Meeting Abstract
Different versions of multiple-choice tests were administered to undergraduate students as part of normal testing in order to evaluate whether the number of alternative answers per question affected the effectiveness of assessment. All versions of the test were equally long, with a total of 30 questions, in which ten questions had three alternative answers, ten questions had four alternative answers, and ten questions had five alternative answers (in all cases one correct answer plus distractors). Each question appeared in all three versions of the test, but with a different number of alternative answers. The point biserial statistic was used to evaluate the discrimination ability of each question. The class was an advanced undergraduate class in human physiology with two sections of >300 students each; three tests were given to each of the two sections during the academic quarter (for a total of six tests). There was no measurable improvement in assessment of student learning (using the point biserial statistic) when more than three alternative answers were provided. Therefore, there appears to be little advantage in providing more than three alternative answers per multiple-choice question, and there are disadvantages, such as needing more time for a test in which students need to consider a larger number of incorrect answers. (Partial funding for AM provided by NSF BEACON DBI=0939454)
