and supplementing it with 12 selected items from Singh and Kay's PRS, so this paper includes BPRS-derived items
Article The evolution of illness phases in schizophrenia: A non-para...
- Badges
- Science topic
- Similar topics
- Instrumentation Engineering
- Instruments