Determination of Diagnostic Cut-Points Using Stochastically Ordered Mixed Rasch Models
Diagnostic tests aim to discriminate between clinically normal and clinically abnormal cases. When raw scores summarizing responses to items in psychiatric or health related scales are used for diagnostic purposes, it is required that measurements are both valid and reliable. Diagnostic tests are also supposed to be simple to use with either markedly higher or lower scores among persons with the specific target disorder than among normal individuals. When diagnostic tests are used for screening in large populations, the requirement of simplicity is of course of utmost importance.
KeywordsLatent Classis Item Response Latent Trait Score Distribution Item Parameter
Unable to display preview. Download preview PDF.