Defining Group Membership: The Impact of Multiple versus Single Ethnic/Racial Categories on Testing Practices.

At the time of registration, students taking the Scholastic Assessment Tests (SAT) voluntarily complete the Student Descriptive Questionnaire (SDQ), which asks a variety of questions, one of which asks students to choose the ethnic/racial category that describes them. choosing only one category. The SAT Program undertook two studies aimed at evaluating the impact of allowing students to indicate more than one ethnic/racial category. Information was collected during 2 large fall administrations of the SAT I from 349,608 students. Students were asked, immediately following testing, to indicate their ethnic/racial background on their answer sheet, and multiple selections were allowed. During the fall administration, students were asked to indicate all categories that described them using essentially the same list found on the SDQ. During the second administration, students were asked to describe themselves using a two-part question, first indicating their ethnicity and then their race. Approximately 93% of test takers responded to the questions on both the answer sheet and the SDQ and about 4.5% moved from a single category to multiple ethnic/racial categories. The effect of the shift from a single category to multiple categories was further examined by comparing differential item functioning (DIF) results using the test takers from the first administration. Results indicate that the mean DIF changed very little when the groups were redefined using their responses on the answer sheet. These analyses provide preliminary information on the impact of allowing students to choose multiple ethnic/racial categories. (Author/SLD)