Are There Catholic School Effects in Ontario, Canada?

Author(s): Scott Davies

Source: Davies, Scott. 2013. “Are There Catholic School Effects in Ontario, Canada?” European Sociological Review 29(4):871-883.


In the 1980s, James Coleman and his colleagues hypothesized that Catholic schools in the U.S. help generate social capital among their students and in turn increase student achievement. Met with as much skepticism as support, the question whether a Catholic school effect exists is still considered largely unanswered. A recent study by Scott Davies tests the theory in elementary schools in Ontario and finds that Catholic schools only produce a modest effect on student achievement.

About the Study
Using a large (55000+) dataset of elementary school students in the province of Ontario, Davies investigates elementary school academic outcomes in Catholic and secular schools in Ontario. Davies argues that the province of Ontario may be a better empirical test for Coleman’s ideas as these schools are more comparable in Canada, with similar governance structures, funding, and student composition. He applies sophisticated statistical methods, OLS and multi-level regression along with propensity score matching, in which he controls for demographics and kindergarten school readiness.

Small Catholic School Effect in Canada
Davies finds a modest Catholic school effect, but much smaller than Coleman’s in the U.S., ranging from 0 to .12 of a standard deviation across models. These findings also show that the Catholic school effect may be generated by family self-selection and that school choice may encourage motivated families to choose schools with good reputations, including publicly funded religious schools.

Some Limitations to Consider
The author notes that the dataset was limited in that it did not distinguish between Catholic and non-Catholic students. As it has been suggested that Coleman’s social capital effect might require a critical mass of Catholic students within a school to create the required sense of community, being able to identify composition is potentially important. While the dataset was large and fairly representative, the use of propensity matching led to a non-random sample where low-income and high-income students might both be underrepresented. Additionally, the U.S. and Canadian contexts may be more different than expected; publicly funded religious schools in Canada may be too similar to the public schools there, not offering enough variation between types of schools. The focus on primary schooling might also be problematic for testing Coleman’s theory as he studied high schools and the findings regarding this effect in elementary schools have not been as strong. 

Keywords: Catholic school effect, Canada, elementary schools

Sector: Catholic,  Public

Outcomes: Academic

Date Posted: 2013-12-21