Sources of Funding and Academic Performance in Economics Principles Courses

Abstract

The authors examine two factors that may affect student achievement in economics principles courses: working for pay and the primary source of funds (employer tuition reimbursement, loans, scholarships, financial aid, self-financing, parental transfers, other) used to pay for college for a sample of students in economics principles classes at a regional, nonresidential public university. After controlling for endogeneity, working for pay has no statistical effect on course performance, but the sources of education funding have differential effects on the course grade, with employer funding having a strong positive and significant influence. Students receiving employer tuition reimbursement score approximately 14 percentage points higher in the course than students funded through parental transfers.

Keywords:

• course performance
• economic principles
• employer tuition reimbursement
• financial aid
• sources of funding for higher education

JEL code:

• A22

Acknowledgments

The authors thank participants at the 2009 annual meetings of the Southern Economic Association for comments and suggestions and Whitney Ruble and the staff in the Office of Institutional Research at Indiana University Southeast for research assistance. The authors also thank Michael Hicks, John Horowitz, Tom Lambert, Eric Schansberg, Nalitra Thaiprasert, Sam Allgood, and three anonymous referees for comments on previous versions of this article.

Notes

1. The percentages in this paragraph are for students at four-year, non-doctoral-granting public institutions.

2. In the Carnegie classification, there are 165 universities classified as medium (3,000–9,999 FTE students) four-year, primarily nonresidential (less than 25 percent of students live on campus) institutions.

3. A limited number of studies have examined the effect of tuition reimbursement on employee turnover in firms, not the academic performance of employees, and found that tuition reimbursement reduces employee turnover while employees are in school but that voluntary and intended turnover increases after the earning of degrees unless employees are subsequently promoted (Pattie, Benson, and Baruch 2006; Benson 2006; Benson, Finegold, and Mohrman 2004; Capelli 2004; Finegold, Benson, and Mohrman 2002).

4. We requested SAT and ACT scores from each student's institutional file, but this information was not available for many of the students in the data set.

5. According to federal (IRS) guidelines, employers can contribute up to $5,250 per year tax-free to pay for an employee's higher education, which may include vocational, technical, or academic classes. Some businesses cap their reimbursement fee at this maximum amount. Others pay 100 percent, 75 percent, or 50 percent of tuition costs, while others compensate according to grades. Employers set rules for the type of employees and courses that qualify for reimbursement (AllBusiness 2011). Tuition and fees are around $2,500 per semester at our university. Major employers in our study area require a student to maintain good academic standing. Their employees taking classes should maintain at least a passing grade (which is C at this institution) or, in some cases, a B or better grade. Employers expecting a B or better grade reduce the reimbursement by a certain percentage if the student's grade falls below B and will not reimburse if it falls below C.

6. The National Post Secondary Aid Study (NPSAS) (NCES 2009) indicates that 6.3 percent of undergraduate students received tuition reimbursement from the student's employer during the 2007–08 academic year. The percentage of undergraduate students receiving tuition reimbursement is slightly higher (7 percent) in our sample. The higher proportion may result from the campus location in an urban area with several large national and regional companies. Cappelli (2004) reviews studies estimating the proportion of students receiving financial assistance from their employers to pay for higher education and employers offering this type of benefit.

7. The survey question asked students to identify the “main source of funds for tuition and fees.” The choices were parents, self, employer through tuition reimbursement, scholarship(s), federal aid (grants/work study), loans, other.

8. Race is often used as a proxy for socioeconomic status. We did not include a dummy variable for race in this analysis because the sample is racially homogeneous. Of the 219 students in the data set, 200 are white, and 19 are nonwhite, primarily Asian or African American.

9. To be a valid instrument, years of work experience must be highly correlated with hours worked and should only impact course performance through hours worked. In appendix A, we report regressions in which the term of hours worked per week is regressed on years of work experience and course grade is regressed on years of work experience. Work experience is a significant predictor of hours worked but not that of course grade. These results suggest that work experience is a strong instrument in that it is a statistically significant and meaningful determinant of hours worked. The estimated coefficient is 0.65 with a p-value of .001.

10. Approval for the survey was granted by the Institutional Review Board. Participation in the study was voluntary, and anonymity of respondents was preserved.

11. It could be one of the three courses Intermediate Algebra, College Algebra, or Calculus.

12. We tested for the presence of multicollinearity by estimating variance inflationary factors for these independent variables in the model. The variance inflationary factors were 1.46 (adjusted GPA), 1.16 (previous grade in mathematics), and 1.43 (prerequisite economics course grade). As the variance inflation factor is below 10 for these variables, we can conclude that multicollinearity is not an issue among these variables. The results from variance decomposition matrix for these variables were 0.468 (adjusted GPA), −0.210 (previous grade in mathematics), and −0.475 (prerequisite economics course grade). But, for an additional check, we estimated the model by omitting the adjusted GPA variable. The results from this model were statistically similar to the results obtained in table 3.