Abstract
Much of the research on Open Educational Resources (OER) focuses on student perceptions of materials and their impact on student learning outcomes. However, in the case of large, general education courses, student interaction with course material may be predicated on their interest in the subject matter, a factor unaccounted for in the current scholarship. This research addresses the gap by exploring how student perceptions of OER vary compared to traditional, commercial textbooks as a function of subject-matter interest. A survey of two introductory American Government courses at Bowling Green State University suggests students with low interest in American government report greater interaction and engagement with the free, online text than their colleagues using a traditional commercial text. In contrast, students reporting a high interest in the American government engage with a traditional text at a greater rate than an open text. Results suggest OER may provide a way to reach students reporting a lower interest in American Government.
Acknowledgements
I would like to thank Debra Leiter, as well as the anonymous reviewers on this article for helping improve the design and scholarship.
Notes
1 The survey questions build on this previous literature and appear in Appendix 1.
2 The survey was approved by [University Name Redacted for Review] IRB on October 31, 2017. IRBNet ID# 1146714.
3 As some of the students in the course were not yet 18, these students could not complete the survey and were given an alternate assignment of equivalent length to earn the five extra credit points.
4 Appendix 2, –A2.6, presents the frequency distributions and row percentages for the summary demographic measures appearing in .
5 If we compare expected grade to grade earned, 35% of Spring students predicted they would earn As, compared to the 22% who actually earned As. In the fall, 14% of students expected an A, however only 6% actually earned the A.
6 Building on the findings of Lawrence and Lester (Citation2018) and this research, future research could explore how the approach and content of the textbooks themselves, beyond format, affects student use and interaction with the material. While both textbooks used in this research are designed for introductory courses in American Government, they are different books. While both texts cover the same general content and were used to accompany the same in-class material, the OpenStax text (Krutz Citation2016) approaches the content by encouraging students to understand what government does, who governs, and how citizens are engaged in the process. The Logic text (Kernell et al. Citation2016) approaches content by focusing on how the organization of government is designed to overcome problems of collective action, emphasizing the structure and function of governance.
7 See Appendix 1 for a list of survey questions, including the political knowledge questions.
8 Following the advice of Montgomery et al. (Citation2018), I completed further robustness checks using a pre-treatment factor as a proxy for interest. Student major functions as a proxy for interest and is recorded by the student and in the official university enrollment data for the course. While not reproduced here due to space considerations, students majoring in political science or related fields (social science majors) reported greater engagement with the traditional text compared to the non-social science majors who reported greater engagement with the OER text. These results confirm, and offer additional support, for subject matter interest conditioning how students interact with the text.
9 These results are not presented in the interest of space, but using the knowledge measure as a dependent variable and estimating an OLS regression controlling for GPA, interest, and semester indicates GPA to be the best, and only significant, predictor of performance on the knowledge measure. The same is true using alternative specifications controlling for frequency of book use and reading habits.
10 A Brant test indicates the assumptions in the ordered logit model are met and the proportional odds assumption is not violated.
Additional information
Notes on contributors
Nicole Kalaf-Hughes
Nicole Kalaf-Hughes is Associate Professor of Political Science at Bowling Green State University. She teaches courses in American politics and political institutions, research methods, and minority politics, using a variety of pedagogical approaches. Her research is in the area of American political institutions, legislative behavior, and race and ethnic politics.