4 Results

Warning View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

4.1 Descriptive results

4.1.1 Group differences in high school and university completion

The results in Table 1 show that the choice of indicators is crucial for studying educational attainment across immigrant groups. While groups differ significantly in their levels of finishing university, there is very little difference in high school completion. The proportion of high school graduates are about the same for all groups, with over 90% having completed high school; this indicates that assessment of educational attainment by high school completion does not reveal substantial group differences. Only by examining the university levels of education do we observe vast inter-group differences.

Table 1
Educational attainment among children of immigrants,1 aged from 25 to 34, by source country/region

We observe wide differences in university completion rates among immigrant groups, with the most educated group—Chinese immigrant children at 69.5%—exceeding almost three times the rate of those near the bottom of the hierarchy—Latin American immigrant children at 23.5%. A striking pattern is noted among the Asian immigrant children, with the most highly educated groups by far being the Chinese, followed by Indians (65.2%), "other Asians" (52.2%), immigrant children from West Asian/Middle East (44.3%), and Filipino immigrant children (39.9%). Immigrant children from Africa also had a high university completion rate at 55.9%. As noted before, over 85% of the immigrants from Africa in our sample, most of who arrived in Canada before the 1980s, had European ethnic ancestry or belonged to visible minority groups other than Blacks. Immigrant children from the Caribbean (26.3%) and Latin America (23.5%) had the lowest levels among non-Western source regions.

For the identified Western source region groups, the highest levels of university completion are observed among "other Europe" (mostly South Europe excluding Italy and Portugal) (45.2%) and the United Kingdom (37.6%), with about one third observed among immigrant children from the United States and Italy. About one quarter of Dutch, Portuguese and German youth obtained a university degree.

Compared to children of Canadian-born parents, children of immigrant parents achieved a clear advantage with regard to university completion rates. Among immigrant groups, children whose parents were from Africa, China, India, West Asia/Middle East, United Kingdom, Eastern Europe, or "other Europe" had significantly higher rates of university completion than children of Canadian-born parents. Moreover, none of the children in the remaining groups had a university completion rate significantly lower than that among children of Canadian-born parents.

The results from the Ethnic Diversity Survey (EDS) should be interpreted with caution for three reasons. First, compared to the estimates from the 2001 and 2006 Census data, which are based on a much larger sample, and thus are more reliable (Table A.5), the EDS overestimates the level of university completion rate for both immigrant children and children of Canadian-born parents. This is particularly true for those from India, "other Europe" and Portugal. However, the census results show a similar pattern to that observed in the EDS for these group differences. Since census data do not contain information on most of our explanatory variables that are available in the EDS, in the following sections the analysis is only based on EDS. Second, the difference in the university completion rate between many groups of children of immigrants and children of Canadian-born parents is not statistically significant in the EDS data. This is mostly due to the small sample size of these groups. The difference in similar magnitude could be highly significant with the census data (Table A.5). Finally, the estimated university rates and group differences can change over a short period of time. As shown in Table A.5, while the rate among children of Canadian-born parents increased 2.4 percentage points from 2001 to 2006, it decreased among children whose parents were from the Philippines, West Asia/Middle East, and "other countries." Over the five-year period from 2001 to 2006, the population size of youth (aged 25 to 34) of immigrant parents (about five times the sample size in Table A.5) increased 84% for those from West Asia/Middle East, and 30% to 64% for those from Africa, Latin America, China, the Philippines, India, and "other Asia". The characteristics of some groups might also change significantly. Given the rapidly changing population size and composition among some national groups of children with immigrant parents, the results from the EDS should be treated as a snap-shot in a particular time period.

4.1.2 Group differences in family background and individual characteristics

Table 2 focuses on the group differences in the family background and individual characteristics: key factors expected to shape the educational attainments of the children of immigrants. The table shows large group differences in parental education. The most educated fathers are from the Philippines (42.7%), India (42.4%), the United States (37.9%), "other countries" (37.4%) and "other Asia" (34.1%), with over one third having obtained a university degree. The proportions of mothers who obtained a university degree are highest among immigrants from the Philippines (41.4%) and the United States, with over one third, followed by "other countries," West Asia/Middle East, India and "other Europe," with just less than one quarter having obtained a university degree. Less than 10% of immigrant fathers and mothers had a university degree among those from Italy and Portugal.

Table 2
Group differences in family background and individual characteristics

The comparison of parental education and children's university attainment reveals a substantial across-generation improvement in university completion rates among children of immigrants. While on average 24% of immigrant fathers finished university education, 37.6% of their children aged 25 to 34 finished their university education. This across-generation improvement seems smaller than that among youth born to Canadian-born parents, i.e., the so-called third-plus generation. About 27.5% of youth aged 25 to 34 whose both parents were born in Canada finished their university education, compared with 15.4% among their fathers. This suggests that the advantage of the second-generation youth in educational attainment over the third-plus generation at least partially originated from the differences in educational attainments among their parents.

Among children of immigrants, there are large group differences in intergenerational mobility (see Charts 1-1 and 1-2). Immigrant children of most groups attain a higher percentage of university education than their parents, with three exceptions. Italian and Portuguese children surpassed their fathers' educational levels by a substantial margin. Some 32.2% of Italian second-generation youth have university degrees, compared with 4.7% among their immigrant fathers. Similarly, 25.8% of Portuguese youths completed university education, although only 4.4% of their immigrant parents did so. The intergenerational mobility among youth of immigrant parents from China, " other Europe," Germany, West Asia/Middle East and Africa is also notable, with close to or over twice as many youth completing a university education relative to their fathers' attainment. This upward pattern is not observed for immigrants from the United

States, the Philippines and " other countries" that are also the only groups in which youth, on average, do not attain a higher proportion of university degrees than their immigrant fathers.

Overall, the group difference in university completion rates is much smaller among children of immigrants than among their parents. This suggests that the Canadian education system tends to reduce the group difference in educational attainment among immigrant parents that was largely associated with the characteristics of immigrant flows from various source regions at different time periods. While groups whose parental educational levels were very low achieved a large intergenerational mobility, those who experienced little improvement across generations still maintained a level near or above the average.

Further differences in family background characteristics reveal the diversity among these groups. The highest proportions (over 90%) who lived in intact families are those from China, India, " other Asia," Italy and " other Europe." Overall, youth in the Caribbean group were more likely to grow up in single-parent households (36.1%) than any of the other groups.

Not everyone grew up speaking either English or French with their parents, and these group differences may be due to the varied language skills in the official languages of the parental generation (Zhou and Xiong 2005). About 71.6% of Portuguese youth grew up speaking only their native language with their parents, followed by those from " other Asia" (67.7%), China (64.3%), India (54.8%) and West Asia/Middle East (54.6%). A lower proportion of Filipinos (32.7%) spoke their native language while growing up, reflecting greater English proficiency for this group. In addition to immigrants from the United States and the United Kingdom, the groups who spoke mainly one of the official languages are those from the Caribbean (almost 100%), the Netherlands (95.7%) and Germany (93.2%).

Table 3 demonstrates considerable group differences in group-level human capital in terms of the average educational level and income and in urban/rural residences among the fathers' generation, as explained in the Data and Methods section. Immigrant parents from the Philippines and the United States have the highest percentage finishing university education among the fathers' generation (around 42%), while Portuguese (2.4%) and Italians (9.8%) have the lowest levels. There is also a large dispersion in log annual earnings of fathers' generation across national origin groups, ranging from 10.37 for those from " other Asia" to 10.77 for those from the United Kingdom. This difference in log earnings is equivalent to about 40% difference in earnings. Compared with Canadian-born parents, most non-U.S./European immigrant parents had lower income even though some groups had much higher university completion rates than those of Canadian-born parents.

The father's generation among most national origin groups predominantly resided in the urban areas, but one third of immigrant parents from the Netherlands and over one fifth of those from the United States, Germany and other Northern/Western Europe lived in rural areas. Over a quarter of Canadian-born fathers resided in rural areas.

Table 3
Group differences in fathers' generation characteristics

Chart 1-1
Comparing children and fathers' educational attainment — Non-U.S./European source regions/countries

Chart 1-2
Comparing children and fathers' educational attainment — U.S./European source regions/countries

4.2 Multivariate analysis

4.2.1 Group differences in university completion

Table 4 shows the extent to which the selected five sets of explanatory variables account for group differences in university completion rates. The first column in Table 4 presents the observed university completion rates as in Table 1. The second and third columns present university completion rates estimated for the logistic regression model and Ordinary Least Squares (OLS) model, respectively, in Table A.1, by assuming each group has the same characteristics as the whole sample in the control variables included in the model. The logit and OLS estimates are almost identical. The fourth column is the difference between the observed (first column) and estimated (third column) university completion rates for each group. This difference shows the portion of a group's advantage or disadvantage that can be 'explained' by the control variables. The remaining (fifth to tenth) columns show the contribution from each set of control variables to the 'explained' portion (see Footnotes 4 and 5 for details).

Table 4
Observed and estimated percentages with university degrees by source country/region

When group differences in demographic factors, parental education, and group human capitals are accounted for, there are relatively little variations among European groups in university completion rates, with those from " other Europe," and Italy having significantly higher rates, while those from the United States and Germany having significantly lower rates than children of Canadian-born parents (Table A.1). In comparison, the variations remain relatively large among non-Western groups. While immigrant youth whose parents are from China and India maintain their significant advantages in university completion rates over children of Canadian-born parents, when the selected sociodemographic and family background variables are accounted for, immigrant youth from the Philippines show disadvantages relative to children of Canadian-born parents (Table A.1 and the third column in Table 4).

Comparing the fourth with the first column in Table 4 shows that differences in demographic factors, parental education, and group human capital account for part of the advantages in university completion rates among some Asian groups. With the average characteristics as the whole sample and assuming the control variables had same effects on the outcome across all groups, the university completion rates would reduce by 69% for children of immigrants from the Philippines, 34% for those from Africa, 30% for those from India, about 35% for those from West Asia/Middle East, and 15% for Chinese immigrants. The results in the fifth to tenth columns suggest that youth of Asian immigrant parents generally benefit from the fact that their fathers' generation tends to be concentrated in metropolitan areas, and high levels of parental education (except Chinese immigrants), and high average educational levels in their fathers' generation. Controlling for demographic factors, parental education, and group human capital had little impact on the university completion rate among Caribbean and Latin American youth, but it tends to lessen the advantage of youth of parents from Africa.

Group differences in demographic factors, parental education, and group human capital also account for a large part of the disadvantages in university completion rates among some European groups. With the average characteristics as the whole sample, the university completion rates would increase for the following national origin groups: 50% for Portugal, 10% for the Netherlands, and 26% for Italy.. Low levels of parental education and average educational levels among the father's generation are the most important variables contributing to the relatively low university-completion rates among the Portuguese.

4.2.2 Group differences in the effects of major explanatory variables

For all groups as a whole, we observe that women tend to have higher university completion rates than men (Table A.1). Youth who did not live with parents, or lived with a lone mother by age 15, have lower university completion rates than those who mainly lived with two biological parents. Youth who lived in large metropolitan areas have higher university completion rates than those who lived in rural areas or small towns. Both mothers' and fathers' educational levels significantly predict youth's university-completion rates.

The effects of mother tongue and family language environment are not statistically significant (Table A.1). This is understandable, since children of immigrants in our study sample finished all their formal education within the Canadian educational system, and should not have had difficulties with the host-country language. It should be noted that many immigrants who came to Canada since the 1970s spok neither English nor French. Although the language difficulty may affect these immigrants' economic performance, the non-English/French family environment had no direct impact on their children's university educational attainment.

Table 5 summarizes group differences in the effects of the major explanatory variables based on separate analysis for non-Western and Western source region groups in Tables A.2 and A.3. Table A.2 shows models separately for non-Western and Western source region groups as well as for children of Canadian-born parents while Table A.3 presents each model separately for five large source region/country groups—China, India, the Caribbean, the United Kingdom and Italy. This allows us to determine the extent to which the predictors are more important for some national origin groups and not for others. In order to show whether the different effects of some explanatory variables across groups are statistically significant, Table A.4 provides the T values for the difference in model parameters for different groups.

As Table 5 shows, males of Caribbean origin are less likely to attain a university degree than their female counterparts. The female advantage is also noted for the Chinese and the Italians, with a 7- to 15-percentage-point difference with their male counterparts. The reverse is observed among those from India (not statistically significant), where 67.0% of males versus 63.5% of females are predicted to obtain a university degree.

The influence of the father's education matters more for Western country origin groups than for youth with non-Western origins. The greatest predicted percentage difference was observed among the Italians, with 30.2% obtaining a university degree for those whose fathers have less than a high school education. This rises to 70.2% for those whose fathers have a university education. Father's university education also makes a large difference for youth of immigrant parents from India, the Caribbean and the United Kingdom. Regardless of whether Chinese fathers only had a high school or a university education, there was little difference in the percentage of those who obtained a degree—69.3% and 77.3%, respectively. Mothers' education makes a large difference for university attainment among youth of immigrant parents from the United Kingdom, and to a lesser extent, from China. The effect of mothers' education is not significant or consistent for youth of immigrant parents from the Caribbean, India and Italy.

Family language environment shows different effects for the non-Western and Western origin groups, with minority-language retention showing to be not beneficial for Western origin groups. Almost 73% of the Chinese youth who retained their native language during childhood attained a university degree in comparison with 55% of those who spoke only either English or French with their parents. This pattern also holds for youth of immigrant parents from India.

Table 5
Group differences in the effects of some explanatory variables