Assessing the impact of research capstone preparation in the economics curriculum

Abstract

It is challenging for economics major programs to support an undergraduate research capstone where the goal is for all majors to complete a research project that creates new knowledge. Our program has found a way to achieve this for all of our majors in the face of resource constraints and varying student skill levels, and has done this without requiring econometrics or a research methods class or sacrificing quality. Our method includes having assignments that require students to use peer-reviewed economics journal articles in most upper-level economic elective courses. We describe our method and “capstone prep” assignment variants, and, using regression analysis on a data set of undergraduate economics majors from 2007 through 2018, find that particular assignments appear to significantly improve capstone outcomes.

Keywords:

• Capstone
• curriculum
• undergraduate research

JEL Code:

• A22

Acknowledgments

The authors acknowledge all their department colleagues over the years who have served as capstone mentors. They also thank the editors and two anonymous referees for their helpful comments.

Notes

1 For example, a student used a theoretical model originally designed to describe the bargaining relationship between wholesale and retail firms to analyze the effects of the Affordable Care Act on bargaining between hospitals and insurance providers.

2 For example, a student developed an econometric model to estimate PGA men’s golf winnings using the estimated marginal products associated with various performance measures to forecast rankings.

3 Mentoring capstones is not included in load for any department member except the capstone instructor. It is recognized as department service.

4 These requirements were introduced in 2011, and the first-attempt completion rate in this course increased from 91.8% to 95.2%.

5 Upper-level elective courses were chosen because these assignments fit well with the field-specific research students undertake in capstone, and to maintain the complete focus of intermediate theory courses on theory.

6 Sample assignments are available from the authors upon request.

7 We have not included six students who, due to scheduling conflicts, took the capstone as an independent study course over this time period.

8 Between 2006 and 2018 PLU’s undergraduate student body averaged 70% white.

9 According to the Integrated Postsecondary Education Data System (IPEDS), 46.51% of degrees in economics were earned by students identifying as white (with 23.3% of degrees earned by nonresident aliens), and males made up 66.4% of 2016 graduates from the largest five economics-major-producing institutions (Data USA 2016).

10 These requirements were changed in fall 2011. Before that semester, almost 30% of students entered the capstone with 8 credits or less of upper-division field courses. After the institution of these requirements, this percentage was reduced to 24% of students. The requirements played a role in this, as did the department’s change in advising to strongly encourage students with more than two semesters remaining to delay enrollment in the capstone.

11 Beyond the literature review (10%), graded components of the capstone course include a proposal (5%), an early draft focusing on the model being used (10%), preliminary (15%) and final drafts (20%) of the completed project, a public presentation of the research (10%), and peer feedback on classmates’ proposals and preliminary drafts (20%). Percentages listed are for the fall 2019 capstone course. While these percentages have varied over the years, the literature review has never represented more than 15% of the course grade, and the final draft has always been worth 30% or less.

12 An alternative measure of these changes (Instance) counts each time the capstone course was offered. Because our data include years in which the capstone was offered twice as well as years in which it was offered only once, this variable does not completely match Semester. All regressions were run using Instance in place of Semester, and the results were not affected.

13 For all regression results, we also ran an instrumental variable model utilizing heteroskedasticity in the absence of instruments, following the methodology of Lewbel (2012) and implemented using ivreg2h in Stata as in Baum and Schaffer (2019). Using this methodology, “Unlike ordinary instruments, identification is obtained even when all the elements of Z are also regressors in every model equation” and where “Z is obtained even when all the elements of Z are also regressors in every model” (Lewbel 2012, 77). All results generated p-values above 0.10 for the Sargan-Hansen test statistic, with one exception noted below.

14 Model 3 in table 3, for ArtOrd, however, had a p-value of 0.08 in the Sargan-Hansen test of endogeneity with respect to Capretake.

15 We also ran the models using the intermediate microeconomic theory course grade as a substitute for GPA. This course has 200 observations with a grade prior to the capstone. The intermediate micro results showed the same pattern of significance as GPA, and all general patterns held in terms of which assignments were statistically significant. In some cases, significance of assignments did decline marginally. The most drastic case was a decline from a 1% level of significance to a 5% level of significance. This marginal decline also may be due to the loss of observations, however.

16 Cultural factors, or differing high school experiences in other systems, also may play a role in the PARA and GRADE results for nonwhite international students, although investigating these specific effects is beyond the scope of this study.