Black students and high school completion in Quebec and Ontario: a multivariate analysis.

Livingstone, Anne-Marie ; Weinfeld, Morton

CONCERNS ABOUT RACISM IN CANADIAN schools and the particular experiences of black students have long attracted the attention of researchers and led to a number of qualitative and quantitative investigations over the years. Qualitative studies have explored the impact of racial discrimination and the role of black parents on black students' engagement in school (Codjoe 2001, 2007; Dei et al. 1997; Howard 2014; James 2011; Kanoute et al. 2014; Kelly 1998; Lafortune and Balde 2012; Solomon and Palmer 2004). Quantitative studies have concentrated on the nature and causes of racial achievement gaps, including two major investigations on black students in Quebec (Balde, Sene, and McAndrew 2011; McAndrew, Ledent, and Ait-Said 2008), and an inquiry on high school completion among immigrant students in the Toronto District School Board (TDSB; Anisefet al. 2010).

Altogether, fewer quantitative studies have been undertaken on racial inequalities in high school performance than on postsecondary education. This can be linked, in part, to the absence of national survey data on adolescents in Canada, in contrast to survey data on adult populations, which are readily available through sources such as the census and the Survey of Labour and Income Dynamics. In the one national survey on Canadian adolescents, the National Longitudinal Survey of Children and Youth, the sample of racial and ethnic minority children remains too small to support statistical comparisons. Researchers interested in investigating racial disparities in high school outcomes must turn to school-level records. While indispensable to research in this vein, school-level data can be time-consuming to secure and consolidate, and possess their own data limitations. Information on race and family income, for example, is generally not stored in school databases and must be extracted through other means.

The findings accumulated to date indicate that black students in Canada are, on the whole, less likely than students in the general population to graduate high school on time (Anisef et al., 2010; McAndrew, Ledent, and Ait-Said 2006). In addition, there is a substantial degree of within-group variability that can be found among black students, with graduation rates varying as a function of income, gender, birthplace, and language. Caution is therefore warranted in interpreting the situation of "black students" as a single group, given the considerable heterogeneity in the historical, cultural, linguistic, and class backgrounds of black families and their children (Livingstone and Weinfeld 2015).

The present paper utilizes the 2006 Canadian Census to examine the high school completion of black students in Ontario and Quebec. The aim is to explore the factors that explain why some black students graduate on time while others are delayed or dropout. The study compares the rate of high school completion of black students who were 18 to 19 years of age and living at home in 2006 with similar groups of white and other racial minority students. The study builds on and extends the research of McAndrew et al. (2006, 2008, 2013) and Anisef et al. (2010) by using the census to expand the range of demographic variables considered, namely income, family structure, race, gender, language, nativity, and immigration. Given its reliance on the census, the study cannot take into account the role of within-school variables. The evidence shows that conditions in schools can either exert a positive or negative impact on black students' motivation and engagement in school (Dei et al. 1997; Jordan, Lara, and McPartland 1996; Livingstone, Celemencki, and Calixte 2014; McAndrew 2015). Studies indicate that black students in Canadian schools report fewer positive relationship with their teachers than white students, and believe that their group receives unfair disciplinary treatment and tougher penalties for transgressing the school's behavioral rules (Fitzpatrick et al. 2013; Ruck and Wortley 2002; Solomon and Palmer 2004).

Across Canada, black families are more likely than other families to be low-income, even when family structures are the same. While blacks of the second generation have attained close to economic parity with native whites, black immigrants, blacks of the third generation or more, and black families with young children have seen their incomes fall or lag in relation to the average (Attewell, Kasinitz, and Dunn 2009; Livingstone and Weinfeld 2015). In 2006, black children were slightly more likely to be living in low-income families compared with 1986. Across the same 20-year period, the low-income rate for children in the general population had declined (Livingstone and Weinfeld 2015). This observation is consistent with a general trend across the country of rising low-income rates for recent immigrants since the 1980s, versus a decline for the native born (Picot and Hou 2003).

The present paper investigates the implications of racial disparities in income and family structure for the high school completion of black youth in Quebec and Ontario, because that is where the majority of these children live. The study does not seek to compare the two provinces, and the analysis is carried out separately for each province mainly because of the differences in provincial education systems. The principal objective of the paper remains to examine the correlates of racial disparities in high school completion.

BACKGROUND LITERATURE

Research on race and educational attainment has a long history in the United States. Early studies conducted on samples from the 1960s showed that controlling for socioeconomic status (SES) not only erased, but overturned the black/white gap in educational attainment (Haveman, Wolfe, and Spaulding 1991). In the years that followed, the racial educational gap would persist after controls for SES, reflecting a period of worsening poverty for many black families and children in the United States due to postindustrial shifts, persistent racial residential segregation, chronic poverty, mass incarceration, and cuts in public spending (Orfield 1996; Western 2006; Wilson 1996). One comprehensive review of the literature in the United States concluded that SES constitutes the most significant and consistent predictor of racial differences in educational attainment, explaining about one-half to two-thirds of the variance in selected outcome variables (Kao and Thompson 2003). As income inequality in the United States has intensified, the racial achievement gap has grown smaller while the income achievement gap has increased (Reardon 2011). Today, a child's success in school in the United States is increasingly dependent on parental income.

A variable associated with SES and with implications for children's educational attainment is family structure. American studies have found that children who belong to single-parent families and children who have experienced a change in family structure (e.g., from single parent to stepparent or from two parent to single parent) perform less well in school than their peers from stable two-parent families. Most studies in this field have also found that lone-parent status affects children's educational attainment primarily by diminishing the economic resources of the family, rather than by altering the quality of parenting (Astone and McLanahan 1991; Boggess 1998; Heard 2007; Pong and Ju 2000; Wang et al. 2004).

A handful of Canadian studies have explored the relationship between SES, family structure, and children's educational attainment in grades K through 12 (Boyle et al. 2007). One longitudinal study begun in Quebec in the 1980s with a sample of children followed from kindergarten to early adulthood observed that adolescent boys who had been "persistently poor" were 4.19 more likely than boys who had "never been poor" to be held back a grade (Pagani et al. 1999). Findings further showed that children experiencing multiple disadvantages--that is, children who had been held back a grade, who belonged to single-parent families, and whose parents had limited formal education--were the group least likely to graduate from high school, whether they were boys or girls (Pagani et al. 2008). A second longitudinal survey involving Ontario children, the "Ontario Child Health Survey," showed that family and neighborhood income both bore a significant, positive correlation with a child's educational attainment (Boyle et al. 2007).

The surveys cited above were conducted principally with native-born children (Boyle et al. 2007). More recent surveys involving immigrant children in Canada have found that the negative effects of poverty are attenuated for immigrant children compared with native-born children. Georgiades, Boyle, and Duku (2007) find that despite their more pronounced rates of poverty, immigrant children show better mental health and higher academic scores than their native-born counterparts. The authors propose that immigrant children may possess a kind of resiliency that protects them from the harms of poverty (Georgiades et al. 2007). It is also plausible that the circumstances of poverty for immigrant versus native children are not altogether the same. Unlike immigrant families, native-born families are more likely to be poor because they are headed by single parents, and poverty for such families tends to be of longer duration (Beiser et al. 2002; Gucciardi, Celasun, and Stewart 2004). In addition, native children face a higher likelihood of growing up in poverty, while the experience may be more transitory for immigrant children, as parents work toward establishing an economic foothold in Canada (Beiser et al. 2002).

Turning to the work of McAndrew et al. (2006, 2008) on black students in French and English high schools in Quebec, it was found that the proportion of black students entering high school in the mid-1990s and graduating on time was significantly below the average. To identify students as "black," the authors singled-out students whose parents were born in a country where the dominant population is of African descent, namely in sub-Saharan Africa (excluding South Africa) and the Caribbean. This remains a safe method of sampling black students, because most black youth in Quebec today remain the children of immigrants. It does, however, exclude black students of the third plus generation, who tend to be enrolled in English schools in the province. In that study, which used school administrative data, results showed that 51.8 percent of black students entering high school between 1994 and 1996 had graduated on time, versus 69 percent of students in the general population. In a more recent investigation on French schools alone, McAndrew et al. (2013) found the rate of high school completion had risen by roughly two points for black students starting high school between 1998 and 2000. The within-group variations were also substantial, with graduation rates ranging from a high of 63.1 percent for black students whose mother tongue is French, and a low of 36 and 42 percent for students whose mother tongue is "other." In both cohorts of students, black students were more inclined than others to resort to the adult education sector in order to finish their diplomas; for the 1998 to 2000 cohort, the figures amounted to 12.3 percent of black students and 7.3 percent of the comparison group (Balde et al. 2011). (1)

Balde et al. (2011) augment their analysis on black students in Quebec with data on private schools and show the rate of graduation for black students and all other students in these schools surpasses the public sector by more than 20 points. For black students, the graduation rate goes up to 75.5 percent for students starting high school between 1998 and 2000. The racial gap in graduation rates in private schools also diminishes, yet remains as high as 11.5 points. Without further research, it is difficult to know the precise reasons for the large discrepancies between the two types of schools. The answer might be found in the interaction between school-level differences and selection factors, such as the characteristics of black parents who can afford the fees that come with enrolling one's child in Quebec's subsidized private schools (e.g., two earners, professional class).

In a regression analysis performed on results for the 1994 to 1996 cohort, McAndrew et al. (2008) find the variables bearing the strongest relationship to high school completion for black students were as follows: placement in "special education," language, gender, birthplace (i.e., Quebec or foreign-born), and birthplace (i.e., Caribbean or African-born). SES was significant for black students when examined alone, but not when placed alongside other background variables. The authors acknowledge that the failure of SES to remain significant in the full model could be due to methodological factors. In constructing their variable for SES, the authors matched each student's postal code with the average income of the Census Enumeration Area in which she or he resided, which may have been too sweeping to pick up racial inequalities in household income.

McAndrew et al. (2008) further conclude that the variable with the most pronounced effect on black student's odds of graduation is "enrollment in special education." According to the latest findings in Balde et al. (2011), fully 37.4 percent of black students who had begun high school between 1998 and 2000 were enrolled in special education in French schools in Quebec, versus 24.2 percent of the comparison group. These results represent a substantial jump from the rate of 14 percent for the 1994 to 1996 cohort of black students in special education; at the time, the racial gap was no more than a few percentage points. If placement in special education continues to be a prime determinant of high school completion, then close to half of all black students in French school will be at risk. These findings should compel policymakers and researchers to investigate why the racial gaps remain as large as they do, why the rate of special education placement for black students has increased so sharply, and what can be done to prevent these inequalities from persisting. Robson et al. (2014) report that among students in Toronto schools, black males were overrepresented in special needs classes and applied streams; as a result of these combined factors, black male students were also the group least likely to enroll in college and university. In the United States, research has consistently shown that racial disproportionality in special education placement cannot be satisfactorily accounted for by differences in students' learning or behavioral competencies and that racial biases and social class also play a significant role (Hosp and Reschly 2004; Oswald et al. 1999; Skiba et al. 2008).

The study by Anisef et al. (2010) used school data from the TDSB to examine the high school completion of immigrant students in Toronto, compared with the majority of native-born English-speaking students. Immigrant students were classified as those born abroad or those whose parents were born elsewhere. Included in this sample were students of Caribbean and Africa origins, whose racial identities remained unknown. In addition, it would appear that black students of the third generation or higher were grouped into the general category of English-speaking native students. (2) While the race of students could not be confirmed, findings from Anisef et al. (2010) do offer some tentative evidence on the trajectories of black immigrant students in Toronto high schools. The results showed that across a five-year period from 2001 to 2006, students of Caribbean and African origin were less likely than students in the comparison group to have finished high school. The likelihood of graduation was especially low for Caribbean-origin students, 40 percent of whom had finished high school by 2006. African-origin students were doing noticeably better, with 77 percent having graduated by 2006, compared to 81 percent of native-born English students.

In a multivariate analysis, Anisef et al. (2010) find that after controlling for the SES of neighborhoods in which students reside and other relevant demographic and academic variables, the gaps in graduation for Caribbean- and African-origin students become nonsignificant. Caribbean-origin students are now as likely as others to graduate from high school, while African-origin students rise above the average. Thus, SES and academic-related disadvantages (e.g., being enrolled in a low academic stream) would seem to explain a large part of the discrepancies in graduation rates between Caribbean- and African-origin students and the general population of students. Compared to others, Caribbean origin students were confronted with the most significant challenges; they were more likely to begin high school late, more often enrolled in a nonacademic stream, facing higher low-income rates, and less likely to belong to a two-parent household.

Qualitative research on the schooling experiences of black students in Canada has helped clarify the processes by which social and organizational processes in school influence the motivation and performance of black high school students. Results of that research reveal the central role of parents in motivating black youth to strive in school and to aim high, in spite of any obstacles they might encounter (Codjoe 2007; Dei 1993; Livingstone et al. 2014; Smith, Schneider, and Ruck 2005). The school factors that bear an important impact on black students are teacher expectations, the representation of black faculty or lack thereof, discriminatory disciplinary practices, Euro-centric curricula, and racial academic tracking (Dei 1993; Dei et al. 1997; Howard 2014; James 2011; Kelly 1998; Smith et al. 2005). In one Montreal study, black students reported that racism actually drove them to work harder in school, because they wanted to defy racial stereotypes of black youth as "underachieving" (Livingstone et al. 2014). Others said they struggled to resist the impact of racism and to retain a positive sense of identity and optimism in the face of the challenges. These black students further claimed that caring teachers and community workers were crucial in lifting their self-confidence and motivating them to excel in school. Experiences with racism within and outside of high school can, therefore, be expected to bear a considerable impact on black students.

The present study examines the high school completion of black youth in Quebec and Ontario and explores the extent to which racial disparities in education can be explained by income, family structure, language, gender, and immigration status. The study seeks to extend and complement previous research on black high school students in Quebec and immigrant students in the TDSB (Anisef et al. 2010; McAndrew et al. 2006; see Table 1). Unlike these previous studies, the present paper examines the joint influence of race, income, and family structure. McAndrew et al. (2006, 2008, 2013) studied race and income, though not family structure. The sample in Anisef et al. (2010) was restricted to students in TDSB schools, and the analysis took into account income and family structure, though not race. The focus in this study is on Ontario and Quebec because the large majority of black Canadians live in those provinces. While it would have been interesting to include black students in Nova Scotia in the study, the sample sizes in the census remain too small to conduct a regression analysis. Throughout, the paper analyzes each province separately and independently, because of the unique immigrant histories and demography of blacks in these provinces, and the differences in educational systems. The purpose in this paper is not to investigate interprovincial differences, but rather to employ census data from the two provinces to probe the causes of racial disparities in high school completion. Our study captures all black students in these provinces, including those who may have attended private schools of various kinds.

METHODOLOGY

The data for the study are taken from the individuals' file of the 2006 Census "Public Use Micro-Data File," a representative sample of Canadians from the census and approximately 2.7 percent of the population. The study utilizes the 2006 data because the accuracy of the 2011 Census, known as the "National Household Survey" remains in question. In the Canadian scientific community, concerns remain about the undersampling of members of disadvantaged groups, racial and ethnic minorities, and residents of small census agglomerations (Grant 2013).

The data on racial identity are based on the "visible minority" variable in the census, which includes a separate code for "black." The proportional representation of black youth 18 to 19 years of age in the 2006 Census was 4.5 percent in Ontario and 2.7 percent in Quebec. The subcategory "Other" comprises individuals who selected more than one racial and ethnic origin in answering the census questionnaire (this group constituted 1.4 percent of the target group in Ontario, and 0.2 percent in Quebec). Individuals of First Nations descent have been left out from the sample, because incorporating them would require a separate study to do justice to the experiences of First Nations youth in Canada.

The sample is made up of respondents who were 18 to 19 years in 2006 and the children of a single- or dual-parent household. The census preselects the age categories that can be statistically manipulated; the choices available are 18 to 19 and 20 to 24 years. It would have been worthwhile to extend the sample's age range to 20 years in order to keep young people in the sample who take longer to graduate, but the census precludes it.

In order to isolate the impact of family structure on children as they were growing up, young people not living at home in 2006 were omitted from the sample. Young people who were living away from home in order to attend college or university would likely continue to be counted as the "dependents" of parents and, therefore, included in the study sample. The proportion of cases that had to be removed from the sample was relatively limited in each province. In Quebec, the percentages of youth "not living at home" in 2006 came to 9.7 percent for black and 5.7 percent for all other youth aged 18 to 19. In Ontario, the missing cases came to 5.6 percent of black youth and 3.4 percent of all youth of the same age. The sample still encompasses over 90 percent of 18- to 19-year-olds in both provinces. The slight discrepancies in the proportions of students living at home across the two provinces could reflect the earlier age of graduation in Quebec versus Ontario (grade 11 versus grade 12), and cost of living variances.

The analysis begins with two descriptive tables on the backgrounds of 18- to 19-year-olds in Quebec and Ontario, both those living at home or not (see Tables 2 and 3). Data are displayed for high school completion, parental education, family structure, and family income both for black youth and for youth of other racial and ethnic backgrounds. Parents in this sample are 25 years of age and higher and were reported to have children at home in the 15 to 24 age group (again, the census preselects the age categories, and there is no possibility to extract the group of parents with children who were strictly 18 to 19 years). In order to keep the focus on youth 18 to 19 years of age, parents who did not have children aged 15 to 24 at home were omitted from the tables. By using this criterion, we are left with 29.6 percent of black parents and 23 percent of all parents in Quebec. In Ontario, the corresponding percentages are 31.4 percent for black parents and 25.5 percent for all parents. The two tables for 2006 incorporate the sampling weights from Statistics Canada.

As indicated earlier, individual tables and regressions have been created for Quebec and Ontario, because of the differences in provincial educational systems. In Quebec, high school ends at grade 11, whereas in Ontario it continues to grade 12 (Ministere de l'Education, du Loisir et du Sport 20063; Ontario Ministry of Education 2011). Readers should keep in mind that 18- to 19-year-olds in Quebec are more likely have completed high school than their counterparts in Ontario. However, as indicated previously, the primary goal of this paper is to examine racial gaps, not interprovincial variations.

Table 4 displays the results of a logistic regression on the odds that a student who is 18 to 19 years old and still living with parents will have at least graduated from high school in Quebec and Ontario by 2005. The reference categories for the dummy variables in the regressions are as follows: "white" (corresponding to "nonvisible minority" and "non-Aboriginal"), "above the Low-Income Cut-Off' (LICO for shorthand), two-parent household, household size of two people, female student, foreign born, residence in a major Census Metropolitan Area (CMA), and an official language spoken at home. Adding CMA to the model removes the confounding effect of urban and rural variations, given that black students in Quebec and Ontario tend to be concentrated in major metropolitan areas, primarily Montreal (88 percent of black youth of this age live in Montreal) and Toronto (home to 70 percent of black youth in the sample).

For reasons of methodology and data limitations, the regression incorporates a variable for nativity and immigration, though not for the different places of birth of black students. While intragroup differences between native, Caribbean-born, and African-born students are important and worthy of attention, the present paper's objective is to concentrate on the sources of interracial variations. Moreover, once the samples of black 18- to 19-year-olds in each province are broken down by birthplace, many of the cells end up with fewer than 100 cases. As the number of native and immigrant black children in Canada continues to grow, it should be possible to use future censuses for such a within-group analysis.

In separate analyses conducted for this study, it was found that native-born black students in Quebec showed a graduation rate that was equal to the provincial average, while Caribbean and African-born black youth were faring less well. By 2005, 20 percent of native-born black youth had not yet finished high school, versus 46.4 percent of Caribbean-born and 35.9 percent of African-born black youth (data not shown). In Ontario, each group of black youth had a lower graduation rate than the provincial average, though it was Caribbean youth who were most likely to have finished high school by 2005. The results indicated that the proportion of black youth who had not finished high school was 30 percent for the Caribbean born, 33.5 percent for the native born, and 42 percent for those born in Africa (data not shown). The reasons why the intraracial patterns should vary across the two provinces remain unclear, and would be worth studying further.

Finally, interaction terms were computed to assess if the variable "black" bore any relationship with other background factors, namely nativity and immigration status, gender, LICO, single-parent status, and language. Not a single interaction term turned out to be statistically significant across the two provinces.

RESULTS

Table 2 displays the educational attainment of young adults 18 to 19 years old and still living at home in 2006 in Quebec. Table 3 does the same for 18-to 19-year-olds in Ontario. As shown, a higher proportion of black youth in Quebec had "not completed high school" by 2005 compared with white youth in the province (29.9 versus 23.3 percent). In Ontario, the proportion of black youth who had not finished high school by 2005 was also higher than white youth, although the margin was slightly lower (34.6 percent for black youth versus 31 percent for white youth). In the two provinces, Asian and South Asian youth had higher high school graduation rates than black youth. In Quebec, Latin American and Arab/West Asian youth showed rates of high school completion comparable to black youth.

The next section, in Tables 2 and 3, display the educational attainment of parents 25 years of age and above with children at home who were 15 to 24 years old. These data are presented for informational purposes, because they cannot be integrated into the regression analysis, yet the evidence of racial inequalities is revealing and worth keeping in mind in interpreting the results of the study. Findings show that black parents in Quebec and Ontario possessed fewer years of education in 2005 than either white or other racial minority parents. In Ontario, similar proportions of black and white parents had completed high school and a postsecondary degree, yet white parents were noticeably more likely to have obtained a bachelor's degree or higher (20.5 percent white parents versus 14.1 percent of black parents). In Quebec, black parents were slightly less likely than white parents to have finished high school, though they were about as likely to have acquired a bachelor's degree or higher (16.2 percent for white parents versus 15.8 percent for black parents). In Quebec, Asian and Arab/West Asian parents were substantially more likely than all other groups to hold a bachelor's degree or higher; in Ontario, the same conclusion applies to Asian, South Asian, and Arab/West Asian parents. Though they are relatively modest, the educational disparities between black and white parents in the two provinces can be expected to have some impact on racial disparities in high school completion.

The third section in Tables 2 and 3 displays the percentages of households living above and below the LICO. Results reveal large racial disparities in income between the households of 18- to 19-years-olds, controlling for family structure. In the two provinces, the proportion of black youth living in a family with an income below the LICO is nearly three times higher than white youth, both for single- and two-parent families. In Quebec, 34.3 percent of black youth were found below the LICO, compared with 11.5 percent of white youth. In Ontario, the percentage was 29 percent for black youth and 9.8 percent for white youth.

The fourth section in Tables 2 and 3 presents data on the household composition of all 18- to 19-year-olds, including those who had left the parental home by 2005. Rows have been added for households with members of an extended family (labeled "extended families") and "nonfamily households," in which young adults may be living alone or sharing a home with nonrelatives. As shown, black youth in both provinces were far more likely than any other group to live in a single-parent home: the percentage of black 18- to 19-year-olds in this group was 46.3 percent in Quebec and 43 percent in Ontario. For white 18- to 19-year-olds, the respective percentages were 21.5 percent in Quebec and 18.8 percent in Ontario.

Table 4 presents the results of the logistic regressions, with the odds ratios for high school completion in Quebec and Ontario. The first column for each province displays the main effects for race where "white" students are the comparison group; the second column presents the odds ratios for the full list of independent variables. As shown, the racial gap in high school completion between black and white students is nonsignificant in Quebec in column 1, and statistically significant in Ontario (p < .05). In Quebec, the odds ratio in column 1 is statistically significant only for Latin American youth (p < .05). In Ontario, the odds ratios for Asian and South Asian youth are highly significant, both in columns 1 and 2. The odds ratio for Arab/West Asian youth in Ontario is nonsignificant in column 1, yet significant in column 2 (p < .05).

The full regression model for Quebec reveals that the variables with the strongest relationship to high school graduation are income, family structure, gender, and residence in a major CMA. When these variables are controlled, the black/white gap disappears. The odds ratio for LICO indicates that a youth from a low-income family--whether she or he is white or black--would have a 59 percent chance of finishing high school by the age of 18 to 19 years compared with a youth from a family above the LICO. For young people in single-parent families, the odds of graduating from high school were 53 percent compared to those from two-parent families. Young people living in an "extended family household" also showed significantly lower odds of graduating compared to their peers in two-parent households. Males are about half as likely as females in Quebec to graduate from high school. Attending school in a major CMA increases the odds of graduation by about a quarter. Speaking a nonofficial language at home in Quebec bears little or no relation to high school graduation, which is consistent with recent findings on immigrant students in Quebec (McAndrew et al. 2013). In addition, nativity and household size are both nonsignificant. Nevertheless, there is a slight positive advantage to being native born.

The full regression model for Ontario displays a pattern of results similar to Quebec, with a few exceptions. Like Quebec, the effects of income, family structure, and gender are all highly significant (p < .001). Adding these variables to the model erases the black/white gap in high school graduation. Unlike Quebec, language turns out to be highly significant in Ontario. Living outside a major CMA in Ontario has a modest negative impact. The odds ratios for Asian and South Asian youth remain strongly significant even with the addition of controls, suggesting that other unmeasured characteristics of these groups bear an impact on their higher levels of educational attainment.

Compared with their peers in better-off families, young people from families earning less than the LICO have a 60 percent chance of finishing high school by the time they are 18 to 19 years old. For young people from single-parent families, the odds fall even lower to 53 percent, when compared to two-parent families. Being part of an extended family represents less of an impediment in Ontario than Quebec; the odds ratio is nonsignificant and 15 percent lower than the control group. The female lead in high school completion remains strong and highly significant in Ontario, though the gap is smaller than Quebec (67 percent). Finally, similar to Quebec, nativity and household size are both nonsignificant in Ontario.

DISCUSSION

The findings of the study reveal that in the two provinces, household income and family structure both play a significant role in determining whether or not black students and all other students will graduate from high school by the age of 18 to 19 years. The gap in high school completion between black and white students was modest and nonsignificant to begin with in Quebec; when income, family structure, gender, and residence in a CMA were entered into the regression, the numerical difference was erased. In Ontario, the black/white gap was relatively larger and statistically significant in the main effects model, though the addition of the control variables brought the gap close to zero. The single divergence between the two provinces rests in the effect of language, or speaking a "nonofficial language at home," which was statistically significant in Ontario and not in Quebec. Results further indicate that gender exerts a fairly strong impact on high school completion in the two provinces, with males faring poorly. While this pattern has been well-documented in the general population, its significance among black youth in Canada is less well-known, due to the limited research (McAndrew et al. 2006; Simmons and Plaza 1998). It is probable that the gender gap is more pronounced for low-income than for middle-class black students, an issue worth studying further (Martino 2008).

The study has demonstrated that the census can be used successfully to explore high school attainment, and the results correspond with the major quantitative studies on this topic in Quebec and Ontario. Given that the school systems in Quebec and Ontario differ, as does the history and demography of the black community in the provinces, the interprovincial similarities in the correlates of black/white gaps in graduation are all the more telling. Nevertheless, the contrasts discussed earlier between native and immigrant black students across the two provinces do suggest that not everything is exactly the same. The subject of interprovincial variations deserves to be investigated further.

The present research builds on and extends the earlier studies of McAndrew et al. (2006) and Anisef et al. (2010) in several ways. McAndrew et al. (2006) examined students in Montreal and the rest of Quebec, while Anisef et al. (2010) concentrated on students attending schools in Toronto. The present study looks at students living anywhere in the two provinces, whether they were enrolled in public or private schools. As indicated earlier, the data sets available to McAndrew et al. (2006) and Anisef et al. (2010) did not include information on race, contrary to the census. McAndrew et al. (2006) and Anisef et al. (2010) assessed the effect of SES on high school completion by constructing an aggregate indicator of neighborhood income, which lacks the precision of census data on household income. The present study has examined the variations in family structures, unlike the other two studies. McAndrew et al. (2006) did not have any data on family structure, while in Anisef et al. (2010) the categories were "two-parent" and "other." Presumably, the latter code merges single-parent and extended families into a single variable. Even though the present study relies on demographic data from the census and not school records, the findings are broadly consistent with McAndrew et al. (2006) and Anisef et al. (2010) in demonstrating the significant relationship between SES and black/white gaps in graduation rates in Ontario and Quebec. If racial disparities in income and family structure were not so significant in the two provinces, the graduation rate of black students would almost certainly be comparable to the average. Given the persistence of racial gaps in high school completion, scholars and policymakers might invest more efforts into analyzing and ameliorating opportunities for black students and other disadvantaged students to succeed. This effort can involve qualitative research as well as quantitative studies using a variety of data sources, including the census.

The finding that income and family structure together explain a large part of any black/white gap in high school completion does not imply that antiblack racism in schools is either inconsequential or nonexistent. Indeed, the dependent variable chosen for the present study--high school completion--may be too broad of a figure and not sensitive enough to detect the effects of racism in school. In addition, high school completion may be a low bar for estimating the impact of racism on black students' educational aspirations and achievements. A more effective method might be to examine the source of racial disparities in more proximate and intermediate indicators of school success, such as teacher-student relations, grades, detentions and suspensions, academic tracking, special education placement, and course selection (Galabuzzi 2014). This research would be strengthened with qualitative studies involving school personnel, parents, and students. A recent TDSB report showed, for example, that black students in grades 7 to 12 were nearly three times more likely than white students to be suspended from school (TDSB 2013). In the United States, one study found that African-American male students were more likely than any other group to cite suspensions and expulsions as their reason for leaving school (Jordan et al. 1996). Therefore, young people's encounters with racial discrimination in school may be deep and far-reaching. These effects may extend beyond high school, limiting black students' prospects of enrolling in postsecondary education and finding stable, well-remunerated jobs (Robson, Brown, and Anisef 2015; Robson et al. 2014).

In interpreting the association between income, family structure, and racial disparities in educational attainment, it is important to keep in mind the closely intertwined relationship between social class and race, gender, and immigration (Robson et al. 2014). As indicated earlier, black families in Canada face higher rates of socioeconomic disadvantage than average, and this is due not only to class, but also racial discrimination, labor market barriers for immigrants, segregation in low-wage work, and gender biases (Livingstone and Weinfeld 2015). It is also because of these combined forces that we find more black women heading households on their own. Black single mothers who are immigrants face the daunting task of having to juggle paid work and taking care of children without the social support systems they depended on in their home countries (Guruge et al. 2015; Whitley and Green 2008).

This study has shown that socioeconomic inequalities impacting black families in Canada play a key role in hampering black students' attainment of a high school degree. It is likely that the effects of racism in high schools are manifest in intermediate outcomes of high school performance, and in the mobility of black students after high school. Future research should attempt to link the experience of racism in high schools with postsecondary and employment outcomes. Studies examining variability in high school completion among black students and across different types of schools would also be instructive. Together, research on these themes would enrich the scholarship on black students and racial stratification in education in Canada, and lay out a clearer path for public policy.

Anne-Marie Livingstone

Johns Hopkins University

Morton Weinfeld

McGill University

We are grateful to the two anonymous reviewers for their insightful and constructive feedback on the first version of this paper. We remain solely responsible for its content .

Anne-Marie Livingstone, Department of Sociology, John Hopkins University, 533 Mergenthaler Hall, 3400 N. Charles Street, Baltimore, MD 21218. E-mail: aliving6@jhu.edu

References

Anisef, P., R. Brown, K. Phythian, R. Sweet and D. Walters. 2010. "Early School Leaving among Immigrants in Toronto Secondary Schools." Canadian Review of Sociology 47(2): 103-28.

Astone, N. and S. McLanahan. 1991. "Family Structure, Parental Practices, and High School Completion." American Sociological Review 56:309-20.

Attewell, P., P. Kasinitz, and K. Dunn. 2009. "Black Canadians and Black Americans: Racial Income Inequality in Comparative Perspective." Ethnic and Racial Studies 33(3):47395.

Balde, A., J. Sene and M. McAndrew. 2011. Portrait des jeunes Quebecois originaires des Antilles et de l'Afrique subsaharienne: Secteur francais. Montreal, QC: Chaire de Recherche du Canada sur l'Education et les Rapports Ethniques, 49 pp.

Beiser, M., F. Hou, I. Hyman and M. Tousignant. 2002. "Poverty, Family Process, and the Mental Health of Immigrant Children in Canada." American Journal of Public Health 92(2):220-27.

Boggess, S. 1998. "Family Structure, Economic Status, and Educational Attainment." Journal of Population Economics 11(2):205-27.

Boyle, M., K. Georgiades, Y. Racine and C. Mustard. 2007. "Neighbourhood and Family Influences on Educational Attainment: Results from the Ontario Child Health Study Follow-Up." Child Development 78(11:168-89.

Codjoe, H.M. 2001. "Fighting a 'Public Enemy' of Black Academic Achievement: The Persistence of Racism and the Schooling Experiences of Black Students in Canada." Race, Ethnicity, and Education 4(41:343-75.

Codjoe, H.M. 2007. "The Importance of Home Environment and Parental Encouragement in the Academic Achievement of African-Canadian Youth." Canadian Journal of Education 30(l):137-56.

Dei, G.S. 1993. "Narrative Discourses of Black/African-Canadian Parents and Canadian Public School Systems." Canadian Ethnic Studies 25(3):45-66.

Dei, G.S., J. Mazzuca, E. Mclsaac and J. Zine. 1997. Reconstructing Dropout: Understanding the Dynamics of Black Students Disengagement from School. Toronto, ON: University of Toronto Press.

Fitzpatrick, V., C. Cote-Lussier, L. Pagani and C. Blair. 2013. '"I Don't Think You Like Me Very Much': Child Minority Status and Disadvantage Predict Relationship Quality with Teachers." Youth and Society 20(10):1-17.

Galabuzzi, G. 2014. "Race and the Streaming of Ontario's Children and Youth." Pp. 185-225 in Restocking the Deck: Streaming by Class, Race, and Gender in Ontario Schools, edited by D. Clanfield, B. Curtis, G. Galabuzzi, A. San Vicente, D. Livingstone, and H. Smaller. Ottawa, ON: Canadian Centre for Policy Alternatives.

Georgiades, K., M. Boyle and R. Duku. 2007. "Contextual Influences on Children's Mental Health and School Performance: The Moderating Effects of Family Immigrant Status." Child Development 78(5):1572-91.

Grant, T. 2013. "Census Changes Produce Flawed Data, Researchers Say." Globe and Mail, October 5, p. A6.

Gucciardi, E., N. Celasun and D. Stewart. 2004. "Single-Mother Families in Canada." Canadian Journal of Public Health 95(l):70-73.

Guruge, S., M. Thomson, U. George and F. Chaze. 2015. "Social Support, Social Conflict, and Immigrant Women's Mental Health in a Canadian Context: A Scoping Review." Journal of Psychiatric and Mental Health Nursing 22:655-67.

Haveman, R., B. Wolfe and J. Spaulding. 1991. "Childhood Events as Circumstances Influencing High School Completion." Demography 28(1):133-57.

Heard, H. 2007. "The Family Structure Trajectory and Adolescent School Performance: Differential Effects by Race and Ethnicity." Journal of Family Issues 28(3):319-54.

Hosp, J. and D. Reschly. 2004. "Disproportionate Representation of Minority Students in Special Education: Academic, Demographic, and Economic Predictors." Exceptional Children 70:185-99.

Howard, P. 2014. "Taking the Bull by the Horns: The Critical Perspectives and Pedagogy of Two Black Teachers in Anglophone Montreal Schools." Race, Ethnicity and Education 17(4):494-517.

James, C. 2011. "Students At Risk': Stereotypes and the Schooling of Black Boys." Urban Education 47(2):464-94.

Jordan, W., J. Lara and J. McPartland. 1996. "Exploring the Causes of Early Dropout among Race-Ethnic and Gender Groups." Youth and Society 28(1):62-94.

Kanoute, F., J. Charrette, J. Vierginat Andre, L. Rachedi and J. Rahm. 2014. "Les dynamiques d'etablissement, les enjeux scolaires, et les ressources communautaires. Le point de vue de parents d'origine haitienne et d'intervenants." Diversite Urbaine 14(2):31-50.

Kao, G. and J. Thompson. 2003. "Racial and Ethnic Stratification in Achievement and Attainment." Annual Review of Sociology 29:417-42.

Kelly, J. 1998. Under the Gaze: Learning to be Black in White Society. Halifax: Fernwood. I

Lafortune, G. and A. Balde. 2012. "Cheminement scolaire des eleves Quebecois originaires des Antilles: un double aper?u a partir des donnees quantitatives et qualitatives." Diversite urbaine 12(11:49-68.

Livingstone, A., J. Celemencki and M. Calixte. 2014. "Youth Participatory Action Research and School Improvement: The Missing Voices of Black Youth in Montreal." Canadian Journal of Education 37(1):283-307.

Livingstone, A. and M. Weinfeld. 2015. "Black Families and Socio-Economic Inequality in Canada: Changes from 1986 to 2006." Canadian Ethnic Studies 47(3):1-23.

Martino, W. 2008. "Boys' Underachievement: Which Boys Are We Talking About?" Ontario: What Works? Research Into Practice (Research Monograph 12), Literacy and Numeracy Secretariat.

McAndrew, M. 2015. La Reussite Educative des Eleves Issus de l'Immigration: Dix Ans de Recherche et d'Intervention au Quebec. Montreal, QC: Les Presses de l'Universite de Montreal.

McAndrew, M., J. Ledent and R. Ait-Said. 2006. "L'ecole quebecoise assure-t-elle l'egalite des chances? Le cheminement scolaire des jeunes noirs au secondaire." Cahiers quebecois de demographie 35(1):123-38.

McAndrew, M., J. Ledent and R. Ait-Said. 2008. La Reussite Scolaire des Jeunes de la Communaute Noire au Secondaire. Montreal, QC: Centre de recherche inter-universitaire sur l'immigration, l'integration, et la dynamique urbaine, Universite de Montreal.

McAndrew, M., J. Ledent, J. Murdoch, R. Ait-Said and A. Balde. 2013. "Le profil et le cheminement scolaire des jeunes Quebecois issus de l'immigration au secondaire: un portrait statistique." Cahiers Quebecois de demographie 42(1):31-55.

Ministere de l'Education, du Loisir et du Sport. 2006. L education au Quebec: En bref. Quebec, QC: Gouvernement du Quebec, Ministere de l'Education, Loisir et Sport, 12 pp.

Ontario Ministry of Education. 2011. Ontario Schools, kindergarten to grade 12: Policy and program requirements. Ottawa, ON: Government of Ontario, Ontario Ministry of Education, 100 pp.

Orfield, G. 1996. Dismantling Desegregation: The Quiet Reversal of Brown versus Board of Education. New York: The New Press.

Oswald, D., M. Coutinho, A. Best and N. Singh. 1999. "Ethnic Representation in Special Education: The Influence of School-Related Economic and Demographic Variables." Journal of Special Education 32:194-206.

Pagani, L., B. Boulerice, F. Vitaro and R. Tremblay. 1999. "Effects of Poverty on Academic Failure and Delinquency in Boys: A Change and Process Model Approach." Journal of Child Psychology and Psychiatry 40(8): 1209-19.

Pagani, L., F. Vitaro, R. Tremblay, P. McDuff, C. Japel and S. Larose. 2008. "When Predictions Fail: The Case of Unexpected Pathways toward High School Dropout." Journal of Social Issues 64(1): 175-94.

Picot, G. and F. Hou. 2003. The Rise in Low-Income Rates among Immigrants in Canada. Research Paper, No. 198. Ottawa: Analytical Studies Branch, Statistics Canada, 57 pp.

Pong, S. and D. Ju. 2000. "The Effects of Change in Family Structure and Income on Dropping Out of Middle and High School." Journal of Family Issues 21(2): 147-68.

Reardon, S. 2011. "The Widening Academic Achievement Gap between Rich and Poor: New Evidence and Possible Explanations." Pp. 91-115 in Wither Opportunity? Rising Inequality, Schools and Children's Life Chances, edited by G. Duncan and R. Murnane. New York: Russell Sage Foundation.

Robson, K., P. Anisef, R.S. Brown and G., Parekh 2014. "The Intersectionality of Post-Secondary Pathways: The Case of High School Students with Special Education Needs." Canadian Review of Sociology 51(3):193-215.

Robson, K., R.S. Brown and P. Anisef 2015. Identifying the Complexity of Barriers Faced by Marginalized Youth in Transition to Post-Secondary Education in Ontario. Hamilton, Ontario: The Gateway Cities Project, Faculty of Social Sciences, McMaster University, 22 pp.

Ruck, M. and S. Wortley. 2002. "Racial and Ethnic Minority High School Students' Perceptions of School Disciplinary Practices: A Look at Some Canadian Findings." Journal of Youth and Adolescence 31(3): 185-95.

Simmons, A. and D. Plaza. 1998. "Breaking through the Glass Ceiling: The Pursuit of University Training among African-Caribbean Migrants and Their Children in Toronto." Canadian Ethnic Studies 30(3):99-119.

Skiba, R., A. Simmons, S. Ritter, A. Gibb, M. Rausch, J. Cuadrado and C.-G. Chung. 2008. "Achieving Equity in Special Education: History, Status, and Current Challenges." Exceptional Children 74(3):264-88.

Smith, A., B. Schneider and M. Ruck. 2005. "Thinking about Makin' It: Black Canadian Students' Beliefs Regarding Education and Academic Achievement." Journal of Youth and Adolescence 34(4):347-59.

Solomon, R. and H. Palmer. 2004. "Schooling in Babylon: When Racial Profiling and Zero Tolerance Converge." Canadian Journal of Educational Administration and Policy 33:1-16.

Toronto District School Board (TDSB). 2013. Suspension Rates by Students' Demographic and Family Background Characteristics. Toronto, ON: Toronto District School Board, 6 pp.

Wang, G., M. Buffalo, S. Bakr and L. Spataro. 2004. "The Effect of Female-Headed Households on Educational Attainment: A Comparative Analysis of White and Black Respondents." Journal of Applied Social Science 21:135-50.

Western, B. 2006. Punishment and Inequality in America. New York: Russell Sage Foundation.

Whitley, R. and S. Green. 2008. "Psychosocial Stressors and Buffers Affecting Black Women in Montreal." Canadian Journal of Community 27(1):37-48.

Wilson, W. 1996. When Work Disappears: The World of the New Urban Poor. New York: Alfred A. Knopf.

(1.) Balde et al. 12011) employ the same methods as McAndrew et al. (2006, 2008) in order to assemble black students into a single group.

(2.) The study does not indicate whether students who have African or Caribbean origins are racially "black" or not. African-born students may be white. North African, or South Asian. In addition, Caribbean immigrant students may either be black, white. South Asian, or of mixed origins.

(3.) Currently renamed as the Ministere de l'Education et de l'Enseignement superieur. Table 1 Comparison of Four Studies on High School Completion McAndrew et al. McAndrew et al. Study (2006, 2008) (2013) Sample characteristics Data source Public schools in Public schools in Quebec Quebec Province Quebec Quebec Age group 1994-1996 cohort of 1999-2003 cohort of seventh graders seventh graders Time of graduation 1999-2001 2004-2006 Independent variables Race Yes No Ethnicity/national Yes Yes origins Income Postal code Postal code Family structure No No Native or foreign born Yes Yes Language Yes Yes School variables Yes Yes Anisef et al. Present study Study (2010) (2006 Census) Sample characteristics Data source Public schools in Canadian census the TDSB (public and private schools) Province Toronto Quebec and Ontario Age group Students who began Youth aged 18-19 and high school in living at home 2001 Time of graduation 2006 2005 Independent variables Race No Yes Ethnicity/national Yes No origins Income Postal code Household income Family structure 1. "Two parent" 1. Two parent 2. "Alternative 2. Single parent family 3. Extended family structure" Native or foreign born Yes Yes Language Yes Yes School variables Yes No Table 2 Profile of Youth 18 to 19 Years of Age in Quebec Either Living at Home or Not White Asian South Asian Black Education of 18- to 19-year-olds No high school 23.3 22.2 18.8 29.9 High school 61.0 57.8 59.4 58.2 Above high school 15.7 20.0 21.8 11.9 Total (%) 100 100 100 100 Education of parents >25 years years No high school 17.6 23.9 25.3 21.6 High school 43.9 28.7 35.5 39.6 College 22.4 16.7 18.4 23.0 Bachelor's or higher 16.2 30.7 20.8 15.8 Total (%) 100 100 100 100 Percent below LICO Two parent 5.4 26.5 30.3 16.4 Single parent 13.3 42.9 33.3 40.3 All families 11.5 29.6 31.0 34.3 Household type Two parent 70.7 77.3 76.0 41.0 Lone parent 21.5 15.9 12.0 46.3 Extended family 2.1 4.5 8.0 3.0 Nonfamily household 5.7 2.3 4.0 9.7 Total (%) 100 100 100 100 Arab/West Other Latino Asian origins Total Education of 18- to 19-year-olds No high school 29.1 28.6 11.1 23.5 High school 64.6 52.4 77.8 60.9 Above high school 6.3 19.0 11.1 15.6 Total (%) 100 100 100 100 Education of parents >25 years No high school 23.8 18.5 37.5 18.0 High school 39.8 28.6 43.8 43.1 College 18.2 20.5 12.5 22.2 Bachelor's or higher 18.2 32.4 6.3 16.7 Total (%) 100 100 100 100 Percent below LICO Two parent 22.9 19.1 40.0 7.0 Single parent 22.2 38.5 0.0 15.8 All families 24.1 30.7 22.2 13.3 Household type Two parent 60.8 67.7 55.6 69.8 Lone parent 34.2 21.0 44.4 22.2 Extended family 2.5 1.6 0.0 2.3 Nonfamily household 2.5 9.7 0.0 5.7 Total (%) 100 100 100 100 Source: 2006 Canadian Census. Table 3 Profile of Youth 18 to 19 Years of Age in Ontario Either Living at Home or Not White Asian South Asian Black Education of 18- to 19-year-olds No high school 31.0 21.5 24.4 34.6 High school 64.5 70.9 65.5 58.5 Above high school 4.5 7.6 10.1 6.9 Total (%) 100 100 100 100 Parental education No high school 13.9 17.9 13.5 13.7 High school 38.3 28.2 26.7 37.4 College 27.3 19.8 23.6 34.8 Bachelor's or higher 20.5 34.1 36.2 14.1 Total (%) 100 100 100 100 Percent below LICO Two parent 5.5 20.8 16.7 12.6 Single parent 14.9 41.5 29.5 41.7 All families 9.8 28.7 19.2 29.0 Household type Two parent 75.1 70.9 79.9 42.8 Lone parent 18.8 13.0 9.6 43.0 Extended family 2.9 9.1 8.3 8.6 Nonfamily household 3.2 7.0 2.2 5.6 Total (%) 100 100 100 100 Arab/West Other Latino Asian origins Total Education of 18- to 19-year-olds No high school 33.9 27.0 27.0 30.0 High school 58.3 68.5 68.8 67.8 Above high school 7.9 4.5 5.3 5.4 Total (%) 100 100 100 100 Parental education No high school 16.4 19.6 20.4 14.2 High school 45.6 25.7 38.1 36.4 College 22.7 18.5 28.4 28.4 Bachelor's or higher 15.3 36.2 13.1 13.1 Total (%) 100 100 100 100 Percent below LICO Two parent 13.7 34.3 15.7 8.6 Single parent 35.9 55.3 25.0 21.1 All families 22.8 38.4 18.4 13.7 Household type Two parent 57.5 73.7 72.8 73.4 Lone parent 30.7 19.2 17.5 18.9 Extended family 7.9 4.6 7.1 4.2 Nonfamily household 3.9 2.5 2.6 3.5 Total (%) 100 100 100 100 Source: 2006 Canadian Census. Table 4 Logistic Regression on the Odds of High School Completion for Young Adults Living at Home, 18 to 19 Years of Age, by Select Independent Variables Quebec Independent variable Odds ratio Odds ratio Race (white) Asian 0.91 1.05 South Asian 1.39 1.72 Black 0.74 1.1 Latin American 0.58 * 0.79 Arab/West Asian 0.58 0.67 Other (multiple or single origins) 2.1 2.61 Above LICO Below LICO 0.59 *** Two parent Lone parent 0.53 *** Extended household 0.47 * Household size (2 members) 3-4 members 1.02 5-6 members 0.86 7 or more 0.73 Female Male 0.49 *** Foreign born Native born 1.22 Residence in major CMA Nonmajor CMA .77 ** French or English spoken at home Nonofficial language spoken at home 0.92 Prob > [chi square] 0.089 0.000 Constant -1.32 -1.97 N 4,012 3,998 Variables Ontario Independent variable Odds ratio Odds ratio Race (white) Asian 1.45 *** 1.85 *** South Asian 1.31 ** 1.47 *** Black 0.76 * 0.96 Latin American 0.78 1.04 Arab/West Asian 1.08 1.53 * Other (multiple or single origins) 1.11 1.11 Above LICO Below LICO 0.6 *** Two parent Lone parent 0.53 *** Extended household 0.83 Household size (2 members) 3-4 members 0.83 5-6 members 0.79 7 or more 0.57 ** Female Male 0.67 *** Foreign born Native born 0.85 Residence in major CMA Nonmajor CMA .9 * French or English spoken at home Nonofficial language spoken at home 0.64 *** Prob > [chi square] 0.000 0.000 Constant -0.86 -1.36 N 7,785 7,734 Source: 2006 Canadian Census. Note: p < .05; ** p < .01; *** p < .001.