Attitudes, Behaviors, and Reasons for Changing Responses Following Answer-Changing Instruction

Contrary to the adage warning against changing test answers, mean gain from changing has been an invariant research finding. Consistency of this gain was tested for students instructed about the research results, and composition of the gain was analyzed by examining the students' reasons for changing. Students in six graduate measurement classes instructed about the answer-changing literature responded to three exams and a questionnaire. Mean gain remained positive and consistent with gain for previously studied uninstructed groups; amount of change was also stable. "Rethinking the item and conceptualizing a better answer" was the most frequent reason given for changing. "Rereading the item and better understanding the question" was the second most cited reason, followed by "rereading/rethinking" combined, and "making a clerical error." For each frequently used reason, wrong-to-right (WR) changes were in the majority. Implications for research and practice are discussed.