Growing Up Canadian - A Study of New Immigrant Children

4. Results

4.1 Description of the Study Population

In 1994, there were 3,129,038 children in Canada between the ages of 4-11 years. Of this total, 7.7 percent, or 240,184 were new immigrants or children of new immigrants. Figure 1 describes the prevalence of poverty among children in the new immigrant and national population groups. The rate of poverty among new immigrant families with at least one child between the ages of four and eleven was more than double that of Canadian families with a child or children in the same age group.

Figure 1 - Distribution of Poverty in New Immigrant and National Populations

4.2 Mental Health of New Immigrant Children

The reliability of each mental health scale was established for each population group. Coefficient Alphas for each of the scales in each of the groups were in the acceptable range, as follows:

Hyperactivity: national = 0.84; new immigrant = 0.79
Emotional disorder: national = 0.74; new immigrant = 0.68.
Conduct disorder: national = 0.78; new immigrant = 0.66.

Figure 2 summarizes the rates of hyperactivity, emotional disorder and conduct disorder among the new immigrant and national sample children.

The rates for hyperactivity and emotional disorder in the national sample of children ages 4-11 yrs. (10.8%, 9.4% respectively) closely approximate those reported by Offord & Lipman (1996). The rates of these disorders among new immigrant children are noticeably lower (3.8% and 5.3% respectively). Our national rate of conduct order (13.2%) was slightly higher than the 10 percent reported by Offord & Lipman (1996); the new immigrant rate for this disorder was substantially lower (6.7%).

Figure 2 - Prevalence of Mental Health Outcomes in New Immigrant and National Populations

4.3 Poverty and Mental Health

Figure 3 describes the prevalence of hyperactivity by poverty status in new immigrant and national groups. In the national group, the rates of hyperactivity were higher among the Poor than among the Non-Poor (15.8%, 10.1% respectively). However, because of sample size constraints, very few cases in the new immigrant group met the 10 percent threshold criterion. Therefore, the estimated prevalence of hyperactivity may not be reliable for this group.

Figure 3 - Prevalence of Hyperactivity by Poverty Status and Population Group

The prevalence of emotional disorder for each population group by poverty status is illustrated in Figure 4. In both groups, this disorder was far more prevalent among the Poor than among the Non-Poor.

Prevalence rates for conduct disorder by poverty status in new immigrants and the national population are presented in Figure 5. Once again, among both the national and new immigrant groups, the prevalence of disorder was higher among the Poor as compared to the Non-Poor.

Figure 4 - Prevalence of Emotional Disorder by Poverty Status and Population Group

Figure 5 - Prevalence of Conduct Disorder by Poverty Status and Population Group

4.4 Poverty, Risk and Protective Factors by Population Group

Table 1 describes the distribution of risk and protective factors according to population group and poverty status. T-tests were used to analyse the significance of differences in scale scores (family drinking problem, parent depression, family functioning, family drinking problem, parenting behaviours) and Chi Square tests for differences in the distribution of categorical variables (single parent status, child care by parents). Prior to performing these analyses, the reliability of scales in the new immigrant and national populations were examined and found to be acceptable (range 0.7 - 0.9).

Compared to new immigrants, families in the national population were significantly more likely to have a drinking problem, family dysfunction, hostile parenting behaviour and to use non-parent child care. Group differences in parental depression scores, levels of positive parenting and in the proportion of single parent families were statistically non-significant.

With the exception of positive parenting, there were significant differences between Poor and Non-Poor population groups on each of the variables. Parents in Poor families were significantly more likely than non-poor parents to be depressed, to exhibit hostile parenting, to use non-parent care and to be single parents. Poor families were more likely to have a drinking problem and to have poor levels of family functioning than their Non-Poor counterparts.

An examination of the distribution of mediating variables by poverty status within the new immigrant and national populations (Table 2) suggests that the greatest differences are between the Poor and Non-Poor national families (columns 2 and 4). Children in Poor national sample families were far more likely than their Non-Poor counterparts to be exposed to mental health risk factors. Positive parenting is the only variable in which no significant difference is observed. Although differences between Poor and Non-Poor new immigrant families (columns 1 and 3) follow the same trend as those found in the general population, there are fewer statistically significant differences and the contrasts between Poor and Non-Poor are not as marked. There are important differences between the Poor new immigrant and Poor national groups (columns 1 and 2). The rates for parental depression, family drinking, hostile parenting and single parent families were significantly lower in Poor new immigrant families compared to Poor national families. With the exception of hostile parenting, there were no significant differences between the Non-Poor new immigrant and national populations (columns 3 and 4).

Table 1 - Distribution of Mediating Factors in Population Groups and by Poverty Status
	New Immigrant (N=1143)	National (N=13848)	Statistical Test	Poor (N=2179)	Non-Poor (N=12812)	Statistical Test
Scales (means)
Parent Depression	4.96	4.74	t=1.26 p=.205	7.80	4.24	t=20.8 p=.000
Family Drinking Problem	0.40	0.45	t=2.76 p=.006	0.61	0.42	t=10.5 p=.000
Family Functionning	8.49	7.90	t=3.56 p=.000	9.73	7.65	t=16.4 p=.000
Positive Parenting Behaviour	12.59	12.79	t=1.86 p=.063	12.75	12.78	t=.389 p=.697
Hostile Parenting Behaviour	8.06	8.92	t=6.77 p=.000	9.22	8.80	t=4.45 p=.000
Categorical Variables (%)
Use of Non-Parent Care	75.4%	70.1%	x ²=14.6 p=.000	85.5%	67.9%	x ²=280.9 p=.000
Single Parent Family Status	16.3%	16.3%	x ²=.001 p=.975	55.1%	9.7%	x ²=2874 p=.000

4.5 Summary of Descriptive Analyses

According to the NLSCY data, new immigrant children in Canada have fewer mental health problems than native-born Canadians. Poverty creates a situation of mental health risk for both groups. For both groups, poverty is associated with mental health risk factors including parental depression, family drinking problems, family dysfunction, hostile parenting behaviour, use of non-parent child care and single parent family status. However, the pattern of comparisons between Poor and Non-Poor new immigrants as well as between Poor immigrants and Poor nationals suggests that poor new immigrant children are less disadvantaged than their counterparts in the poor national sample.

Table 2 - Distribution of Mediating Factors within Population Groups by Poverty Status
	Poor New Immigrant (N=352) [1]	Poor National (N=1827) [2]	Non-poor New immi. (N=791) [3]	Non-poor National (N=12021) [4]	Test of Difference [1] - [2]	Test of Difference [3] - [4]	Test of Difference [1] - [3]	Test of Difference [2] - [4]
Scales (means)
Parent Depression	6.23	8.16	4.55	4.25	t = 6.0 x=.000	t = 1.4 x=.135	t = 4.8 x=.000	t =28.4 x=.000
Family Drinking Problem	0.47	0.63	0.39	0.43	t = 4.3 x=.000	t = 2.1 x=.034	t = 2.1 x=.035	t =10.5 x=.000
Family Dysfunctioning	10.21	9.76	7.92	7.67	t = 1.5 x=.127	t = 1.3 x=.208	t = 6.9 x=.000	t =15.2 x=.000
Positive Parenting	12.39	12.79	12.56	12.78	t = 1.8 x=.079	t = 1.9 x=.059	t = .72 x=.474	t = .13 x=.900
Hostile Parenting	8.04	9.43	8.04	8.84	t = 6.9 x=.000	t = 5.1 x=.000	t = .02 x=.985	t = 6.1 x=.000
Categorical Variables (%)
Child Care by Parents	84.5%	85.7%	71.3%	67.7%	x=.308 x=.579	x=4.56 x=.033	x=23 x=.000	x=249 x=.000
Single Parent Family Status	33.0%	59.4%	8.9%	9.8%	x=85 x=.000	x=.566 x=.452	x=105 x=.000	x=2931 x=.000
Note: As multiple comparisons inflate the probability of Type I errors, the Bonferroni procedure suggests that p < 0.01 be used as the level of significance in this table.

4.6 Multivariate Analyses

Multivariate analyses were used to examine whether the association between poverty and family characteristics might explain the relationship between poverty and children's mental health. These analyses included age, gender and hypothetical mediating variables. "Family drinking problem" was dropped from these multivariate analyses because it was very highly correlated with family dysfunction. Age and gender were included because earlier analyses had demonstrated significant differences in mental health outcomes in girls and boys. Other studies have reported higher rates of conduct and emotional disorder among children 8 to 11 compared to children aged 4 to 7 (Offord & Lipman, 1996).

Nine multiple linear regression models were prepared for each of the mental health variables: Hyperactivity, Emotional Disorder and Conduct Disorder. Model 1 included poverty and immigrant status. Because the interaction term between poverty and new immigrant status became significant after controlling for the mediating variables, this interaction term is included in Model 1 for Emotional Disorder and Conduct Disorder. In Models 2 to Model 8, age and gender (female), parent depression, family dysfunction, positive parenting, hostile parenting, child care by parents, and single parent family were separately added to Model 1. Model 9 includes all relevant variables.

The differences in R-square between each of these models and model 1 indicate the contribution of the added variable(s) in explaining the variance of the mental health outcome over and above the effect of poverty and new immigrant status. In addition, the differences between each of these models and Model 1 in the coefficients of Poor, New Immigrants, and their interaction term indicate the modifying effect of the added variable(s).

4.6.1 Hyperactivity

Models for hyperactivity regressed on the predictor variables appear in Table 3. According to Model 1, poverty significantly increased, and new immigrant status significantly reduced hyperactivity scores. The R² for Model 1 (.013 or 1.3%) suggests that poverty and new immigrant status explain a very small proportion of the variance in hyperactivity. The R²s in Model 6, which includes hostile parenting (.202), and in Model 3, which includes parental depression (.052), suggest that hostile parenting and parental depression make a substantial contribution to explaining the variance in hyperactivity.

Following changes in the effect of poverty on hyperactivity from model to model suggests that parental depression (resulting in a reduction of 51.6% in Model 3), single parent family status (-47.3% in Model 8), family dysfunction (-23.7% in Model 4), and hostile parenting (-22.6% in Model 6) account for a large proportion of the mean difference in hyperactivity between the Poor and Non-Poor. Hostile parenting reduces a large proportion of the mean difference between the new immigrant and national populations (-34.2% in Model 6), suggesting that the relatively low prevalence of hostile parenting behaviours in new immigrant families helps account for new immigrant children's mental health advantage. As shown in Model 9, the joint effects of all the selected mediating variables reduces the difference in mean scores of hyperactivity between the Poor and Non-Poor by 74.2 percent and reduces the difference between new immigrants and the national population by 29.7 percent.

Model 9 also indicates that younger children, girls, and children in families using non-parent child care are less likely to be hyperactive than older children, boys and children in families where care is primarily provided by parents. Parental depression, family dysfunction, hostile parenting behaviour, and single parent family status are significantly associated with the occurrence of hyperactive behaviour.

Table 3 - Regression Models For Hyperactivity
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
Constant	4.55***	6.02***	3.98***	3.72***	5.85***	0.95***	4.94***	4.46***	2.19***
Poor	0.93***	0.91***	0.45***	0.71***	0.93***	0.72***	1.03***	0.49***	0.24**
New Immigrant	-1.11***	-1.20***	-1.05***	-1.14***	-1.13***	-0.73***	-1.10***	-1.02***	-0.78***
Age		-0.11***							-0.09***
Female		-1.28***							-1.06***
Parent Depression			0.13***						0.07***
Family Dysfunction				0.11***					0.03***
Positive Parenting					-0.10***				0.001
Hostile Parenting						0.41***			0.36***
Childcare by Parents							-0.58***		-0.25***
Single Parent Family								0.94***	0.55***
R²	0.013	0.049	0.052	0.037	0.020	0.202	0.018	0.020	0.244
% Change in the effect of ¹
poor		-2.1	-51.6	-23.7	0	-22.6	10.8	-47.3	-74.2
new immigrant		8.1	-5.4	2.7	1.8	-34.2	-0.9	-8.1	-29.7
Note: ¹ defined as ((coefficient in model i/coefficient in model 1)-1)100. significant at " x < 0.05; x < 0.01; and * x < 0.001.

4.6.2 Emotional Disorder

Models for the regression of emotional disorder on the predictor variables appear in Table 4. As in the case of hyperactivity, Model 1 indicated that poverty significantly increased and new immigrant status significantly reduced the risk of emotional disorder. The interaction term between poverty and new immigrant status was not significant; however, it has been included in this model because this interaction term became significant after controlling for modifier variables. The R² for Model 1 (.013) suggests that poverty and new immigrant status explain only a very small proportion of the variance in emotional disorder. The R²s in Model 6, which included hostile parenting (.139), and in Model 3, with parent depression (.093), indicates that hostile parenting and parent depression make substantial contributions to explaining the variance in emotional disorder.

The interaction term between poverty and new immigrant status became significant in Model 3 (parent depression), Model 6 (hostile parenting), and Model 8 (single parent family), suggesting that these three variables significantly modify the impact of poverty and new immigrant status on emotional disorder.

To illustrate the modifying effects of the control variables, mean scores for emotional disorder were calculated for children in 4 family groups: Poor immigrant, Poor national, Non-Poor immigrant, and Non-Poor national. The changes in mean score differences between contrasting groups between Model 1 and each of the subsequent models were calculated and appear in the bottom portion of Table 4. According to these data, parental depression, single parent family status, and to a lesser extent, family dysfunction and hostile parenting account for a large proportion of the difference in mean Poor and Non-Poor national population emotional disorder scores (79.5%, 61.9%, 24.6% and 20.8%, respectively). Parental depression, hostile parenting, and single parent family status reduced a large proportion of the difference in emotional disorder between Poor immigrants and Poor nationals (72.0%, 72.0%, and 54.7%, respectively) .

Table 4 - Regression Models For Emotional Disorders
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
Constant	2.54***	1.42***	1.96***	1.89***	4.34***	0.42***	2.68***	2.46***	-0.72***
Poor	0.68***	0.73***	0.14***	0.51***	0.68***	0.54***	0.72***	0.26***	-0.03
New Immigrant	-0.77***	-0.71***	-0.82***	-077***	-0.79***	-0.58***	-0.76***	-0.77***	-0.57***
Poor x New Immigrant	0.30	0.22	0.69***	0.25	0.28	0.45***	0.30	0.56***	0.72***
Age		0.15***							0.15***
Female		0.05							0.17***
Parent Depression			0.14***						0.10***
Family Dysfunction				0.08***					0.01***
Positive Parenting					-0.14***				-0.01
Hostile Parenting						0.24***			0.21***
Childcare by Parents							-0.20***		-0.23***
Single Parent Family								0.85***	0.43***
R2	0.013	0.029	0.093	0.041	0.016	0.139	0.014	0.024	0.208
Mean scores of emotional disorders¹
poor immigrant	2.76	2.76	2.64	2.56	2.72	2.96	2.79	2.65	2.79
poor national	3.22	3.25	2.77	3.08	3.23	3.25	3.25	2.86	2.64
non-poor immigrant	1.77	1.81	1.81	1.79	1.76	1.77	1.77	1.83	2.03
non-poor national	2.54	2.52	2.63	2.56	2.55	2.53	2.53	2.60	2.60
% changes in difference between ²
poor vs. non-poor immigrant		-3.6	-15.7	-22.0	-2.5	0.5	3.9	-16.8	-23.1
poor vs. non-poor national		7.0	-79.5	-24.6	-0.3	-20.8	5.4	-61.9	-95.0
poor, immigrant vs. national		5.6	-72.0	12.1	9.9	-72.0	-0.4	-54.7	-132.1
non-poor, immigrant vs national		-7.4	6.9	0.9	3.0	-24.4	-0.8	0.4	-25.2
Note: ¹ - calculated based on the corresponding regression model, assuming that all the control variables equal their means. ²- defined as: ((mean difference between the two contrasting groups in model i/mean difference between contrasting two contrasting groups in model 1)-1)100 significant at x < 0.05; x < 0.01; and * x < 0.001.

Model 9 illustrates that the joint effects of the mediating variables reduce the difference in emotional disorder between Poor and Non-Poor nationals population by 95 percent. These mediating variables reversed the differences between the Poor national and Poor immigrant families. These effects can also be observed in Figure 6, which presents new immigrant and national group mean scores for Model 1 (unadjusted) and Model 9 (adjusted). Unadjusted mean score differences between Poor and Non-Poor, and between New Immigrant and National groups are pronounced. However, with the mediating variables controlled, differences between Poor and Non-Poor national families disappear (as indicated by the virtually flat line in Figure 6), and the slight difference between Poor national and Poor new immigrant families is statistically non-significant. However, the mental health disadvantage for Poor versus Non-Poor immigrants, and for Non-Poor nationals compared to Non-Poor immigrants remains statistically significant.

Figure 6 - Emotional Disorders in Population Groups: Effect of Mediating Variables

This pattern of findings suggests that the effect of poverty on emotional disorder of children in national families is mainly indirect, that is, that it operates through associated parental and family characteristics. The overall difference in children's emotional disorder between Poor and Non-Poor national families might be primarily attributed to the disadvantages experienced by Poor national families, including elevated rates of parental depression, single parent family status, family dysfunction, and hostile parenting. In new immigrant families, the effect of poverty appears to be more direct. Family characteristics apparently explain very little about the higher risk of emotional disorder among children in Poor new families relative to children in Non-Poor families.

In addition, it appears that new immigrant status is salient for Non-Poor families, but not for Poor families. Children in Non-Poor new immigrant families were less likely to have emotional disorders than those in the Non-Poor national population, and this difference remained statistically significant even after controlling for the mediating variables. This suggests that the risk of exposure to mental health risk factors in Non-Poor national and new immigrant families is very similar. The effect of mediating variables in reducing the differences in children's emotional disorder between Poor immigrant families and Poor national families indicates that poverty in national families is much more likely to be associated with parental depression, hostile parenting, and single parent family status.

Model 9 also revealed that age (older), gender (female), parent depression, family dysfunction, hostile parenting, use non-parent care, and single parent family status were associated with a higher risk of emotional disorder. With other mediating variables taken into account, the effect of positive parenting was not significant.

4.6.3 Conduct Disorder

Regression models for conduct disorder appear in Table 5. Again, Model 1 indicates that poverty significantly increases and new immigrant status significantly reduces the risk of conduct disorder. The interaction term between poverty and new immigrant status was not significant. The R² in Model 1 (.011) suggests that poverty and new immigrant status explained a very small proportion of variance in children's emotional disorder. The R²'s in Model 6, which includes hostile parenting (.190), and in Model 3, with parent depression (.036), indicates that hostile parenting and, to a lesser extent, parental depression add substantially to explaining the variance of Conduct Disorder.

Considered separately, none of the mediating variables had a strong enough effect to produce a significant interaction term between poverty and new immigrant status. However, as indicated by the changes in the mean score differences (bottom portion of Table 5), single parent family status, parent depression, as well as hostile parenting and family dysfunction reduce the difference in children's conduct disorder between Poor and Non-Poor national families (54.9%, 47.7%, 27.2%, and 21.2% in Models 8, 3, 6, and 4 respectively). Hostile parenting, single parent family, and parent depression also attenuate the difference in children's conduct disorder between Poor immigrant families and Poor national families (54.1%, 28.7% and 25.1% in Models 6, 8, and 3 respectively). With the inclusion of all the mediating variables in Model 9, the interaction term between poverty and new immigrant status becomes statistically insignificant. Comparing Model 1 to Model 9, these joint mediating effects reduce the difference in Conduct Disorder between Poor and Non-Poor national population by 95.3 percent, and reduce the difference between the Poor national population and Poor immigrants by 77.9 percent.

Table 5 - Regression Models For Conduct Disorder
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
Constant	1.35***	2.02***	1.12***	0.96***	1.86***	-0.47***	1.41***	1.31***	- 0.25*
Poor	0.46***	0.45***	0.24***	0.36***	0.46***	0.33***	0.47***	0.21***	0.02
New Immigrant	-0.50***	-0.54***	-0.52***	-050***	-0.51***	-0.34***	-0.50***	-0.50***	-0.39***
Poor x New Immigrant	0.04	0.007	0.12	-0.07	-0.004	0.10***	-0.04	0.12***	0.28*
Age		0.06***							-0.05***
Female		0.49***							-0.37***
Parent Depression			0.06***						0.02***
Family Dysfunction				0.08***					0.01***
Positive Parenting					-0.04***				0.01*
Hostile Parenting						0.21***			0.19***
Childcare by Parents							-0.08***		-0.10**
Single Parent Family								0.50***	0.36***
R²	0.011	0.032	0.036	0.031	0.015	0.190	0.011	0.018	0.214
Mean scores of emotional disorders¹
poor immigrant	1.28	1.23	1.25	1.16	1.29	1.45	1.28	1.21	1.31
poor national	1.81	1.78	1.65	1.73	1.81	1.69	1.82	1.59	1.42
non-poor immigrant	0.86	0.79	0.89	0.87	0.84	1.02	0.85	0.89	1.01
non-poor national	1.35	1.33	1.41	1.37	1.35	1.36	1.35	1.39	1.40
% changes in difference between ²
poor vs non-poor immigrant		4.7	-14.9	-31.2	7.8	0.7	1.7	-23.4	-29.7
poor vs non-poor national		-2.0	-47.7	-21.2	0.2	-27.2	2.4	-54.9	-95.3
poor, immigrant vs national		2.4	-25.1	6.4	-3.7	-54.1	1.1	-28.7	-77.9
non-poor, immigrant vs national		8.4	4.4	0.8	2.4	-32.3	0.4	0	-20.9
Note: ¹ - calculated based on the corresponding regression model, assuming that all the control variables equal their means. ²- defined as: ((mean difference between the two contrasting groups in model i/mean difference between contrasting two contrasting groups in model 1)-1)100 significant at x < 0.05; x < 0.01; and * x < 0.001.

Figure 7 presents the mean scores for Model 1 and Model 9 for each of the four family groups. Additional tests indicate that, after controlling for the mediating variables, differences in children's Conduct Disorder between Poor and Non-Poor national families and between Poor immigrant and Poor national families are no longer significant. However, differences between Poor immigrants and Non-Poor immigrants and differences between Non-Poor immigrants and Non-Poor national families persist.

Figure 7 - Conduct Disorders in Population Groups: Effect of Mediating Variables

Similar to the patterns associated with Emotional Disorder, the overall difference in children's Conduct Disorder between Poor and Non-Poor national families may be explained by the disadvantages experienced by Poor national families such as increased likelihood of parental depression, single parent family status, family dysfunction, and hostile parenting. Among national families, the direct effect of poverty on children's conduct disorder is not observed. Poverty is more closely associated with higher levels of disadvantage in the family environment which increases the risk of this outcome. By contrast, in new immigrant families, the effect of poverty is more direct. Family characteristics do little to explain differences in children's conduct disorder among Poor and Non-Poor new immigrant families.

Once family characteristics are taken into account, children in Non-Poor new immigrant families are less likely to have Conduct Disorders than their counterparts in national families. The differences in children's Conduct Disorder between Poor immigrant families and Poor national families disappears after controlling for mediating variables, suggesting that Poor national families are at a disadvantage in terms of parent depression, hostile parenting, and single parent family status over Poor immigrant families.

Model 9 in Table 5 also suggests that age (younger), gender (male), parent depression, family dysfunction, hostile parenting, and single parent family status are risk factors for conduct disorder. Positive parenting is significantly associated with a reduced level of conduct disorder without control for other mediating variables. However, with other mediating variables taken into account, the relationship between positive parenting and conduct disorder becomes marginally significant.

Further analyses indicated that the relationship between positive parenting and conduct disorder changed from negative to positive, once hostile parenting was controlled. At the bivariate level, hostile parenting was negatively related to positive parenting (r=-.23, p<.01), but hostile parenting was more strongly associated with children's conduct disorder (r=.43, p<.01) than positive parenting (r=-.07, p<.01). After the overlap effect of hostile parenting and positive parenting was taken into account, the weakly significant yet positive relationship between positive parenting and children's conduct disorder probably indicates the effect of over-protection or excessive indulgence. This may also explain why child care provided by parents is associated with a higher risk of Conduct Disorder.

In summary, several consistent patterns emerged from the multivariate analyses.

First, parent depression, single parent family status, and to a lesser degree, hostile parenting, and family dysfunction, explain a large proportion of the differences in children's mental health in Poor and "Non-Poor families. In the cases of Emotional Disorder and Conduct Disorder, the modifying effects of these variables were manifested primarily between Poor and Non-Poor national families. Similarly, hostile parenting, followed by parent depression and single parent family status, explained a large proportion of the differences in children's mental health between new immigrant and national families. In the cases of Emotional Disorder and Conduct Disorder, the modifying effects of these variables were limited to Poor new immigrant and Poor national families.

Second, parent depression, family dysfunction, hostile parenting, and single parent family status had deleterious impacts on children's mental health, while the effects of positive parenting and child care type were weak and varied with mental health outcomes. Among the mediating variables, hostile parenting and parental depression were the most strongly and consistently associated with variations in mental health outcomes.

Contents

Growing Up Canadian - A Study of New Immigrant Children - October 1998