The multidimensional forced-choice item fails to control for social desirability due to the additive scoring method

The multidimensional forced-choice (MFC) item format contains multiple stimuli matched on social desirability (SD) that represent different traits. In several studies subjects have been asked to identify the more desirable stimulus in these MFC items. Since they are matched, the preference should on average be equal. But all studies show that they are not equally preferred. This study indicates that subjects then tend to identify the more applicable stimulus as the more desirable one. It is argued that the matching on SD causes the stimulus samples of the different traits to diverge in mean attractiveness or difficulty. As a result, the additive scoring of MFC items perpetuates the SD response bias.