TL;DR
This study investigates whether attending private or religious high schools impacts college academic performance at Ball State University. While religious high school attendees show a slight initial advantage in GPA, this effect diminishes significantly by junior and senior years, suggesting that any initial gains are not sustained long-term.
What Problem Does This Paper Address?
The debate surrounding school choice and the potential benefits of private versus public education is a long-standing one. Proponents of school vouchers argue that they can improve educational outcomes by allowing parents to choose better schools and by fostering competition among schools. However, research on the actual impact of private schooling on student achievement has yielded mixed results. Many studies focus on high school metrics like standardized test scores or graduation rates, which can be influenced by factors like "teaching to the test" or easier passing requirements. This paper aims to address this gap by examining a more long-term measure of educational success: college academic performance, specifically GPA. The authors seek to understand if the choice of high school, particularly private or religious institutions, has a lasting effect on how students perform in their undergraduate studies.
Methods
The researchers employed a quantitative approach using regression analysis to model the relationship between high school background and college GPA. They utilized a parsimonious linear model to focus specifically on the impact of school choice. The dependent variable is the student's Grade Point Average (GPA). Independent variables include:
- AGE: Student's age.
- HRS: Number of credits taken by the student.
- GEN: Gender (coded as 1 for female, 0 otherwise).
- REL: Indicator variable for attending a religious high school (1 if attended, 0 otherwise).
- PRIV: Indicator variable for attending a non-religious private high school (1 if attended, 0 otherwise).
- ENT: A measure of the student's performance on their entrance exam (either SAT or ACT, or the average if both were taken).
This model was initially applied to the entire undergraduate student body and then re-estimated specifically for junior and senior classes to assess the long-term impact.
Datasets
The study uses data from 15,270 undergraduate students registered at Ball State University during the Spring semester of 2002. This dataset includes information on student demographics, high school background (public, religious, private), and academic performance metrics like SAT, ACT, and GPA. The authors also reference national averages for SAT and ACT scores for comparison.
Key Insights and Ideas
- Short-term vs. Long-term Gains: The most striking insight is that while religious high school attendees show a statistically significant, albeit small, GPA advantage over public school students in the overall sample, this advantage disappears entirely for junior and senior students. This suggests that any initial academic boost from religious schooling is not sustained throughout a student's college career.
- Private School Neutrality: Attending a non-religious private high school did not show a statistically significant difference in GPA compared to public schools. This challenges the notion that all private schooling inherently leads to better college performance.
- Entrance Exam Dominance: The entrance exam score (SAT/ACT) and gender were found to be the most significant predictors of college GPA, far outweighing the impact of high school type. This highlights the importance of prior academic preparedness and inherent differences in academic aptitude.
- Potential for Socialization: The diminishing effect of religious schooling over time suggests that other factors, possibly socialization within the university environment or the public/private school students "catching up," play a role in leveling the academic playing field.
Key Results
- Students who attended religious high schools had a 0.055 to 0.073 higher GPA on a four-point scale compared to public school students when considering the entire undergraduate body.
- Students who attended non-religious private high schools showed no statistically significant difference in GPA compared to public school students.
- For junior and senior classes specifically, the REL and PRIV variables were not statistically significant, indicating no discernible difference in GPA based on high school type for these upperclassmen.
- A 1% increase in the entrance exam score led to a 2.847% increase in GPA for the entire student body.
- Female students had a GPA that was 0.30 higher on average than male students.
- The R-squared values for the regressions were around 0.27, indicating that while the model explains a significant portion of the variance in GPA, a substantial amount remains unexplained by the included variables.
Limitations and Future Work
- Generalizability: The study is limited to one university (Ball State University) in a specific region (Muncie, Indiana). The authors acknowledge that the United States may differ from this specific context and suggest that further research using panel data from various universities across different locations is needed for more generalizable conclusions.
- Selection Bias: While the authors argue that selection bias is unlikely because all students self-selected into Ball State and income was not a prohibitive factor, this is a common concern in such studies and could still influence outcomes.
- GPA as a Sole Measure: The authors concede that GPA is not a perfect measure of total educational benefits, just as SAT scores are not perfect measures of high school benefits. Other aspects of college success are not captured.
- Attrition Rates: While the study notes that GED students have a higher attrition rate, a more in-depth analysis of why students from different high school backgrounds might drop out could provide further insights.
- Policy Implications: The authors suggest that policymakers might consider educational policies aimed at changing public and private school environments to better reflect the perceived structure and discipline of religious schools. They also propose that the discussion around vouchers might benefit from focusing on other factors like per-student costs, given the small differences observed in college performance.
Summary updated