Does Attendance Enhance Political Science Grades?

Abstract

This article tests a relationship between class attendance and final grades in several political science courses that I taught at the University of Georgia, University of Vermont, and University of Central Arkansas between the Fall 2000 and Spring 2006 semesters. The study employs ordinary least square estimators to test the foregoing hypothesis. Relying on a sample of up to 951 undergraduate students and controlling for grade-point average (GPA), test scores, number of credit hours, and gender, the study finds that attendance influences students' final grades. It also finds that students' GPA and tests predict final grade outcomes. Gender and students' number of credit hours are, however, not significantly related to final grades.

Keywords:

• attendance
• gender
• GPA
• grades
• political science

Acknowledgments

I would like to thank Ross Burkhart and the anonymous reviewers for their valuable comments.

Notes

∗p < 0.05; N = 951.

∗p < 0.001; two-tailed tests; Bs are unstandardized betas; standard errors are in parentheses.

∗p < 0.05; N = 131.

∗p < 0.05; two-tailed tests; Bs are unstandardized betas; standard errors in parentheses.

The main reason that I have been taking attendance in all my classes since Fall 2000 is to encourage student learning. I have all along assumed that being present in class is the first step in the learning process. However, the idea of testing the relationship between attendance and grades occurred to me only while I was giving an exam in the Spring of 2005.

The number of observations, N, varies from a correlation to a regression analysis. In the former, N corresponds to the maximum number; in the latter, N is affected by the data for each variable. For instance, students in my two seminar classes did not have to take tests; and the first three classes I taught between Fall 2000 and Spring 2001 had to take only one test. Thus, while the data on Attendance and Final Grades equal to the maximum N, the data on the other variables are not.

It seems that my students, on average, do better on Test #2 than on Test #1; they seem to have a better idea of what kind of questions I might ask on Test #2 (and I often tell them that my test formats would be the same); as a result, they seem to be more prepared to do well in the second test than they do in the first. This also holds, as it will be clear soon, with the data at the University of Central Arkansas. It then follows that students' familiarity with their professors' exam types may, in part, contribute to their success in subsequent exams.

Why Attendance impacts Test #2 but not Test #1 is not immediately clear, and a further investigation of this issue is beyond the scope of this study.

In addition, in analysis not shown here, I tested whether the interaction of Gender and Attendance affects Final Grades. In other words, I tested whether Attendance benefits each group differently in their academic performance. I found that there was little or no impact of this interactive variable on Final Grades.

See Romer (1993), for instance, for suggestions that attendance be mandatory.