The effect of announced quizzes on exam performance.

We administered announced, multiple-choice quizzes to 1 of 2 sections of a Psychology of Learning course in each of 2 consecutive semesters. We administered the quizzes between the first and second exams and discontinued them between the second and third. At the end of each semester, students completed a questionnaire about their studying and whether the quizzes helped them prepare for the exam. We found that attendance was higher in the quiz sections during the quiz portion of the semester. In both semesters, students reported that the quizzes caused them to study more and feel better prepared for the exam. However, there was no difference in exam performance between the quiz and no-quiz sections.

**********

Brief, frequently administered exams, or quizzes, serve many instructional purposes. One of the primary reasons for frequent testing is to motivate students to study on a more regular basis. Regularly scheduled exams, spaced several weeks apart, seem to produce studying behavior that is characteristic of fixed-interval (FI) schedules. In this schedule, the target behavior is virtually absent at the start of the interval and does not reach a high level until the interval is almost over. Once the interval is completed and reinforcement is collected, the behavior then stops and this "scalloped" pattern repeats itself. The study behavior of students, in which exams serve as the reinforcers (or at least the opportunity to earn reinforcement), so often resembles the FI pattern seen in the lab, that introductory psychology textbooks often use studying as an example of an FI schedule (e.g., Passer & Smith, 2001; Weiten, 1998). Upper division texts on the psychology of learning (Domjan, 2003) also use this example, and there is empirical support for this phenomenon (Mawhinney, 1971).

Quizzes may reduce the FI scallop and induce more frequent studying. More frequent studying (e.g., reading the textbook) may produce at least two important results. First, it requires the students to engage in distributed rather than massed practice. It is well established that distributed practice is more effective (Dempster, 1996). Second, if the quizzes contain material not yet covered by lecture or discussion, prior reading should make the classroom time more beneficial. Indeed, there is evidence that reading compliance among college students is not high (Sappington, Kinsey, & Munsayac, 2002) and has in fact been declining during the 1980s and 1990s (Burchfield & Sappington, 2000).

It seems intuitively obvious (at least to us and colleagues with whom we have discussed this issue), that quizzes will produce earlier and increased studying by students and thus the two results discussed previously will be achieved. However, faculty may be reluctant to give quizzes because of the belief that students find them aversive. This seems especially true for unannounced or "pop" quizzes. Thorne (2000) cited several studies confirming student's aversion for pop quizzes and found that they were made much less aversive when given for extra credit. However, he presented no data that students' exam scores improved as a result. Sporer (2001) also gave quizzes for extra credit and reported that the quizzes reduced anxiety about the exams and improved exam scores, although she presented no data. Additionally, Wilder, Flood, and Stromsnes (2001) found that extra-credit pop quizzes increased attendance and that students liked the quizzes. The students also reported that the quizzes helped them keep up with the course material. However, the quizzes did not improve exam performance. Ruscio (2001) found that randomly presented quizzes increased students' reading as defined by the percentage of students that passed the quizzes. However, he did not present any data that showed that the quizzes improved performance on scheduled exams.

There seems to be, at best, only anecdotal evidence that pop quizzes improve exam performance. There is evidence that students perform better when tested more frequently (Beaulieu & Zar, 1986; Fulkerson & Martin, 1981) but other research found that more frequent testing has no effect (Beaulieu & Frost, 1989). Marchant (2002) gave announced and unannounced quizzes and found that students gained almost 20 percentage points on the quiz scores when they were announced and reported reading the assigned articles more thoroughly if they expected a quiz. This study showed that announced quizzes can cause students to study more. Announced quizzes also increased attendance for the dates when the instructor gave the quizzes (Hovell, Williams, & Semb, 1979).

Research into the effect of regular, announced quizzes on exam performance has produced inconsistent results. Several studies have found that announced quizzes improved performance on regularly scheduled exams (Geiger & Bostow, 1976; Noll, 1939) but others have found no effect (Beaulieu & Utecht, 1987; Lumsden, 1976). One possible reason for the lack of consistency is that there are relatively few empirical studies that have addressed the effect of announced quizzes on exam performance. The purpose of the present study was to provide additional evidence on this topic.

Method

Participants

Participants were undergraduates enrolled in four sections of a Psychology of Learning class taught by the first author in two consecutive semesters In the first (fall) semester, the two sections had 20 and 17 students complete the course. In the spring semester, 23 and 21 students completed the course.

Procedure

In each semester there were three essay exams, each covering the most recent material. We gave the students a list of 6-8 essay questions several weeks before an exam. The exams consisted of one of the essay questions worth half the total test points. The students did not know which essay would be chosen. The other half of the test consisted of short essays that covered material from the remaining essays. The students had a choice of answering five out of seven short answers. The exams were graded strictly for content.

During each semester, we assigned the same textbook, the same lists of essay questions and the first author gave virtually the same lecture to both sections. During the spring semester, we adopted the latest edition of the same textbook. For the first exam, the essay question for each section was different but of comparable difficulty. Six out of seven of the short essay items were identical. Scores for this exam allowed us to compare the sections. Between the first and second exam, we administered multiple choice quizzes to one of the sections determined by a coin flip. We informed students in this section that a 10-point multiple choice quiz would be given at the beginning of each class and told them what material each would cover.

There were a couple procedural differences between the fall and spring semesters. During the fall semester, a few quizzes primarily covered material already covered in lecture or material not yet covered. Most were a mixture of both. We also informed the students that out of the seven quizzes, the lowest two would be dropped and that each quiz would contain 11 or 12 questions so there was some extra credit. There were no make-ups for quizzes.

In the spring semester, there were a total of eight quizzes, and we dropped the lowest two. An important change we instituted was that each quiz consisted solely of material that had not yet been covered in lecture. Thus in the spring semester, students could not rely on their notes and had to read ahead in the book to perform well on the quizzes.

The second author administered the second exam so the first author would be blind to section when grading the essay exams. To further insure this condition, she assigned the students a code number that was attached to the exam instead of a name. In order to administer exams of equal difficulty to each section, she administered two exams of comparable difficulty to each section. Within each section, she randomly gave each student one or the other exam. Thus, half of the students in each section took one version of Exam 2, and the other half took the second version. Student identities were not revealed until after the first author graded all the exams.

We discontinued the quizzes after the second exam and gave students in both sections a brief questionnaire to assess their opinions about the quizzes. They were then debriefed about the study.

Results

We organized our results section into four main sections. In the first section we present analyses to ascertain if the two sections of the classes used were equivalent in those factors that could potentially create a confound in the study. In the second section we present analyses to ascertain whether research trends found in past research were replicated in our study. In the third section we present analyses that will ascertain if quizzes had an impact on the student's performance on course examinations/tests. Finally, we examine the results of the survey.

Equivalence of Sections

To determine if the two sections of the classes across the two semesters were similar in academic ability, 2 (Quiz vs. No Quiz) x 2 (Fall vs. Spring semester) ANOVAs were performed comparing overall cumulative credits earned, cumulative GPA, and performance on Exam 1. None of these analyses revealed statistically significant differences (all p values > .20) for any of the main effects or interactions. In fact, the means of the quiz and no-quiz groups were virtually identical for all three of these measures. The quiz group had 80.35 earned credits, a 3.07 G.P.A. and scored 71.60 on Exam 1. In the no-quiz group, these means were, respectively, 83.35, 3.05 and 71.10. In addition, 6 out of 7 of the short answers were identical on Exam 1. The means of this part of the exam for the quiz and no-quiz sections were, respectively, 38.56 and 38.45. We concluded that these groups did not differ academically and any differences in performance between the groups on Exam 2 would be due to quizzing.

Replication of Previous Research Trends

Attendance. The mean percentage of classes attended prior to each test appears in Table 1.

In both semesters, attendance in the no-quiz sections declined from Exam I to Exam 2 by 7-8%. However, in the quiz sections, the decline was either less (fall) or actually increased slightly. Taken together, quizzes appeared to increase attendance, which is consistent with other studies (Hovell et al., 1979; Wilder et al., 2001).

Increased Studying. Students' self-reports of studying appear in Table 2.

Probably the most compelling finding in these data is that 80% of the students in the fall quiz section reported studying three or more hr per week compared to 38.9% of the students in the no-quiz section. However, in the spring semester, the amount of preparation time was more similar between the sections in that 33.3% of the students in the quiz section reported studying three or more hr per week compared to 26.4% of the students in the no-quiz section. If we assume that increased studying is equal to an increase in reading, then our data are consistent with Ruscio (2001) for the fall semester but not for the spring semester. We have no explanation for this difference between the two sections. However, in both semesters, the quizzes may have at least motivated a minimum of regular studying. In the fall and spring, 100% and 93.3% of the quiz sections, respectively, reported studying at least 1 hr per week compared to 77.8% and 73.7% of the no-quiz sections. This result is consistent with Marchant (2002), who found that quizzes increased self-report of studying.

Performance on Course Examinations/Tests

Now that we have established that the four sections of classes were similar academically, we examined whether the quizzes affected overall exam performance. We performed an independent t-test comparing exam scores of students who received quizzes prior to Exam 2 to those of students who did not. This analysis was not statistically significant, t(74) = .267, p > 0.70. The means and standard deviations are shown in Table 3. Analyses of performance on each type of essay (long versus short answer) also revealed no differences. In addition, scores on the multiple-choice extra-credit exam did not differ between quiz and no-quiz sections in either semester nor did the scores on the subset of items covering material from the quiz period (data not shown).

Student Opinion Questionnaire

Table 4 presents students' opinions about quizzes. A striking finding in both semesters is that a vast majority of students in the quiz sections thought that quizzes helped them prepare for Exam 2 (item 1) and that quizzes increased the amount of studying they did for Exam 2 compared to Exam 1 (item 2). Interestingly, although a majority of the no-quiz students thought that quizzes would have increased their studying between Exams 1 and 2, a much lower percent thought that quizzes would have helped them prepare for the second exam. In addition, a much higher percentage of the students in the quiz sections wanted quizzes between Exams 2 and 3 (item 3). One difference between the semesters was the response to item 4. In the fall, the percentage of students who felt prepared for Exam 2 was much higher in the quiz section, but this pattern was reversed in the spring semester.

Discussion

Consistent with previous research, these results show that quizzes increased attendance (Hovell et al., 1979; Sporer, 2001; Wilder et al., 2001) and self-reported studying (Marchant, 2002; Ruscio, 2001). The quizzes also resulted in increased self-report of preparation for the exam and increased studying between Exams 1 and 2 (Table 4, items 2 and 3). These results provide some support for the notion that quizzes reduce the FI scallop and promote more regular studying.

However, in spite of these results, the quizzes did not have any effect on exam performance. In addition, quiz content was not a factor. Regardless of whether quizzes covered new and old material (fall semester) or consisted entirely of material not yet covered in lecture (spring semester), they had no effect. The failure of quizzes to improve exam performance is puzzling, but it is consistent with some previous research (Beaulieu & Utecht, 1987; Lumsden, 1976).

An interesting result in both semesters is the discrepancy between the students' opinion of the quizzes and the actual effects of the quizzes. The vast majority of students in the quiz sections reported that the quizzes increased their studying and helped them prepare for the exam. However, self-report of increased studying may not reflect increased studying (Sappington et al., 2002).

It may be appropriate to speculate about why the quizzes had no effect. One possibility is that the total point value of the quizzes was too low. In the fall semester quizzes represented only 14.3% of the total grade and 16.7% in the spring. Perhaps the quizzes would have had an effect on exam performance if they were worth a higher percentage of the total points. In addition, dropping two of the quiz grades may also have reduced quiz effectiveness.

Another speculation is that the differences in format between the exams and quizzes may have contributed to these results. Because students were given the essay questions for each exam several weeks prior to it, perhaps they organized their studying by outlining each essay according to material in the book and their notes. Students would have had to read some of the text quite closely but they could have ignored other sections. On the other hand, studying for the quizzes perhaps involved a more general reading of the both the notes and the text. If these assumptions are true, then perhaps studying for the quizzes provided less benefit than we had anticipated. The merit of these speculations can certainly be tested by additional research.

In conclusion, although announced quizzes may produce some desirable effects such as increased attendance and self-report of increased preparation for exams, this study provided no evidence that quizzes do in fact improve exam performance.

Correspondence concerning this article should be addressed to Julian L. Azorlosa, Department of Psychology, 40 Peoples Bldg., West Chester University, West Chester, PA 19383; Email: [e-mail]

References

Beaulieu, R. P., & Frost, B. E. (1989). Impact of examination frequency on achievement and effect. Journal of Instructional Psychology, 16, 145-150.

Beaulieu, R. P., & Utecht, K. M. (1987). Frequently administered formative tests and student achievement. Journal of Instructional Psychology, 14, 195-200.

Beaulieu, R. P., & Zar, M.C. (1986). The effects of examination frequency on student performance. Journal of Instructional Psychology, 13, 81-89.

Burchfield, C. M. & Sappington, J. (2000). Compliance with required reading assignments. Teaching of Psychology, 27, 58-60.

Dempster, F. N. (1996). Distributing and managing the conditions of encoding and practice. In E. L. Bjork & R. A. Bjork (Eds.), Memory (pp. 317-344). San Diego: Academic Press.

Domjan, M. (2003). The principles of learning and behavior, (5th ed.). Belmont, CA: Wadsworth/Thomson Learning.

Fulkerson, F. E., & Martin, G. (1981). Effects of exam frequency on student performance, evaluations of instructor and test anxiety. Teaching of Psychology, 8, 90-93.

Geiger, O. G., & Bostow, D. E. (1976). Contingency-managed college instruction: Effects of weekly quizzes on performance on examination. Psychological Reports, 39, 707-710.

Hovell, M. F., Williams, R. L., & Semb, G. (1979). Analysis of undergraduates' attendance at class meetings with and without grade-related contingencies: A contrast effect. The Journal of Educational Research, 73, 50-53.

Lumsden, E. A. (1976). The effect of brief, pre-lecture quizzes on performance on regularly-scheduled tests, class participation and student evaluation of lectures. Journal of Instructional Psychology, 3, 40-44.

Marchant, G. J. (2002). Student reading of assigned articles: Will this be on the test? Teaching of Psychology, 29, 49-51.

Mawhinney, V. T. (1971). A comparison of students studying: Behavior produced by daily, weekly, and three-week testing schedules. Journal of Applied Behavior Analysis, 4, 257-264.

Noll, V. H. (1939). The effect of written tests upon achievement in college classes: An experiment and a summary of evidence. Journal of Educational Research, 32, 345-358.

Passer, M. W., & Smith, R. E. (2001). Psychology: Frontiers and applications. New York: McGraw-Hill.

Ruscio, J. (2001). Administering quizzes at random to increase students' reading. Teaching of Psychology, 28, 204-206.

Sappington, J., Kinsey, K., & Munsayac, K. (2002). Two studies of reading compliance among college students. Teaching of Psychology, 29, 272-274.

Sporer, R. (2001). The no-fault quiz. College Teaching, 49, 61.

Thorne, B.M. (2000). Extra credit exercise: A painless pop quiz. Teaching of Psychology, 27, 204-205.

Weiten, W. (1998). Psychology: Themes and variations (4th ed.). Pacific Grove, CA: Brooks/Cole.

Wilder, D. A., Flood, W. A., & Stromsnes, W. (2001). The use of random extra credit quizzes to increase student attendance. Journal Instructional Psychology, 28, 117-120.

Julian L. Azorlosa and Catherine H. Renner, Department of Psychology, West Chester University.

Table 1
Class Attendance in the Quiz and No-Quiz Sections

Section   Attendance Prior   Attendance Between   Attendance Between
             To Test 1         Tests 1 and 2        Tests 2 and 3

No-Quiz        90.25%              82.60%               80.83%
Quiz           91.25%              89.96%               81.66%

Table 2
Hours Studied Per Week For Quiz and No-Quiz Sections

                           Semester

                   Fall             Spring

Hrs Per Week  Quiz    No-Quiz   Quiz    No-Quiz

Less than 1    0.0%    22.2%     6.6%    26.3%
1-2 Hr        20.0%    38.9%    60.0%    47.4%
3-4 Hrs       73.3%    22.2%    33.3%    21.2%
5-6 Hrs        6.7%    11.1%     0.0%     5.3%
More than 6    0.0%     5.6%     0.0%     0.0%

Table 3
Performance on Test 2 for Quiz and No-Quiz Sections

Section   Mean    SD      N

No-Quiz   77.30   13.38   40
Quiz      76.47   13.76   36

Table 4
Student Quiz Opinion Questionnaire

Opinion Questions          Fall Semester         Spring Semester

                             Quiz     No-Quiz      Quiz     No-Quiz
                           (n = 15)   (n = 18)   (n = 15)   (n = 19)

The quizzes helped          80.00%     50.00%     73.33%     42.11%
me/would have helped
me prepare for Exam 2

The quizzes increased/      86.67%     77.78%     73.33%     68.42%
would have increased the
amount I studied between
Exam 1 and 2

I wish we had quizzes       66.67%     38.89      50.00%     26.32%
between Exams 2 and 3

I studied enough for        93.4%     76.7%       53.3       68.4
Exam 2