The Use of Peer Modeling to Increase Self-Efficacy in Research Methods Courses

Abstract

One of the biggest challenges students face in any undergraduate methods course is a lack of confidence in their mathematical abilities, leading to a struggle for both retention of information and for continued involvement in research-based courses. In my article, I present a new approach to improving self-efficacy in undergraduate methods students and show that self-reported measures of learning improved as well as the completion of senior theses and directed research opportunities increased in the subsequent semester over previous years. Specifically, instead of teaching the course through academic articles and lessons from a textbook, I used a paper from the same course written by a former student as an exemplar of how to write a research paper. The peer-modeling approach improved student evaluations of the course immediately following the semester, increased the number of students involved in independent research subsequent to the course, and showed a high self-reported level of retention and use of research methods over a year after the course was completed.

Keywords:

• methods
• quantitative analysis
• self-efficacy
• statistics
• teaching

Notes

Students in the class asked the author a couple of questions about the paper, but because she had secured a prestigious job after graduation they were much more interested in that than the paper itself. I have no way of testing this, but I would be surprised if the visit had any effect on student self-efficacy or outcomes.

See Figure 1 for a discussion of the evaluation scores.

The exemplar paper in its original form is available at https://www.academia.edu/8396779/Presidential_Power_and_the_Bully_Pulpit_-_Student_Version. It has been updated and edited multiple times, and the new version is available upon request from the author.

I include this question because I believe it taps into clarity in general. Rather than simply being about the instructor’s grading, this question seems to look at whether students feel the assignment was clear and whether they were clear about the expectations for what a good paper looks like.

As one of the anonymous reviewers noted, Section 2 was generally higher than Section 1 across the board. My impression is that this comes from two factors. The first is that Section 2 had more students who took courses with me both before and after the research methods course. My impression is that these students were generally more positive toward my teaching style, which is why they took multiple courses with me. This would explain the small differences in two of the three questions where Section 2 was higher than Section 1. However, the largest difference was in terms of “instructor presented the material in an organized manner.” Because Section 2 was taught after Section 1, I would likely attribute this to repetition. Simply having an opportunity to go over the material once probably improved the presentation.

The full-text of all responses appears in Appendix B (see supplementary files online).

One of the anonymous reviewers pointed out that in such a small classroom, these changes could be driven by only one student. I present the percentages of responses in each category in Appendix C (see supplementary files online), and show that overall, in a course of 10 students, I generally saw 30% of students increase their scores. While this is only a few students, in a small classroom in a small liberal arts college, this is a substantial improvement from the first time teaching the course. Additionally, the first time teaching the course, at least one student in each section had a strong negative response to the course while this was nonexistent in the course where I used the exemplar. Taken together, while this sample is too small to apply conventional statistical tests it is highly supportive of the idea that there is an effect of the peer-modeling approach.

The analysis of the papers based on the rubric was done with the author blinded to the identity of the students and which semester the papers were from so as to not bias results. All efforts were taken to be consistent between papers, and the papers were all sufficiently far in the past that there would be no real way of identifying which paper was from which semester. Thanks to the anonymous reviewers for making this suggestion.