Challenge quizzes: A unique tool for motivation and assessment

Abstract

We present a blended formative/summative assessment tool – ‘challenge quiz’ – developed to support mastery learning by students without placing undue burden on instructors. The formative nature of the challenge quiz motivates students to modify study behaviors while the summative component encourages students take responsibility for knowledge acquisition. This mastery-based testing approach serves to bring the student's objective of a quality grade in line with the instructor's objective of quality learning. Using a student survey administered to Principles of Microeconomics students at two US institutions, the characteristics and motivations of students are quantified, providing insight into student learning behaviors and outcomes. Empirical analysis indicates that students taking challenge quizzes tend to engage in more study methods and perform better than their initial in-class quiz, despite the increased rigor of the assessment. Changing study behaviors in advance of the challenge quiz had a significant positive impact on quiz grade improvement.

Keywords:

• mastery learning
• assessment
• student motivation

Notes

 1. Numerous studies show that research outcomes are the primary determinant of faculty salaries, promotion and tenure. See for example, Siegfried and White (1973a, 1973b), Katz (1973), Konrad and Pfeffer (1990) and Gomez-Mejia and Balkin (1992).

 2. There is currently no ongoing teacher training program in economics, and past efforts have varied in terms of intensity of experience and frequency. For more information on such programs see Goodman, Maier, and Moore (2003), Salemi (2010), and Salemi, Saunders, and Walstad (1996).

 3. Notable exceptions include Walstad (2006) and Walstad, Curme, Carson, and Ghosh (2010).

 4. Walstad (2006) suggests using self-graded quizzes as a formative assessment tool. While this method provides student feedback on their learning, it does not have the same incentive structure as the challenge quizzes employed herein.

 5. As with all examples of mastery learning, affording students the opportunity for grade improvement may distort the evaluation function of assessment. The parameters of the challenge quiz assessment minimize this possibility.

 6. Data were collected during three semesters at University A and two semesters at University B. Student grade point averages were also collected for all semesters at University A and during the latter semester at University B. The survey was administered during the final exam period for each class. Students who withdrew from the course prior to the final exam are therefore not included in the sample, potentially creating a sample selection issue. Because the number of withdraws is typically very small (3% or less), any sample selection bias is likely insignificant.

 7. A similar but less dramatic pattern is found for University A students. The average improvement of those that took challenge quiz 6 was lower than for other challenge quizzes (14 points compared with 18). However, only three students at University A took this final challenge quiz and the difference in average improvement is not statistically significant.

 8. While exploring potential causes of these differences is beyond the scope of this study, we recognize that they may reflect factors inherently related to dissimilarities between the two student bodies including aptitude and motivation.

 9. All students in all classes are afforded the opportunity to partake in the challenge quiz assessment. Hence, selection into the treatment (challenge quiz) group can be considered endogenous, and we cannot definitively identify exogenous factors that lead to challenge quiz decisions or outcomes. Our investigation is therefore a comparison of students who took the challenge quiz and those who did not, predicated on the existence of unobservable factors that may influence the challenge quiz decision.

10. The details for this test, as well as other methods of estimation, in the context of the challenge quiz study are provided in the Appendix.

11. OLS and logit regression models reveal qualitatively identical results in terms of factors associated with taking a challenge quiz.

12. For University A students that took two challenge quizzes, improvement is measured as the average change in score over both quizzes. Similar results were obtained when improvement is limited to the change in score for the first challenge quiz only.

13. Dummy variable for taking challenge quiz 3 at University B or challenge quiz 6 at University A.

14. For University A students that took two challenge quizzes, improvement = change in score on first challenge quiz taken.

15. When the expected value of a particular combination of x and y is exceptionally low (less than 1) the Chi square test may be inappropriate (SAS, 2008). While alternative tests are available (e.g. Fisher's exact test), a more straightforward remedy is to combine low response categories.