Religiosity and social trust: evidence from Canada

Abstract

Using the latest wave of the Canadian Ethnic Diversity Survey, I investigate whether religious identity and religious intensity associate with the degree to which people trust others, controlling for a wide range of characteristics. The analysis shows that Canadian Roman Catholics are appreciably less trusting than mainline Protestants, and religious nones are situated in between these two groups. With regard to religious intensity, I find that higher commitment negatively correlates with trust in unknown others for Roman Catholics. The reverse is true for Protestants. Results also show stark cross-denominational variations within Protestantism, as two highly committed denominations of Mennonite and Pentecostal are found to be the most and the least trusting religious groups in Canada. No non-Christian religious minority is found statistically significantly less trusting than Canadian Roman Catholics. Considering particularized trust in one’s neighbours and co-workers yields comparable conclusions.

Keywords:

• Generalized trust
• particularized trust
• religion
• Canada

Notes

¹ Sander (2002) maps the predetermined General Social Survey categories onto a quantitative measure as follows: never equals 0; less than once a year equals 0.5; about once or twice a year equals 1; several times a year equals 3; about once a month equals 12; two to three times per month equals 30; nearly every week equals 40; and every week or more often equals 52. To grant the same range of variation to the variable ‘Importance of religion’, its values are also rescaled to vary between 0 and 52, before using it in the regressions.

² Some scholars believe that Mennonites, along other Anabaptists denominations, should be considered outside Protestantism. No consensus can be established due to sparsity of stances in which this denomination is studied in relation to other Christian groups. Ziegler (2015) suggests that evangelical Protestant category of the American General Social Survey is a close identifier. In this paper, I follow the classification of the EDS, which includes them as a branch of Protestantism.

³ In the EDS, for both personal and household income, about 20% of data is missing due to no-response. However, the mean wage rate and household income calculated using this data did not differ much from the official reports. Still, to insure against possible biases, all the estimations have been made with and without the natural logarithm of household income as an explanatory variable. The results showed that the inclusion of this variable, which means the exclusion of observations with missing data, does not impact the variables of interest. Moreover, the EDS public-use microdata do not include the province of residence. Previous studies on regional variations in social trust in Canada document discrepancies between urban and rural areas, and between English and French Canada. The inclusion of dummies for francophone and living in non-metropolitan areas likely account for the larger part of regional differences.