Within-District Effects of Catholic Schooling on 12th-Grade Math Achievement

Abstract

Using a propensity score matching method, and regression modeling based on the 2002 Education Longitudinal Study, this study found a significant Catholic school, mathematics achievement effect among those 12th graders who were least likely to attend Catholic school. This result is evident within districts after we used the School District Demographics System map data to locate Catholic schools within school district boundaries. Furthermore, the Catholic school effects were statistically significant for students in districts that allowed publicly funded private education.

• school choice
• Catholic school
• mathematics achievement
• school district
• secondary education
• propensity score matching

Notes

1. 1. Sander and Krautmann (1995) applied the Heckman (1979) procedure on the 1980 High School and Beyond data. They found negative Catholic school effect on dropout rates, but not on years of schooling. Gamoran (1996) also used the same method on the 1988 National Educational Longitudinal Study (NELS: 88) and did not find any Catholic school effect on math test score.

2. 2. The instrumental variables (IV) used include the Catholic affiliation (Evans & Schwab, 1995), Catholic school availability, and the proportion of Catholics in the county (Neal, 1997). An extension of this IV approach can be found in Altonji, Elder, and Taber (2005a, 2005b).

3. 3. Morgan (2001) used the propensity score matching to estimate Catholic school effects and found low-income students who were less likely to attend private schools than were high-income students benefited more academically from attending Catholic schools.

4. 4. Mathematics is a subject more related to school instruction than other academic learning (Borman & D’Agostino, 1996; Murnane, 1975), and previous research has shown that students who scored higher on mathematics tests were more likely to attend competitive 4-year colleges (Hoffer, 1995). It is worth noting that the ELS contains students’ reading achievement as well, but it is only available in the base year and not in the follow-up study. Our cross-sectional analysis of reading achievement produced results consistent with the results of math achievement. For simplicity of presentation, we only report our findings on math achievement here.

5. 5. The zip code–wide entropy index H is defined as:, where is the population of the census-block group i, E and are the diversity score of zip code and census-block group i, respectively., where is the proportion of racial/ethnic group r for block group i.T denotes the zip code population, and n is the number of block groups within a zip code. Each zip code covers multiple block groups. The entropy ranges from 0 to 1, representing the least segregated to the most segregated neighborhoods.

6. 6. GIS was used to calculate the distance between the centroids of the zip codes of students’ residences and the schools these students attended. The average home-to-school distance was approximately 6 miles. Based on this result, we constructed a 6-mile radius from each centroid and counted the number of other schools within that area. Note that this variable may not capture school supply in the case where there is no choice within district and that students are required to attend a particular school.

7. 7. Our result holds for reading achievement and when prior math score (10th-grade mathematics) was removed from our logit model.