Philosophies, within the context of Psychometrics, refer to theoretical beliefs regarding the nature of measurement. One of these beliefs is held tight, as an orthodoxy (Rasch measurement), such that empirical associations contrary to the orthodoxy are deemed heretical and violations of measurement. Most philosophies, however, take less extreme positions regarding the nature of measurement. It is useful for you to align yourself with one of these philosophical orientations from the perspective of assessment development. You may, however, borrow procedures from each of these traditions to evaluate the quality of your measure.
We will present 4 philosophical orientations, although this is not intended to be a comprehensive or exhaustive listing. There are, no doubt, useful perspectives on the nature of measurement that have not yet been explicitly articulated.
Tradition <-c("CTT", "IRT", "Rasch", "Network Psychometrics")Philosophy <-c("observed scores a function of person 'ability' and measure error", "probability of item endorsement a function of person 'ability' & item characteristics", "deviations from empirical prediction indicate poor measurement", "manifest associations reveal substantive Psychological phenomena")phil <-as.data.frame(cbind(Tradition, Philosophy))kableExtra::kable(phil)
Tradition
Philosophy
CTT
observed scores a function of person ‘ability’ and measure error
IRT
probability of item endorsement a function of person ‘ability’ & item characteristics
Rasch
deviations from empirical prediction indicate poor measurement