Using forward search algorithm for person fit analysis in general cognitive diagnosis models

Abstract

Cognitive diagnosis models (CDMs) are psychometric models used to identify the strengths and weaknesses of examinees based on multidimensional attribute patterns. However, these latent attribute profiles may be inaccurately estimated because of an atypical test performance reflected in the response patterns. Person fit assessment is then performed to obtain information regarding the aberrant behavior of the examinees. The aim is to extend the forward search algorithm to general CDMs, particularly the G-DINA (generalized deterministic inputs, noisy and gate) model, to identify aberrant response patterns. Methods to select the initial set, and to progress and monitor the search are explored. Forward plots of goodness-of-fit statistics and Cooks distance are examined to observe drastic changes. A simulation study is conducted to determine the performance of the proposed method on different scenarios.