SEPARABILITY OF ITEM AND PERSON PARAMETERS IN RESPONSE-TIME MODELS

This paper discusses two forms of separability of item and person para meters in the context of response time (RT) models. The first is ''sep arate sufficiency'': the existence of sufficient statistics for the it em (person) parameters that do not depend on the person (item) paramet ers. The second is ''ranking independence'': the likelihood of the ite m (person) ranking with respect to RTs does not depend on the person ( item) parameters. For each form a theorem stating sufficient condition s, is proved. The two forms of separability are shown to include sever al (special cases of) models from psychometric and biometric literatur e. Ranking independence imposes no restrictions on the general distrib ution form, but on its parametrization. An estimation procedure based upon ranks and pseudolikelihood theory is discussed, as well as the re lation of ranking independence to the concept of double monotonicity.