To Guess or Not to Guess? – A Study of the Impact of Guessed Answers of the Adaptive Behavior Assessment System, Third Edition (ABAS-3)

The Adaptive Behavior Assessment System, Third Edition (ABAS-3) is widely recognized as a valuable tool for assessing adaptive behavior. In addition to indicating how frequently a certain behavior is displayed, respondents may also specify if they guessed their answer. Despite the author’s suggestions, it has not yet been empirically established if the presence of guessed answers represents a discriminant variable to continue scoring in the adaptive skill area that presents four or more guessed answers. Therefore, the present study aimed to establish if there are differences between protocols with guessed and no guessed answers in ABAS-3. To achieve this goal three different studies were conducted for Parent, Teacher, and Adult Forms. The Item Response Theory approach was adopted to assess the Differential Item Functioning (DIF) between participants who had given no guessed answer (NGG) and those who had given at least one guessed answer (GG). In the subscales with DIF items, these constituted less than 10% of the total items, or DIF’s effect size was negligible, therefore, all the subscales of ABAS-3 were invariant with respect to membership in NGG and GG. The present study makes an important contribution in interpreting ABAS-3 profiles with guessing answers.