An item response theory analysis of two frequently used measures of identity development.

Abstract

Identity development—exploring options and making commitments—is an important process related to human functioning across the lifespan. An accurate understanding of identity development processes requires precise measures, but commonly used questionnaires have not been subject to intensive psychometric analyses. We investigated the psychometric properties of two such measures, the Utrecht Management of Identity Commitments Scale and the Dimensions of Identity Development Scale. Previous analyses have treated the response scales as interval rather than ordinal, which may not be reasonable given the measures’ Likert-type response scales. Accordingly, we evaluated their measurement precision by conducting multidimensional item response theory analyses of data from six studies of secondary and postsecondary students in The Netherlands and the United States (total N = 4,844; 36.00% boys/men, 63.07% girls/women; 0.02% nonbinary or missing gender data; 62.39% completed the measure in Dutch, 37.61% in English; 52.66% postsecondary school; 47.34% secondary school; racial, ethnic, and nationality information varied across studies). Graded response models showed that a limited range of the latent attributes was precisely measured, and the quality of items varied considerably. The measures functioned mostly similarly between gender groups, but there was substantial differential item functioning based on school level and language of the measure. We originally sought to create shortened versions, but the shortened versions provided no improvement over the low quality of the longer versions.