ARCHIVED – Health Status and Social Capital of Recent Immigrants in Canada: Evidence from the Longitudinal Survey of Immigrants to Canada

Regression Results and Discussion

Table 4 presents the survey means and standard errors of the variables used in our empirical models for all immigrants. Time variant variables and time invariant variables are differentiated in the table. The regression results from the GEE models are reported in Table 5.

Table 4:   Survey Means of Variables in the Final Specification Estimation, All Immigrants

N=7656; n=22320

Time invariant variables

Time variant variables

Source: LSIC (2005).

Let us first look at the regression results for all the LSIC immigrants (see Table 5). 

Column 1 reports the regression results of the specification without social capital variables. The estimated effects of the demographic variables (e.g., age and gender) are consistent with the theoretical explanations and the findings from the existing empirical studies. Males are more likely to report good health than females. Immigrants in the older age groups are more likely to rate their health status as fair or poor. The marginal effects related to the 15-19 age group (reference group) decline as age increases. Immigrants in the oldest age group of 65 and over are less likely to rate their health status as healthy compared to those aged 15-19. However, region of birth does not have a significant effect on the health status of recent immigrants. 

As discussed previously in the descriptive analysis, the health status varies across immigration categories. The regression results confirm that the health status is significantly different across immigration categories, controlled for other characteristics. Compared to family class immigrants, skilled worker principal applicants are more likely to report good health, while refugees are more likely to report poor health.

In terms of perceptions of official language ability variables, health status varies significantly across groups. Compared with no official language ability, being proficient in English is associated with a higher likelihood of reporting as healthy. However, the ability to speak French does not have a significant effect on health. The effect of level of education at landing is also not statistically significant during the initial four years after landing.

For the variables of accessibility to the Canadian health care system, immigrants who have had problems accessing health care services are more likely to rate their health as fair or poor. This may reflect the role of immigrants’ ability to effectively identify and access health care services (including preventive care) in positive perceptions of health status.

Looking at family income quartiles, there are quite large differences in health status across groups. Immigrants in the lowest family income quartile are more likely to report poor health. Furthermore, the employment status of both respondents and their spouses is positively associated with immigrants’ health. 

When including time indicators in the regressions, we confirm the “healthy immigrant effect” – when compared with the situation in wave 1, immigrants are less likely to report being in good health in wave 2 and wave 3 (the likelihoods are 3 percent points and 4.5 percent points, respectively, lower than that in wave 1). [ Note 9 ]

Columns 2 to 4 of Table 5 estimate models with social capital variables. As an initial step, column 2 only adds general indicators of the existing networks including relatives and friends in Canada upon landing, and development of new networks after landing captured by whether an immigrant made new friends in Canada. The addition of these indicators does not change other effects much. Also, making new friends in Canada shows a positive relationship with health status of immigrants. To further investigate which elements play a role among networks, column 3 includes all network content indicators in the model. The results confirm what column 2 indicates – friendship networks matter, particularly frequency of contact and ethnic diversity of the networks. Column 4 presents the final specification with social capital effects.

In terms of social capital variables, our results indicate that friendship networks have a significant effect on the respondents’ self-reported status of health. Both the frequency of contact with friends and the ethnic diversity of friends have significant and positive effects on immigrants’ health.

Immigrants who have more diverse friendship networks and who are in contact with their friends more frequently are more likely to report being in good health. However, it is important to note that we do not find significant effects of family and relative networks or group and organization networks on health for all immigrants.

Table 5:   GEE Population-Averaged Estimations of Probability of Being Healthy for All Immigrants in the First Four Years in Canada

Notes: * p<0.1; ** p<0.05; *** p<0.01; marginal effects for dummy variables are for discrete change from 0 to 1; reference categories are in brackets.

Source: LSIC (2005).

Regression results for family class immigrants

In order to investigate the different ways in which social capital affects the health status of immigrant sub-groups, we also present the regression results of our GEE model for the reference group of family class immigrants (see Table 6). 

Noteworthy regression results indicate that family and relative networks, friendship networks, and group and organization networks all have significant effects on the health status of recent family class immigrants.

Unlike the results from the model for all immigrants, for each of the three networks the effects of frequency of contact with network members or units are all statistically significant. Compared to those who do not have contact with friends or take part in organizational activities regularly, family class immigrants who interact with friends or groups on a daily basis are more likely to report that they are healthy.

However, frequency of contact with family sponsors is associated with a lower likelihood of reporting a positive health status. This may be partly due to the fact that a large proportion of family class immigrants are sponsored parent and grandparent immigrants (PGPs), who tend to be much older than the average family class immigrant. [ Note 10 ] The LSIC shows that the majority of PGPs lived with their family sponsors during the initial years after landing, and elderly PGPs are more likely to live with their sponsors (Zhao 2007b). In addition, a significant number of PGPs (37 percent at six months after landing, and 34 percent at two years after landing) who live with their sponsors provide unpaid labour for their sponsors, such as maintaining a house and caring for family members, which might be a factor that negatively affects their health. Another possible explanation for this finding may rest in the negative potentiality of social capital. As indicated in the literature review, social capital is not inherently positive; it may be the case that the frequency of contact with family members is connected to increased demands for time, resources, and energy by these networks. These excessive demands may adversely affect the health of recent immigrants.

The regression results also indicate that social network size and social network diversity do not have any significant effect on family class immigrants’ health.

Table 6:   Social Capital Effects on Probability of Being Healthy in the Initial Four Years in Canada, Family Class Immigrants

Notes: * p<0.1; ** p<0.05; *** p<0.01. Reference categories are in brackets; marginal effects for dummy variables are for discrete change from 0 to 1; GEE population-averaged regression is used. The regression also includes controls for sex, age group, area of birth, incidence of problems accessing Canadian health care system, CMA of residence, ability of official languages, family income, employment status of the respondent and the spouse, and education. 

Source: LSIC (2005).

⁹ To investigate the time effects, we control for two types of time variables: (1) number of weeks in Canada since landing (continuous variable), and (2) wave variables (dummy variables). Both methods provide us with similar regression results: the respondents’ self-rated health status is significantly and negatively related to increased time in Canada. This is consistent with our findings in the descriptive analysis that the health of recent immigrants declines over time. In Table 5 we only report results from the models with wave indicators used.

¹⁰ In the LSIC, PGPs account for 34 percent of family class immigrants. In 2001, the average age at landing of family class immigrants was 34 years, while the average age at landing of PGPs was 52 years – calculations based on data extracts on 31 March 2009 from CIC’s Permanent Resident Data System (PRDS).

• Previous
• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8 (current)
• 9
• 10
• 11
• 12
• Next