Measured Proficiency of Ethnic Groups in Canada

Jason Malloy and I have individually collected a large number of papers and research reports from countries around the world detailing ethnic and racial differences. I have sent some of the stuff to Richard Lynn, lost a number of reports due to hard drive failures, and simply haven’t got around for various reasons (time, health, other priorities, etc.) to posting on the remainder. In response to an article by Chanda Chrisala, James Thompson recently suggested that it would be informative to look at ethnic differences in other American countries. As such, I will comment on a few studies from Canada and Brazil. Regarding Canada, there seems to be no published detailed ethnic data for the nation as a whole — though many reports discuss the Aboriginal/overall Canada gap. The country has a number of national longitudinal surveys which most likely contain the relevant variables, but as far as I am aware no has looked into the issue. Nonetheless, since the 1980s the Toronto public schools have published research reports which decompose math and reading pass rates by linguistic, ethnic, and racial background.

Here I will summarize Brown et al. (2008), Brown et al. (2014), Presley and Brown (2011), and Yau et al. (1992) in table form. Brown and Siney (2008) report 2006 reading, math, science, and writing attainment level rates for grades 7 and 8 as well as reading, math, science, and geography rates for grade 9, and Ontario Secondary School Literacy Test (OSSLT) pass rates for grade 10. I converted the grades 7 to 9 level 3 and 4 math and reading and grade 10 OSSLT pass rates into standardized differences using the method of thresholds and then normed them on the Toronto White (European) mean. As a result of the heterogeneity in the sample, the standard deviations would have been inflated by about 20%, reducing the White/Other gaps by the same amount. I did not make corrections for this or compute composite scores. The OSSLT seems to be highly language loaded; moreover the pass rates were based on all 10th graders who could have taken the exam (those who were absent for it, who deferred it, and who took it). On account of both of these factors, this is a questionable index. Thus, I excluded it when computing average scores. Brown et al. (2015) report 2012 grade 7 and 8 reading, math, science, and writing attainment level rates as well as grade 10 OSSLT pass rates for Afghans, Spanish language speakers, Portuguese language speakers, Somalis, and Aborigines. White rates were not provided, but rather Toronto district rates were. As such I estimated the 2010/2012 White rates based on Brown and Siney’s (2008) report. I computed the percent Toronto district over percent White rates for 2006 and used this to estimate 2010/2012 White rates given the 2010/2012 district ones. For OSSLT I reported standardized differences computed using the estimated 2012 White rates, which came out lower than the 2006 ones (less plausible), and using the 2006 rates (more plausible, since it’s more likely that increased immigration depressed the 2012 Toronto OSSLT rates than that the White ones dropped; for Ontario as a whole, the 2006 to 2012 rates stayed the same and were commensurate with the 2006 Toronto White rates). Presley and Brown (2011) reported 2006 and 2010 math rates for grade 6 Portuguese language speakers. To summarize the results, I employed the same strategy as above. Yau et al. (1993) reported 1991 proficient rates for all students in the Toronto public school system. I used their “high in both English and Math rates” from their Table 4 for ethnic and language groups and from Figure 23 for racial divisions. This strategy generally had the effect of inflating gaps, but I did not feel like combining scores across ability levels. For scores based on Brown et al. (2014), Presley and Brown (2011), and Yau et al. (1992), I reported sample sizes in parentheses next to the groups.

Overall, the results confirm the existence of a Black/White gap in Canada. The deficit exists for African born, Caribbean born, and Canadian born self-identified Blacks. The 2006 results also show little to no South Asian/White differences, consistent with findings from the U.K and continental Europe. As found elsewhere, (north) East Asians generally perform better than Whites. Between 2006 and 2012 the proficiency of Somali speakers improved some. This could be because a greater percent were Canadian born in 2012 (82%) than 2006 (not reported) or because in 2008 the Toronto school district passed and has since acted on an initiative “for improving the success of Black students”. Since percent pass rates are flexible, with dedicated effort, it’s possible to raise them significantly, especially when subgroups are small. Alternatively, the difference could owe to sampling. White subgroups such as Italians performed low in 1991. It is not clear why. To note, the 1991 results are not directly comparable with the 2006 ones. Also, It’s not clear how well the scores differences would generalize across Canada.

References

Brown, R. S., & Sinay, E. (2006). student census: Linking demographic data with student achievement. Toronto District School Board Research Report. Organizational Development Department, Research and Information Services. (pdf)

Brown, R., Newton, L., & Tam, G. (2015). The Toronto District School Board’s student group overviews: Aboriginal heritage, Afghan, Portuguese-speaking, Somali-speaking, and Spanish-speaking students. (Research Report No. 14/15-31). Toronto, Ontario, Canada: Toronto District School Board (pdf)

Presley, A., & Brown, R. S. (2011). Portuguese-speaking students in the TDSB: An overview (No. 11/12, p. 01). Report. (pdf)

Yau and others (1992.The 1991 Every Secondary Student Survey, Part III:Program Level and Student Achievement No. 205.Toronto Board of Education (Ontario) .ResearchDept. (pdf)