Although statistics instructors have attempted to improve the cognitive aspects of instruction, particularly with respect to assessments, relatively little attention has been paid to non-cognitive issues, including students' attitudes, feelings, beliefs, perceptions, motivations, and interests. Thus, the purpose of the present study was to determine (1) the methods of statistics assessments that students most prefer; (2) the methods of statistics assessment that students feel induce the least amount of anxiety; (3) the methods of statistics assessments that students most rate as inducing higher-order thinking; (4) the characteristics of students with the most negative overall attitudes toward statistics assessments; and (5) how students rate performance assessment and authentic assessments. Findings revealed that students appear most to prefer statistics examinations in which at least limited supporting material is permitted. Interestingly, assessments in which some form of supporting material is allowed appeared to be more popular than are examinations with no time constraints. However, examinations that are untimed and in which supporting material is allowed were regarded best as inducing the least amount of anxiety, as increasing levels of performance, and as promoting higher-order thinking. Overall, students tended to rate performance assessments the most highly. Unfortunately, authentic assessments were not rated as higly as were performance assessments. Canonical correlation analyses suggested that age, the number of college-level mathematics courses, the number of years elapses since students' last statistics class, and levels of statistics anxiety are determinants of these attitudes. The implications of these findings are discussed, and recommendations for future research are presented.
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education