ABSTRACT
The forms and parameters of fitted item response theory (IRT) can be understood as reflecting patterns that emerge from sociocognitive phenomena in circumscribed spaces of people and practices. What is the situated meaning of person-proficiency parameters θ? How might regularities arising from cognitive and social interactions be reflected in item parameters? What regularities we might expect from sociocognitive perspective might run counter to the standard IRT interpretation? The resulting perspective is applied to five areas of IRT research, namely differential item functioning, item parameter drift, instructional sensitivity, construct shift, and learning effects. Approaches to studying the implications of particular social and cognitive factors on the stability of IRT parameters and possible distortions of inferences through the model are noted.
