Absenteeism and academic performance in an introduction to business course.
The relationship between absenteeism and academic performance was investigated among 172 students enrolled in an Introduction to Business course. The students who were absent from class on the random days attendance was taken performed significantly lower on subsequent tests. In addition, the total number of days absent from class was negatively correlated with performance on a comprehensive final exam. Finally, the number of days absent was found to be the second most important factor after GPA, in predicting student performance on the final exam. These findings suggest there may be some value in adopting intervention strategies designed to encourage attendance.
In a traditional university setting, class meetings are the primary means of delivering educational services. However, as almost anyone who has ever taught a large section of an introductory course can confirm, student attendance at these meetings is far from perfect. As educators, we want our students to attend class. We enjoy teaching and interacting with students and work hard to make our classes worthy of the students' time and energy. We know who our best students are because they attend class regularly, but we often have a hard time putting a face to the names of the students that fail. Based on this anecdotal evidence, we assume that students who attend class regularly benefit from the lectures, films, and learning activities designed to facilitate their acquisition of knowledge. Conversely, we assume that excessive absences from class results in poor academic performance. These assumptions, however, are not based on empirical evidence.
To encourage class attendance, we have always spent part of the first class meeting telling our students that it is important for them to attend class, that they will learn more if they attend class, and that students who attend class regularly generally earn higher grades for the course and vice versa. Unfortunately, when we searched the literature for empirical evidence to back up our claims, we failed to find any studies specifically examining the relationship between students' absenteeism during an Introduction to Business course and their subsequent performance on exams. Moreover, studies over the past fifteen years that have examined the relationship between absenteeism and academic achievement in related fields such as economics, finance, and operations management have produced different, sometimes contradictory, results.
Several studies have linked class absences with lower grades. For example, Brocato (1989) found a strong negative correlation between absences and grades among undergraduate students enrolled in Principles of Macroeconomics and Intermediate Macroeconomics courses he taught over a four-year period. Park and Kerr (1990) also found that attendance was a determinant of student performance in a Money and Banking course, but not as important as a student's GPA and percentile rank on the American College Test (ACT).
Browne et al. (1991), however, found that students who studied on their own did just well as students that attended a typically structured Principles of Microeconomics class on the Test of Understanding College Economics (TUCE). Their conclusion: "Apparently, instructors' classroom activities have negligible impacts on student performance, measured by multiple choice items tapping memory, application and simple analysis competencies" (Browne et al. 1991, p. 29).
Based on attendance counts taken in all undergraduate economics courses taught at three "relatively elite" universities, Romer (1993) concluded that absenteeism was rampant, with typically about one-third of the students absent from class. (This finding may explain why so many studies of absenteeism and academic performance have students who are enrolled in economics course as their subjects.) Romer also reported that regression estimates of the relation between attendance and performance in one large Intermediate Macroeconomics course suggested that attendance might substantially affect learning. Considering only students who did all of the problem sets (in order to control for the impact of student motivation to do well in the course) and controlling for prior grade point average, he found the difference in performance between a student who attends class regularly and one who attends class sporadically to be approximately one full letter grade.
Durden and Ellis (1995) also found that attendance does matter for academic achievement in a Principles of Economics course. However, their findings suggest that the effect is nonlinear, becoming important only after a student misses four classes during the semester. They concluded that what really seems to matter is excessive absenteeism.
Chan, Shum and Wright (1997) studied the effect of class attendance on student performance in a Principles of Finance course using Tobin's (1958) TOBIT model and a two-stage equation developed by Heckmen (1976, 1979). They found a significant positive relationship between attendance and student performance in the TOBIT model, but only a weak relationship between attendance and student performance in Heckman's two-stage model. Furthermore, they found that a mandatory attendance policy did not improve performance.
Lai and Chan (2000) also examined the relationship between mandatory class attendance and student performance in two sections of a Principles of Microeconomics course, one with a mandatory attendance policy and the other without one. Attendance was found to be positive and statistically significant at the 1 percent level. For every 1 percent attendance, an average student received almost a 1 percent increase in his/her course average (on a 100 percent scale). In contrast to Chan, Shum and Wright (1997) findings, however, they found a mandatory attendance policy boosted an average student's performance by 6.7 percent.
More recently, Marburger (2001) investigated the relationship between students' absenteeism during a Principles of Microeconomics course and their subsequent performance on exams. He found that students who missed class on a given day were significantly more likely to respond incorrectly to questions relating to material covered that day than students who were present. In contrast, Peters, Kethley, and Bullington (2002) found that class attendance did not affect students' exam performance in an introductory operations management course. Finally, Durden and Ellis (2003) found that class attendance and motivation were related in their study of 252 students enrolled in a Principles of Economics course. However, they note that if motivation is not controlled for, the effect of absence from class on performance may be overstated.
After reviewing the literature, we began to question our assumptions about attendance and academic performance. Like many university professors who teach large, introductory courses, we had no attendance policy. After all, taking attendance in large classes is difficult and time consuming. Moreover, we prefer to treat our students as adults who are responsible for their choices and attending class is one of those choices.
We began to wonder, however, if our laissez faire approach to attendance was actually encouraging students to miss class. Although we told our students on the first day of class that attendance was important, we kept our comments relatively brief and only repeated them after reporting the results of tests. Attendance was sometimes poor, suggesting that many students did not believe our claims about its importance.
Thus, this study set out to answer three important questions as they related to our Introduction to Business course. First, what was the extent of absenteeism? Second, how much, if any, does absenteeism affect student learning? Finally, in light of our research findings as they relate to answering the first two questions, should we make any changes to our course policies to combat absenteeism?
Subjects and Setting
The subjects were 172 undergraduate students enrolled in a large section of Introduction to Business at a medium-size, state university located in the upper Midwest. An equal number of men and women were enrolled in the course. A review of university records revealed that 70 students were classified as freshman, 80 as sophomores, 15 as juniors, and 7 as seniors. In terms of their major, forty-five of the students had declared business majors (30 in Business Administration, 6 in Marketing, 5 in Accounting, and 1 each in Economics, Finance, Human Resources Management, and Management Information Systems). Seventy-one students were non-business majors. The three most popular non-business majors for students enrolled in the course were Education (15 students), Mass Communication (13 students) and Nursing (eight students). A further 56 students enrolled in the course had yet to declare a major. Table 1 presents an academic profile of the subjects based on their high school (percentile) rankings, ACT scores, grade point averages (GPA), number of credit hours attempted, number of credit hours completed, and current credit hours (i.e., course load).
The course was taught over a 15-week semester in the spring of 2003. The teaching methods included lectures, video case presentations, and large-class discussions. At the first meeting of the class, the instructor stressed that class attendance was an important component of the educational environment and that students were expected to be present for each class session. It was further explained to the students that attendance would be randomly checked throughout the semester; however, no incentives were offered for class attendance nor were students penalized for not attending class. Thus, while class attendance was monitored, it was neither rewarded nor punished.
Academic Performance and Attendance Measures
The assessment activities in the course included four individual case write-ups (approximately 500 words in length), four tests consisting of 50 multiple-choice questions, and a comprehensive final examination consisting of 100 multiple-choice questions. Each case write-up accounted for five percent of a student's overall grade for the course and students were required to complete a minimum of two case write-ups. Each test accounted for 10 percent of a student's overall grade for the course and students were required to sit a minimum of three tests. For students exceeding the minimum assessment requirements, only the scores greater than their final exam score were included in the calculation of their overall grade for the course. Thus, the final exam was weighted at 40-60 percent of a student's overall grade for the course.
All the questions on the tests and final exam came directly from a test bank written by the authors of the textbook, which was required reading for the course. Thus, a student who had never attended class and relied exclusively on the textbook to prepare for the tests could have answered all the questions correctly.
Using an unannounced random schedule, class attendance was recorded at four points during the semester, once prior to each test. On the days attendance was taken, a sign-up sheet was circulated on which students were required to print their name and student number as well as sign their name. A head count was also taken to insure that no one signed in for a student who was absent.
Of the 172 students who enrolled in the course, 21 withdrew and three did not sit the final exam, yielding an overall retention rate of 86.05 percent. The overall class average for the course was 76.51 percent, with a high of 94.85 percent, a low of 31.50 percent, and a standard deviation of 10.88. The distribution of grades was as follows: A - 11 (7%); B - 57 (39%); C - 45 (30%); D - 20 (14%); F - 14 (11%). The overall pass rate for the course was 89.86 percent.
Table 2 presents the absenteeism rates recorded on the four random days that class attendance was taken. The percentage of the students absent on the days attendance was taken ranged from a high of 35.98 percent to a low of 30.60 percent, with an overall weighted average of 32.49 percent. However, class attendance did improve as the semester went on. Table 1 also presents the absenteeism rates by gender, which shows that male students where far more likely than female students to be absent from class.
Absenteeism and Test Performance
Students' scores on Tests 1-4 were classified into two groups based on whether they were present in class the unannounced random day attendance was taken prior to each test. We used t-tests to test for the differences in the mean test scores for those students present and absent. The results of these t tests are presented in Table 3. Clearly, the students present in class outperformed the students absent from class on each test. An examination of the range of scores also reveals that no students who were absent from class on the random day attendance was taken scored 90 percent or above (the cutoff for a grade of A) on the subsequent
test. Thus, class attendance appears to be a necessary condition for scoring an A on a test in this course. Furthermore, a student's total number of absences was negatively correlated with his or her final exam score (r = -.47, p<.0001).
The Relative Impact of Absenteeism on Performance
To assess the impact of absenteeism relative to other factors that may explain a student's final exam performance, we developed a regression model that incorporated the following factors: High School Rank, ACT score, GPA, Credit Hours Attempted, Credit Hours Earned, Current Credit Hours, Case Count, Test Count, Gender, Absent Overall and Major (Business versus non-business).
Unfortunately, the student records accessed for this study had 25 missing values for high school rank and 15 missing values for ACT score, so missing values for these variables were inputted using regression on each other along with GPA. The initial regression model parameter estimates are presented in Table 4. A cursory look at Table 4 suggests that GPA, number of absences, ACT, and high school rank are the significant explanatory variables.
Using stepwise model selection methods, we arrived at the following parsimonious model predicting final exam performance.
E(Final Exam | GPA, ACT, H.S. Rank, Absent Count) = 46.24 + 9.85*(GPA) + .744*(ACT) - .104*(H.S. Rank) - 1.85*(Absent Count) (<.0001) (.0018) (.0116) (.0014) N = 148 [R.sup.2] = .614 adj-[R.sup.2] = .603 (p-values in parentheses)
Here we can see that a student's GPA, Absent Count, ACT score, and High School Rank explain 61.4 percent of the variation in their final exam score. The partial correlations of the covariates are presented in Table 5. These partial correlations indicate that the number of days absent from class is the second most important factor after GPA in predicting student performance on the final exam.
In this study, we set out to answer three questions with regard to our Introduction to Business course: What was the extent of absenteeism? How much, if any, does absenteeism affect student learning? And considering the answers to the first two questions, should we make any changes to our course policies to combat absenteeism?
Concerning the first question, on a typical class day roughly one-third of the students enrolled in the course were not in class. This figure is comparable to the rate of absenteeism reported by Romer (1993) in economics courses and leads to the same conclusion--"absenteeism is rampant" (p 173). In regard to the second question, we found a very strong statistical relationship between absenteeism and academic performance. Specifically, the students who were absent from class on the random days attendance was taken performed significantly lower on subsequent tests. In addition, the total number of days absent from class was negatively correlated with student performance on the comprehensive final exam. Finally, after GPA, the number of days absent was the second most important factor in predicting student performance on the final exam.
Faced with similar research findings regarding the relationship between absenteeism and academic performance, Romer (1993) suggested experimenting with making class attendance mandatory. We caution against such an approach. Mandatory attendance policies, which impose an academic penalty for failure to attend class regularly, may have some unintended consequences. For example, the quality of classroom decorum may decline, due to the presence of resentful and disinterested students. As Stephenson (1994) notes, "a captive audience is not an ideal learning environment" (p.307).
Having found a significant negative relationship between absenteeism and academic performance, as educators we feel challenged to identify measures that will encourage class attendance. Approaching the problem of student absenteeism from an organizational behavior modification perspective (Luthans & Kreitner, 1985), we offer the following suggestions based on a functional analysis of the antecedents and consequences of the desired behavior--class attendance:
1. To establish the proper antecedents, at the first class meeting students should be informed of the empirical relationship between class attendance and academic performance. 2. Applying the business axiom, "What gets measured gets done," we recommend instructors take attendance at the beginning of every class session. Fortunately, there are a of variety interactive student response systems on the market today (e.g., EduCue, eInstruction, and TurningPoint) that make the task of taking attendance, even in large classes, a relatively quick and simple exercise. 3. Instructors should consider awarding bonus points for class attendance to positively reinforce the desired behavior. Beaulieu (1985), for example, found attendance rates of students enrolled in a sophomore-level, undergraduate management course were higher for those experiencing positive consequences versus those experiencing negative consequences, although the difference was not statistically significant. Alternatively, one might base a small portion of the overall course grade (e.g., 5-10 percent) on class participation. In order to eliminate the element of compulsion, the class participation mark would be dropped when computing the course grade if it were a student's lowest component grade. 4. Instructors may also want to consider giving short quizzes at the end of every class session. Again, this is relatively easy to accomplish with an interactive student response system. Such in-class assessments serve to reinforce the day's lesson as well as class attendance. To eliminate the element of compulsion, alternative homework assignments should be made available for students needing to make up any quizzes they missed due to an absence. 5. Finally, we recommend instructors maintain records on class attendance and provide feedback to students on their behavior. For example, when reporting test results, also report correlation between attendance and test performance to reinforce the message that students who attend class regularly generally perform better on tests than those who do not.
In formulating the above recommendations, we placed the emphasis on positive reinforcement, which should enhance students' learning by encouraging class attendance. It should be noted, however, that class attendance alone does not guarantee that learning will take place. Some students who attend class regularly still struggle academically. However, the best instructor, no matter how clear in providing explanations and examples, will certainly not be successful in teaching the academically challenged student who fails to show up for class. Future research should assess whether strategies designed to increase class attendance actually do so, and if so, whether academic performance improves.
Beaulieu, R. (1985). Positive and punitive contingencies and the control of absenteeism, College Student Journal, 19(2): 242-245.
Brocato, J. (1989). How much does coming to class matter? Some evidence of class attendance and grade performance, Educational Research Quarterly, 13(3): 2-6.
Browne, M. N., J. H. Hoag, et al. (1991). The impact of teachers in economic classrooms, The Journal of Economics XVII: 25-30.
Chan, K. C., C. Shum & D. Wright (1997). Class attendance and student performance in Principles of Finance, Financial Practice & Education, 7(2): 58-64.
Durden, G. & L. Ellis (2003). Is class attendance a proxy variable for student motivation in economics classes? An empirical analysis, International Social Science Review, 78(1 & 2): 42-46.
Durden, G. C. & L. V.Ellis (1995). The effects of attendance on student learning in Principles of Economics, American Economic Review, 85(2).
Heckman, J. (1976). The common structure of statistical models of truncation, sample selection, and limited dependent variables and a simple estimator for such models, Annals of Economic and Social Measurement, 5: 475-492.
Heckman, J. (1979). Sample bias as a specification error, Econometrica, 47(1): 153-162.
Lai, P. & K. C. Chan (2000). Should class attendance be mandatory?, Atlantic Economic Journal, 28(3): 377.
Luthans, F. & R. Kreitner (1985). Organizational Behavior Modification and Beyond, Glenview, Ill.: Scott, Foresman..
Marburger, D. R. (2001). Absenteeism and undergraduate exam performance, Research in Economic Education, 32(2): 99-109.
Park, K. H. & P. M. Kerr (1990). Determinants of academic performance: A multinominal logit approach, Journal of Economic Education: 101-111.
Peters, M., B. Kethley, et al. (2002). The relationship between homework and performance in an introductory operations management course, Journal of Education for Business, 77(6): 340-344.
Romer, D. (1993). Do students go to class? Should they?, Journal of Economic Perspectives, 7(3): 167-174.
Tobin, J. (1958). Estimation of relationship for limited dependent variables, Econometrica, 26(1): 24-36.
Stephenson, K. (1994). Correspondence, Journal of Economic Perspectives, 8(3): 207-208.
Woolen, T. C. (1998). Factors influencing student learning in Introductory Accounting classes: A comparison of traditional and nontraditional students, Issues in Accounting Education, 13(2): 357-373.
Daniel A. Sauers, Winona State University
Gloria J. McVay, Winona State University Brant D. Deppa, Winona State University
Table 1: Subjects' Academic Profile High ACT GPA School Rank Mean 62.93 21.81 2.51 Max 99.42 33 4.00 Min 11.39 13 0.12 SD 20.58 3.19 0.81 Credit Credit Current Hours Hours Credit Attempted Cumulative Hours Mean 40.82 38.19 15.50 Max 162 133 22 Min 12 3 10 SD 22.42 22.93 1.88 Table 2: Absenteeism Rates Day 1 Day 2 Day 3 35.98% 31.17% 31.69% Male Female Male Female Male Female 43.37% 28.40% 36.71% 25.33% 35.62% 27.54% Day 4 Overall 30.60% 32.49% Male Female Male Female 33.33% 27.69% 37.50% 27.24% Table 3: Student Test Performance N Range Mean t statistic df p-value Test 1 Present 105 94 - 42 70.63 3.103 162 .0023 Absent 59 88 - 36 64.92 Test 2 Present 106 100 - 44 77.75 4.511 152 .0001 Absent 48 82 - 48 69.83 Test 3 Present 97 90 - 36 69.15 3.412 140 .0008 Absent 45 82 - 34 62.13 Test 4 Present 93 92 - 42 68.54 4.004 132 .0001 Absent 41 86 - 30 59.66 Table 4: Parameter Estimates from Initial Regression Model Term Estimate Std Error t Ratio Prob>|t| Intercept 44.757707 9.971578 4.49 <.0001 HS Rank -0.110657 0.043247 -2.56 0.0116 ACT 0.771172 0.242873 3.18 0.0019 GPA 9.895801 1.321115 7.49 <.0001 Credit Hours Attempted 0.038679 0.053653 0.72 0.4722 Credit Hours Earned -0.013917 0.052088 -0.27 0.7897 Current Credit Hours 0.155418 0.354358 0.44 0.6617 Gender [Female] 0.371662 0.678264 0.55 0.5846 Case Count -0.127628 0.778061 -0.16 0.8699 Test Count -0.495223 1.874011 -0.26 0.7920 Absent Count -1.980165 0.602504 -3.29 0.0013 Major [Business] -0.465244 0.709629 -0.66 0.5132 Table 5: Partial Correlations of the Covariates with Final Exam Score Covariate Partial Correlations Absent Overall -0.2626 GPA 0.6038 Imputed ACT 0.2574 Imputed HS Rank -0.2092