Collaborative testing: a second chance to learn.

Abstract

This study examines the impact of collaborative testing on exam scores and students' engagement with course content. Numerous studies have documented the efficacy of collaborative learning, and several researchers have examined collaborative testing. However, few studies compare individual and collaborative testing procedures on the same exam. The results of this study demonstrate the efficacy of collaborative testing and will supplement the small body of research on this alternative testing method as compared to a traditional methodology, when used on the same exam.

Introduction

Whatever term you choose: collaborative learning, cooperative learning, active learning, peer mediated instruction, group or team learning, etc. students learning together has been widely promoted and accepted in all levels of education. Evidence of the efficacy of these pedagogical models abounds (Barkley, Cross, & Major, 2005; Michaelsen, Knight, & Fink, 2004; Johnson & Johnson, 1994; Goodsell, Maher, Tinto, Smith, & MacGregor, 1992; Johnson, Johnson, & Smith, 1991; Slavin, 1991). While some researchers are skeptical, (Bacon, 2005; Kelmett, Young & Berril, 1999; Punchocar & Fox, 2004; Stuart, 1994) group learning is rapidly moving from something alternative and non-traditional, to standard operation procedure in many classrooms. This is due in part, to the increased level of collaboration in the "real world," outside the academy, the world into which we are preparing our students to enter.

Literature Review

By the late 1980s researchers were moving beyond examining the efficacy of collaborative learning, to exploring the possible advantages of allowing students to work collaboratively on quizzes and exams. Lambiotte et al. (1987) compared four different testing methods with psychology students, including collaborative testing. They found collaborative testing resulted in higher scores due to students providing more information in their answers as compared to individual testing. In 1993, Meinster and Rose's developmental psychology students were given four exams, two of which were taken individually, and two that had collaborative testing procedures. They found that most students performed better on the collaborative format, although they did not observe a carry-over phenomenon, in other words students who may have scored high on their collaborative tests, did not subsequently score as high on their next individual exam. However, Jensen (1996) as well as Jensen, Johnson and Johnson (2002), found that administering collaborative quizzes during the semester resulted in improved scores on individually completed major exams. Zimbardo, Butler, and Wolfe (2003) noted an increase in exam scores when they compared the first exam of a course, which was taken individually, to subsequent exams that were taken collaboratively. Similar results were found by Skidmore and Aagaard (2004). Breedlove, Burkett, and Winfield (2004) observed that taking tests collaboratively raised scores, compared to exams that were given individually, for concept and knowledge questions, but did not raise scores for theory questions. While most researchers have concluded that collaborative testing is advantageous, Webb (1993) gave both individual and group exams in her math classes, and while the mean performance was higher in the collaborative testing format, ultimately she concluded that "scores from group assessment may not be valid indicators of many students' individual competence" (p. 150).

Only a limited number of researchers have compared individual and collaborative testing formats on the same exam. Applegate (1995) found that test scores improved 80% of the time when students worked in groups to complete tests that they have already completed individually. Rao and DiCarlo (2000) asked multiple choice questions during class sessions to verify whether or not students understood course content. Students first answered these questions individually, and then answered the same question a second time, after consulting with classmates. Correct answers consistently increased after students worked with their peers. Similar results were noted by Cortright, Collins, and DiCarlo (2005) and Rao, Collins, and DiCarlo (2002). Mitchell and Melton (2003) allowed their nursing students ten minutes at the end of class to work with a randomly assigned partner on exams. Students could change their original answers after working in these pairs and Mitchell and Melton found this period of collaboration resulted in an increase in student comprehension, as well as higher test scores.

Some researchers who may not have seen significant improvements in test scores with collaborative testing have found other benefits to the group testing format. For example, Russo and Warren (1999) observed that their students demonstrated improved communication skills and decreased anxiety levels with collaborative testing procedures. Additionally, Muir and Tracy (1999) found that allowing students to write essay exams collaboratively led to more reflective thinking, decreased anxiety, and a modest increase in test scores.

The purpose of this study is to examine the impact of collaborative testing on exam scores and students' engagement with course content. The results will supplement the small body of research that compares this alternative test-taking method to a traditional methodology, on the same exam.

Research Methodology

Subjects were students in two Introduction to Theatre classes during the 2004-2005 academic year. This class fulfills a Fine Arts Requirement as part of the General Education program at the State University of New York College at Brockport, a medium-size liberal arts public institution. This class is not intended for theatre majors. Typically, the majority of students are freshman and sophomores and generally equal numbers of male and female students enroll each semester. The students were told about the research study and asked to complete an informed consent form. Fall semester 151 students agreed to take part in the study and Spring semester 153 students participated.

All students were given the same four unit exams. These exams were not cumulative and consisted of 20 questions (multiple choice, matching, and true/false) that were answered using a bubble-in answer sheet. Each exam was given twice during a class period. Both Fall and Spring semesters, the first of these two attempts at each exam was administered in a traditional format, in other words no books or notes and no talking. Once all students had turned in their answer sheet from the first attempt, the class started the second attempt. (Time was not a problem as typically students completed each attempt in a little over fifteen minutes). During this second attempt all students (both Fall and Spring semesters) were allowed to consult their text books and class notes. In order to isolate student collaboration as an independent variable, students in the Fall semester class were asked to work collaboratively on this second attempt by forming groups of three to four students. Group members were not required to select the same answer choice, but they did need to discuss each question. Students did not stay in the same group for each exam. Spring semester students did not work collaboratively on their second attempt, instead they worked individually.

Since I could safely anticipate that test scores would be higher on the second attempt (when students could consult their books and notes, whether or not they worked collaboratively) I used the following grading formula to ensure, as best as possible, that exam grades would reflect individual student knowledge and effort. Grades on the exams were calculated by weighing the first attempt (the traditional format) as 75% of a student's grade on the exam, and the second attempt as 25% of the grade. I encouraged students to study as usual, despite this two-attempt exam format, by presenting a "case study" of a hypothetical student who chose not to study. If this student answered 50% of the questions correctly on the first attempt, and then scored 100% on the second attempt, the student's final grade, as it turns out, would be a 62%, or a D-.

Results

In order to examine the impact of collaborative test taking a t-test was performed, to compare the mean improvement in test scores, from the first attempt to the second attempt, between Fall and Spring semesters. The improvement in test scores was significantly higher Fall semester, when students were allowed to work collaboratively in groups on the second attempt (t = 8.970 or p<.001). It is also compelling to look at scores for individual exams. For example, during Fall semester on Exam One, 7 students (or 4.6% of test takers) scored 100% on the first try (traditional format) and during Spring semester, 4 students (or 2.6% of test takers) scored 100% on their first attempt. But on the second attempt the scores were quite different. Fall semester, when students collaborated, 83 students (or 55% of tests takers) scored 100% on Exam One, whereas during Spring semester, when students worked individually, only 18 students (or 11.8% of test takers) scored 100% on the exam. Grades on subsequent exams were similar. In other words, even though all students were allowed to consult their class notes and textbooks on the second attempt, far more students had a perfect score on their second attempt during the Fall semester, when students worked collaboratively than Spring semester, when students worked individually.

Discussion

It is important to note that due to the 75% / 25% grading formula, the increased test scores for the second attempt, in both Fall and Spring semesters, did not appreciably raise students' final grade in the class. The significant increase in test scores during Fall semester can be attributed to students working collaboratively on the exams. The benefits of this two-attempt testing format, however, went beyond improvement in test scores. In both semesters students were able to engage with the material presented on the exam two times. But student engagement was observably enhanced during Fall semester, when students collaborated on their second attempts. During these collaborative sessions students' conversations were extremely dynamic and lively. As I walked around the room students were actively defending their answer choices to group members who may have picked an alternate answer. Animated and passionate discussions about course content literally erupted for the fifteen to twenty minutes during which students collaborated on the second attempt. After handing in their exams and second answer sheets, student groups would frequently ask me which answer was correct, and upon learning the correct answer, respond with a gleeful "Yes," or turn to other group members saying, "I told you so."

The results of this study in terms of test scores, as well as the evidence observed of students' engagement with course materials, may be explained by the cognitive perspective theory of learning, which Slavin (1996) notes, "holds that interactions among students will in themselves increase student achievement for reasons which have to do with mental processing of information" (p. 48). In their book, The Collaborative Dimension of Learning, Hamm and Adams write, "when students talk and reason together to complete a task or solve a problem, they become more involved in thinking and communicating" (p. 2). Johnson & Johnson (1994), explain:

   Knowledge is social, constructed from cooperative efforts to learn,
   understand, and solve problems. Group members exchange information
   and insights, discover weak points in each other's reasoning
   strategies, correct one another, and adjust their understanding on
   the basis of others' understanding. (p. 40)

In addition to students having a second chance to engage with the material, this open-book, open-notes procedure for the second attempt had an added benefit; I no longer had to use class time going over the answers to exams. Students in both semesters were double checking their answers themselves, during the second attempt. And while I did not administer an official student satisfaction survey about this two-test format, many students noted that they enjoyed the testing procedures on their teacher evaluation forms.

This study documents several advantages for both students and instructors of collaborative testing. But further research to test students' retention of course content between individual and collaborative testing methods would be worthwhile.

References

Applegate, J. (1995). Cooperative learning in graded tests. The American Biology Teacher, 57(6), 363-364.

Bacon, D. R. (2005). The effect of group projects on content-related learning. Journal of Management Education, 29(2), 248-267.

Barkley, E. F., Cross, K. P., & Major, C. H. (2005). Collaborative learning techniques: A handbook for college faculty. San Francisco: Jossey-Bass.

Breedlove, W., Burkett, T., & Winfield, I. (2004). Collaborative Testing and Test Performance. Academic Exchange Quarterly, 8(3), 36-40.

Cortright, R. N., Collins, H. L., & DiCarlo, S. E. (2005). Peer instruction enhanced meaningful learning: Ability to solve novel problems. Advances in Physiology Education, 29(2), 107-111.

Goodsell, A., Maher, M., Tinto, V., Smith, B. L., & MacGregor, J. (1992). Collaborataive learning: A sourcebook for higher education. University Park, PA: National Center on Postsecondary Teaching, Learning, and Assessment.

Hamm, M., & Adams, D. (1992). The collaborative dimensions of learning. Norwood, NJ: Ablex Publishing Corporation.

Jensen, M. S. (1996). Cooperative quizzes in the anatomy and physiology laboratory: A description and evaluation. Advances in Physiology Education, 16(1), 48-54.

Jensen, M. S., Johnson, D. W., & Johnson, R. T. (2002). Impact of positive interdependence during electronic quizzes on discourse and achievement. The Journal of Educational Research, 95(3), 161-166.

Johnson, D. W., & Johnson, R. T. (1994). Learning together and alone: Cooperative, competitive, and individualistic learning. Boston: Allyn & Bacon.

Johnson, D. W., Johnson, R. T. & Smith, K. (1991). Active learning: Cooperation in the college classroom. Edina, MN. Interaction Book Company.

Kennett, D. J., Young, A. M., & Berrill, D. P. (1999). Is cooperative learning effective for high achieving entrance students? Implications for policy and teaching resources. Journal of Research and Development in Education, 33(1), 27-35.

Lambiotte, J. G., Dansereau, D.F., Rocklin, T. R., Fletcher, B., Hythecker, V. I., Larson, C. O., O'Donnel, A.M. (1987). Cooperative learning and test taking: transfer of skills. Contemporary Educational Psychology, 12(1), 52-61.

Meinster, M. O., & Rose, K. C. (1993, March). Cooperative testing in introductory-level psychology courses. Teaching of Psychology: Ideas and Innovations, Proceedings of the annual conference on Undergraduate Teaching of Psychology, Ellenville, NY. (ED 365 406).

Michaelson, L. K., Knight, A. B., & Fink, L. D. (2004). Team-based learning: A transformative use of small groups in college teaching. Sterling, VA: Stylus Publishing.

Mitchell, N. & Melton, S. (2003). Collaborative testing: An innovative approach to test taking. Nurse Educator, 28(2), 95-97.

Muir, S. P. &Tracy, D. M. (1999). Collaborative essay testing. College Teaching, 47(1), 33-36.

Punchoear, J. M. & Fox, P. W. (2004). Confidence in individual and group decision making: When "two heads" are worse than one. Journal of Educational Psychology, 96(3), 582-591.

Rao, S. P. & DiCarlo, S. E. (2000). Peer instruction improves performance on quizzes. Advances in Physiology Education, 24(1), 51-55.

Rao, S. P., Collins, H. L, & DiCarlo, S. E. (2002). Collaborative testing enhances student learning. Advances in Physiology Education, 26(1), 37-41.

Russo, A. & Warren, S. H. (1999). Collaborative test taking. College Teaching, 47(1), 18-21.

Skidmore, R. L. & Aagaard, L. (2004). The relationship between testing condition and student test scores. Journal of Instructional Psychology, 31 (4), 304-313.

Slavin, R. E. (1996). Research on cooperative learning and achievement: What we know, what we need to know. Contemporary Educational Psychology, 21(1), 43-69.

Slavin, R. E. (1991). Synthesis of research on cooperative learning. Educational Leadership, 8(5), 71-82.

Stuart, M. A. (1994). Effects of group grading on cooperation and achievement in two fourth-grade math classes. The Elementary School Journal, 95(1), 11-21.

Webb, N. M. (1993). Collaborative group versus individual assessment in mathematics: Process and outcomes. Educational Assessment, 1 (2), 131-152.

Zimbardo, P. G., Butler, LD., & Wolfe, V. A. (2003). Cooperative college examinations: More gain, less pain when students share information and grades. The Journal of Experimental Education, 71(2), 101-125.

Davida Bloom, SUNY Brockport

Davida Bloom, Ph.D., is an Assistant Professor in the Department of Theatre at State University of New York College at Brockport