Morton Beiser, Feng Hou, Violet Kaspar, and Samuel Noh

August, 2000

Culture, Community and Health Studies

Centre for Addiction and Mental Health

Department of Psychiatry, University of Toronto

250 College Street, Toronto, ON M5T 1R8

Tel: (416) 979-4988

Fax: (416) 979-0564

E-mail: morton_beiser@camh.net

Project funded by Human Resources Canada

 

Abstract

 

The present study examined the effects of movement into or out of poverty on developmental behaviours of 4-to-9 year old children. Analyses were based on a longitudinal sample of 6200 children from the first two cycles of the National Longitudinal Survey of Children and Youth, covering the period from 1994/95 to 1996/97. Results indicated that among non-immigrant families, disadvantages of children living in persistent poverty sustained over time. Simply moving out of poverty was not sufficient to improve children=s developmental outcomes unless it was accompanied by a substantial improvement in living standard. Moreover, small gains in absolute income might elevate the behavioural problems of children living in persistent poverty. Absolute income increases among persistently poor non-immigrant families were related to new employment of the parents and decreases in welfare dependence. The newly employed parents could not make enough income to pull the family out of poverty, they were also probably no longer able to spend a lot of time with the child. By comparison, among persistently poor immigrant families, absolute income increases tended to reduce children=s behavioural problems.

Immigrant and Non-Immigrant Children's Developmental Outcomes

Research has consistently demonstrated that poverty is a risk factor that threatens the health status, cognitive abilities, behaviours, and educational attainment of children (Aber et al. 1997; Lipman and Offord 1995). However, the reasons why poverty affects some, but not all poor children remain unclear.

Developmental psychologists examining resilience and vulnerability have identified parental characteristics, negative events, as well as personal and familial resources as factors that mediate the effects of poverty (Brooks-Gunn 1995; Elder, Nguyen and Caspi 1985; Huston 1991; McLoyd 1989). Social-economic studies attempt to explain differential effects of poverty by looking at the persistence and severity of poverty, sources of income, allocation of financial and non-monetary resources within the family, and contextual factors such as immigrant status, race, ethnicity, and neighbourhood characteristics(Blau 1999; Guo 1998; Lefebvre and Merrigan 1998; Mayer 1997; McLeod and Shanahan 1996). In order to develop effective policy, it will be necessary to adopt a research approach sensitive to both developmental issues and contextual factors.

Using data from the first two cycles the National Longitudinal Survey of Children and Youth (NLSCY), this present study draws from both developmental and social-economic approaches to examine the effects of changes in poverty status on children=s behavioural outcomes. Controlling for changes in parenting and other familial characteristics, this study compares children=s developmental behaviours among families in four different situations: a. Families poor over both cycles, b. Families which change from poor at time 1 to non-poor at time 2, c. Families which change from non-poor to poor, and d. Families which are non-poor in both cycles. We hypothesize that children in persistently poor families will have the worst developmental outcomes, followed by those living in families that recently became poor.

This study further examines the extent to which the effects of changes in poverty status depend on the amount of change in family income and the sources of change. We hypothesize that the amount of income change may moderate the effects of changes in poverty status: the larger the increase in income associated with movement out of poverty, the better children=s developmental outcomes; the larger the decrease in the amount of income associated with movement into poverty, the worse the developmental outcomes. Furthermore, the effects of income changes may not be as profound as those of changes in family structure and employment status -- the two primary sources of income changes.

The present study also compares the effects of change in poverty status in immigrant and non-immigrant families. Studies of the effects of poverty on immigrant and receiving-society children suggest a paradox. Although immigrant families are typically poorer than their receiving country counterparts (National Council of Welfare 1998; US Department of Health and Human Services 1998), immigrant children are, on the whole, at least as healthy as majority culture children, and often out-perform them in school (Beiser et al. 1995; Hernandez 1999; Klimidis et al. 1994; Chang et al. 1995; Zhou 1997). Furthermore, evidence from many U.S. studies, and a previous study using cycle 1 data of the NLSCY suggest that relative health status tends to decrease from foreign-born children, to native-born children of immigrant parents, to children of non-immigrant parents, even though poverty rates also decrease in the same direction (Hernandez 1999; Hernandez and Charney 1998; Beiser, Hou, Hyman and Tousignant 2000). These results suggest the possibility that immigrant status protects children, at least temporarily, from many of the deleterious health consequences of poverty (Harris 1999; US Department of Health and Human Services 1998).

Previous studies using cycle 1 NLSCY data found that poverty had different concomitants in immigrant and receiving society families. Among poor majority culture families, there were higher rates of parental depression, single-parent status, and ineffective parenting than among either non-poor families in the general population or among immigrant families, whether poor or not. Poor immigrant children may have a mental health advantage over their receiving society counterparts because they have a more supportive family environment (Beiser, Hou, Hyman and Tousignant, 2000). Some US studies also found that poverty in immigrant families was not necessarily associated with characteristics that were repeatedly found in poor non-immigrant families, such as single parent status, a large number of siblings, and low rates of father=s labour force participation (US Department of Health and Human Services 1998).

Immigrant poverty is primarily due to unemployment and underemployment in the first few years of resettlement. After an initial period of high unemployment, immigrants in Canada eventually achieve higher rates of labor force participation and higher employment income than native-borns (Beiser et al. 1997; deVoretz 1995). Thus, poverty may be a transient feature of resettlement. For many receiving country families, however, poverty is probably not part of an unfolding process, but the end stage of a cycle of disadvantage. Among the majority culture, the concomitants of poverty include not only financial burden, but, in addition, social isolation and compromised self-esteem (Beiser, Johnson and Turner 1993).

Based on these findings, we hypothesize that immigrant families are more likely to move out of poverty, mainly through employment and increases in market earning, than are receiving society families. We also hypothesize that immigrant children will experience greater improvement in developmental behaviours once their families move out of poverty.

Data for this study derived from cycles one (1994/95) and two (1996/97) of the NLSCY, a nationwide study of approximately 23,000 children ranging from newborn to 11 years of age at the time of cycle one. The matched longitudinal component contained a total of 14,102 children. The study focussed on a sub-sample of 6,218 children who were between 4 and 9 years old at the time of cycle 1. Children younger than 4 years old were excluded because this group had different measures of behavioural problems at cycle one of the survey. The 10 to 11 years old age group can not be combined with the 4 to 9 years group, because the former had self-reported measures of developmental behaviours at both cycles, while the latter had parent-reported measures of developmental behaviours at both cycles. Furthermore, in the 10 to 11 years old group, the sample size of children in immigrant families was not large enough to meet Statistics Canada=s guidelines of reliable estimate.

The original NLSCY sampling strategy relied on household selection through a multi-stage stratified cluster probability sampling procedure. A weighting procedure was designed by Statistics Canada to compensate for the differential representation of population groups and was used in the present study.

For the 4 to 9 years old age group, children of immigrant families were identified to be compared against the remaining children who constitute the national comparison sample. The children of immigrant families consist of those who entered Canada as immigrants and those born in Canada into a family in which at least one of the parents was an immigrant. There were 1336 (21.5%) children of immigrant families in this age group. Among them, 260 (158 before weighting) were born in foreign countries.

Emotional problems and conduct problems constituted the study outcomes. Emotional problems, characterized by feelings of anxiety and/or depression, were measured with an eight item scale. Sample items included: Aseems to be unhappy, sad, or depressed@, Anot as happy as other children@, Atoo fearful or anxious@, Aworried@, Acries a lot@, and Aappears miserable, unhappy, tearful, or distressed@. The person most knowledgeable (PMK, mostly the mother) answered these questions, and for each question chose never or not true, sometimes or somewhat true, or often or very true. The score of the scale ranges from 0 to 16, a higher score indicating a higher level of emotional problems.

Conduct problems are characterized by aggression, either physical or indirect, or a violation of social norms. Observations making up the six item scale include Agets into many fights@, Aphysically attacks people@, Athreatens people@, and Acruel, bullies or mean to others@. The score of the scale ranges from 0 to 12.

Poverty and Changes in Poverty Status

Poverty status in the present study is a relative indicator of low income based on income adequacy, a five-category measure developed by Statistics Canada. Low-income families were defined as those belonging to the lowest or lower middle income adequacy category. The lower middle income adequacy refers to the situations when Ahousehold income is 10,000-14,999 and household size is 1-2 persons; or household income is 10,000-19,999 and household size is 3-4 persons; or household income is 15,000-29,999 and household size is 5 or more persons@ (HRDC 1996: Appendix 4 p. 69). Among all the children in the matched longitudinal sample, about 18.9% would be classified as living in low income families at cycle 1 survey.

Changes in poverty status include relative change and absolute change based on household income measured at both cycle 1 and cycle 2 of NLSCY. Relative change classifies families into four categories: Persistently Poor (poor at both cycles), Newly Poor (from non-poor at cycle 1 to poor at cycle 2), Newly Non-poor (from poor to non-poor), and Non-poor (non-poor at both cycles). Absolute change is the difference between cycle 1 and cycle 2 logarithmic average household income. Average household income is total household income divided by number of persons in the household. The logarithmic transformation of average household income is used to adjust for possible non-linearity. The difference of logarithmic incomes at cycle 2 and cycle 1 is equivalent to the logarithmic transformation of the ratio of cycle 2 income to cycle 1 income.

Changes in family income may result from a diversity of sources. Government welfare is an important income source to the very poor. Dependence on welfare is also an indicator of the intensity of poverty. Since previous studies found different effects for market income versus welfare payment, the present study creates a variable to represent changes in the dependence on welfare. Based on the questions about main source of household income, this variable is coded as: Persistent Dependence B welfare payments are the main source of income at both cycles; Increased Dependence B welfare payments become the main source of income in cycle 2; Decreased Dependence B welfare payments are no longer the main source of income at cycle 2; and No Dependence- welfare payments are not the main income source at both cycles.

Changes in employment status of parents lead to increases or decreases of family income. Possible changes will be reflected by a four category variable coded as: Move Into Employment B one or both parents move from unemployment or not in labour force at cycle 1 into employment at cycle 2; Move Into Unemployment B one or both parents become unemployed; Stable Employment; and No parent working at both cycles.

Family formation or breakdown is a major factor contributing to the movement of families with children into or out of a low income situation (Statistics Canada 1998). In the present study, we created a four-category variable to capture the change in family structure: Family Formation -- single, divorced, or widowed parents at cycle 1 become married or cohabited at cycle 2; Family Breakdown B married or common-law parents at cycle 1 become separated, divorced or widowed at cycle 2; Both Parents at Both Cycles; and Single Parents at Both Cycles. Due to small sample size for immigrant children in the age group of 4 to 9 years old, we combined the categories of Family Formation and Both Parents at Both Cycles in the analyses for immigrant children.

The study included three parental factors: ineffective parenting, parental depression, and family function. Ineffective parenting is a scale based on seven questions about how parents interact with the child. The measure of parental depression draws upon the CES-D, a widely-used survey instrument originally developed at the US National Institute of Mental Health (Radloff, 1977). To ease respondent burden, the NLSCY employed a 12 question, abbreviated version of the CES-D.

Family functioning was measured using the Family Assessment Device, a 12 item questionnaire developed by the Chedoke-McMaster Hospital group (Epstein, Baldwin and Bishop, 1983). The scale assesses communication, problem solving, affective responsiveness, affective involvement, and family roles. In the analysis, this variable was coded in a way that a higher score reflects a higher level of family dysfunction.

Control variables included the age and education of the PMK, as well as age and gender of the child. We also created a dummy variable, nonwhite (1=nonwhite, 0=white), as a crude measure to control for racial/ethnic heterogeneity among immigrant families. This variable was not included in the multivariate regression models for non-immigrant children, since this group contained only about 3.2% of non-white children. Among immigrant children, 41.1% were non-white. For immigrant children, we also included country of birth (foreign-born=1, Canadian born=0) and parent=s length of immigration as control variables.

The present study first described movements into and out of poverty, and changes in welfare dependence and employment status. This study further examined the bivariate associations of changes in poverty status with children=s developmental behaviours.

Multivariate analyses used the conditional change or static-score model (Finkel 1995, Kessler and Greenberg 1981). This model includes the prior level of the dependent variable to take into account the possible negative correlation between initial scores on a variable and its subsequent change. The prior level of the dependent variable in the model can also serve to control, at least partially, for omitted variables that influence the dependent variable (Finkel 1995).

For each outcome, the multivariate model includes initial score of the developmental behaviour, and all the selected control variables. It also contained following groups of variables representing changes in family economic situation and parental characteristics: (1) three dummy variables constructed from the four-category variable for changes in poverty status, with the non-poor at both cycles as the reference group; (2) the conditional interaction terms between absolute income change with each of the three dummy variables representing changes in poverty status; (3) three dummy variables derived from the four categories of changes in welfare dependence, with no dependence at both cycles as the reference; (4) three dummy variables derived from changes in employment status, with stable employment as reference; (5) dummy variables based on changes in family structure, using two parents at both cycles as the reference; and (6) changes in ineffective, parental depression and family dysfunction.

1. Economic Changes in the Family

During the period from 1994/95 to 1996/97, both non-immigrant and immigrant families improved, on average, their economic status, a trend consistent with the recovering national economy (Table 1). However, immigrant families slightly lagged behind non-immigrant families in moving out of poverty. About 14.5% of immigrant families were poor at both cycles of the survey, compared with 10.1% of non-immigrant families. The differences between immigrant and non-immigrant families were statistically significant (p<.000).

While the proportion of non-immigrant families depending on social welfare as the major source of family income decreased during this period, more immigrant families became dependent on social welfare. For about 7.8% of immigrant families, social welfare was the major income source in 1994/95, compared with 9.5% two years later. The corresponding percentage decreased from 8.8% to 7.0% for non-immigrant families. Again, these differences were statistically significant (p<.000).

More people, regardless of immigrant status, entered into the labour market in a growing economy. However, the improvement in employment rate was stronger among non-immigrant families. Furthermore, the labour market was more volatile for immigrants than for non-immigrants. Although adults in immigrant families were less likely to have no jobs at both cycles, they also had a smaller proportion of stable employment and a higher proportion of new unemployment than non-immigrant adults.

Changes in family poverty status were closely related to changes in welfare dependence. The relation was stronger among non-immigrant families (Cramer=s V (CV)= .44, contingency coefficient (C)=.60, p<.000) than among immigrant families (CV= .34; C=.51; p<.000). When both were persistently poor, non-immigrant families were far more likely to depend on welfare than immigrant families. Among persistently poor immigrant families, 20.6% persistently depended on social welfare, 11.3% decreased their welfare dependence, 21.1% increased their welfare, and 46.9% never depended on welfare in the two-year period. The corresponding percentages were 47.3%, 13.0%, 10.2%, 29.5%, respectively for persistently poor non-immigrant families.

Among those who recently became poor, non-immigrant families were also more likely to persistently depend on welfare or become dependent on welfare than immigrant families. Among newly poor immigrant families, 1.3% persistently depended on social welfare, 11.7% decreased their welfare dependence, 16.9% increased their welfare, and 70.1% never depended on welfare. The corresponding percentages for newly poor non-immigrant families were 11.7%, 6.6%, 18.7%, and 63.0%, respectively. Thus, the larger proportion of low-income among immigrant families was the primary reason contributing to their slightly higher overall dependence on social welfare than non-immigrant families. Within the same income level, immigrant families were less likely to depend on welfare than non-immigrant families.

Changes in poverty status were also related to changes in employment. Again, this association was stronger among non-immigrant families (CV = .38, C= .55, p<.000) than among immigrant families (CV=.30, C=.46, p<.000). Among the persistently poor, 49.1% of non-immigrant families had no parent working in the two-year period, compared with 25.7% of immigrant families. Furthermore, moving out of poverty was more strongly associated with new employment of parents among non-immigrant families than among immigrant families.

2. Changes in Poverty Status and Developmental Outcomes

Table 2 first provides some background information about differences in developmental outcomes between immigrant and non-immigrant children aged 4 to 9. At cycle 1, immigrant children had significantly lower levels of conduct problems. At cycle 2, the advantages of immigrant children over non-immigrant children extended to both outcomes.

As Table 2 also shows, at the bivariate level, changes in poverty status had no consistent associations with changes in emotional and conduct problems among both immigrant and non-immigrant children. In particular, children living in persistently poor families or in families recently falling into poverty did not necessarily experience deterioration in developmental outcomes. Children in families recently moving out of poverty also did not necessarily experience improvement in developmental outcomes.

Children's developmental behaviours at both cycle 1 and cycle 2 by changes in poverty status revealed a different pattern. Non-immigrant children in persistently poor families clearly had disadvantages in both emotional and conduct problems, and their differences with children in persistently non-poor families were significant. Thus, the disadvantages of non-immigrant children living in persistent poverty sustained over time, although prolonged exposure to poverty may not further escalate children=s developmental problems. In comparison, immigrant children living in persistently poor families did not have any significant disadvantages in developmental outcomes at both cycles of the survey.

The above bivariate analyses could not indicate whether the effects of changes in poverty status are conditioned by absolute income changes and influenced by changes in income sources, family structure, and parental characteristics. Tables 3 presents results of multivariate analyses for both developmental outcomes. Models 1 and 2 are for non-immigrant children, while Models 3 and 4 are for immigrant children.

As shown in Model 1 for emotional problems in Table 3, among non-immigrant families, there was a positive and significant interaction between persistent poverty and absolute increases in income. This positive interaction is rather counterintuitive: absolute income increases elevated children's emotional problems in persistently poor families. Figure 1 illustrates that among persistently poor families absolute income increases heightened children's emotional problems. As explained later in the discussion section, this counterintuitive result may entail some important underlying dynamics.

Results in Model 1 also indicated that children living in non-immigrant families that became recently dependent on social welfare as the major income source tended to have escalated emotional problems. Furthermore, single-parent status, regardless of the recency or duration of this status, was associated with elevated levels of children=s emotional problems. Increases in ineffective parenting and parental depression elevated children=s emotional problems.

Model 2 for conduct problems among non-immigrant children yield several results similar to those for emotional problems. First, increases in ineffective parenting and parental depression tended to aggravate children=s conduct problems. Second, children living in families which became increasingly dependent on social welfare had higher levels of conduct problems. Third, among persistently poor families, absolute income increases heightened children=s conduct problems.

Model 2 also revealed some patterns that were not observed in the model for emotional problems. First, among families recently moved out of poverty, absolute income increases helped to ameliorate children=s conduct problems. As Figure 2 illustrated, newly non-poor families on average showed about 91% increases in absolute income. Only among those families whose absolute income almost doubled in the two-year period, did children have lower levels of conduct problems than those in persistently non-poor families. This result suggests that transitions out of poverty without substantial increases in income would not be enough to reduce children=s conduct problems. Second, changes in absolute income also had a conditional effect among families that fell into poverty. The larger the decrease in absolute income, the larger the difference in children=s conduct problems between newly poor and persistently non-poor families. Third, children with parents who were unemployed at both cycles had higher levels of conduct problems. Fourth, breakdowns of two-parent families were associated with reduced levels of children=s conduct problems.

The associations between changes in children=s developmental outcomes and changes in family economic situation were rather different for immigrant families than for non-immigrant families. As shown in Model 3, after controlling for changes in family structure and parental characteristics, immigrant children living in persistently poor families had lower levels of emotional problems than those living in persistently non-poor families. On the other hand, increases in welfare dependence elevated immigrant children=s emotional problems. Unemployment of parents, whether it was persistent or recent, was associated with lower levels of children=s emotional problems. Changes from living with two parents to living with single parents tended to escalate immigrant children=s emotional problems. Increases in ineffective parenting and parental depression tended to aggravate immigrant children=s emotional problems. Although country of birth and race made no difference, immigrant children=s emotional problems increased as parents stayed longer in Canada after immigration.

Results in Model 4 revealed that immigrant children in newly poor families tended to have lower conduct problems than those in persistently non-poor families. Furthermore, absolute income decreases among persistently poor and newly poor families elevated children=s conduct problems. Similar to the results for emotional problems, immigrant children with newly unemployed parents had lower levels of conduct problems than children with persistently employed parents. Decreased welfare dependence also increased children's conduct problems.

The results in Model 4 also revealed that changes in family structure had no significant effects on immigrant children's conduct problems. Increases in ineffective parenting elevated children's conduct problems. Foreign-born children tended to have higher levels of conduct problems than children born to immigrant parents, while parents= length of residence and race made no significant difference.

Although economic changes in the family had rather weak and inconsistent overall effects on children=s developmental outcomes, the context in which economic changes manifest or do not manifest their effects deserves detailed discussions.

The finding that children=s developmental problems did not deteriorate among persistently poor families seems encouraging. However, among non-immigrant children in the 4 to 9 years old age group, those living in persistently poor families had sustained disadvantages in both emotional problems and conduct problems.

A salient finding of present study was the conditional effect of absolute income change among families exhibiting recent movements out of poverty. Simply moving out of poverty may not be sufficient to improve conduct problems for non-immigrant children aged 4 to 9 years. A substantial improvement in living standard, in terms of approximately doubling the absolute income, is needed in order to observe a significant impact of changes in poverty status. This finding highlights the importance of sustained and substantial improvement in family income. Thus, government programs for job training and placement should incorporate support for long-term education and skills upgrading.

A different conditional effect of absolute income was found among persistently poor non-immigrant families in which increases in absolute income intensified children's emotional problems and conduct problems. Detailed analyses suggest a tentative explanation: small economic gains might not compensate the associated psychological harm to parents and children. While some of persistently poor families had economic gains, the increases were typically small, and kept the family within the poverty zone. Furthermore, parents in persistently poor families who showed increases in absolute income were more likely to move into employment and consequently receive less or even be disqualified for welfare payments.

Thus, while the transition from welfare dependence to employment resulted only in small gains in absolute income that were not enough to pull the family out of poverty, parents might experience elevated stress from working at low-paying jobs. In addition, these working parents might spend less time with their children than they did previously. Therefore, measures to help poor families reduce welfare dependence and move into employment must be supplemented by comprehensive child-focussed supports. (Cauthen and Knitzer, 1999; Gomby and Larner 1995).

The scenario for immigrant children is different: immigrant children in persistently poor families seem to benefit from small absolute income increases. Absolute income increases tended to reduce conduct problems for immigrant children living in persistently poor families. The difference in this conditional effect between immigrant and non-immigrant children may be related to differential associations between changes in poverty status and other economic changes. Our descriptive analyses indicated that persistently poor immigrant families were far less likely to depend on welfare than their non-immigrant counterparts. In addition, among persistently poor families, immigrant parents were much more likely to be employed during the two-year study period than non-immigrant parents. Thus, gains in absolute income among persistently poor immigrant families were less likely to be associated with changes in income sources.

Furthermore, immigrant families in persistent poverty tended to have lower average income than their non-immigrant counterparts, probably because they were not protected by the social welfare to the same extent as non-immigrant families were. Thus, it is possible that poor immigrant families fell deeper in poverty and their children suffered a larger degree of material deprivation. Consequently, poor immigrant children could benefit more directly from small increases in absolute income.

When other economic and parental factors were controlled, persistent dependence on welfare had no significant detrimental impact on children=s developmental outcomes. Increases in welfare dependence tended to increase immigrant and non-immigrant children=s emotional problems, and increase non-immigrant children=s conduct problems.

Persistent unemployment of parents tended to increase non-immigrant children=s developmental problems. For immigrant children, however, parents= unemployment did not seem destructive to children=s developmental outcomes. For instance, when parents became unemployed, immigrant children=s emotional problems and conduct problems reduced. If both parents were without jobs at both cycles of the survey, immigrant children tended to have lower levels of emotional problems than those whose parents hold persistent employment. It is possible that parental attention is much more indispensable for immigrant children since immigrant families tended to have smaller social networks to support the care of their children, and smaller economic capacity to access other child care services.

Government policies should incorporate the special needs of immigrant families. Our results found that immigrant families were more likely to experience poverty but poor immigrant families were less likely to access, and derive benefits from the social welfare system than non-immigrant families. Probably because poor immigrant children suffered a higher extent of material deprivation, they responded to increases in family income more directly and positively than poor non-immigrant children in terms of the improvement in developmental behaviours. This finding highlights the urgent needs to reduce the depth of poverty among new immigrant families with young children. Although many new immigrant parents may not meet current regulations to receive social welfare, their children should not suffer as a consequence. Immigrant children are the fastest growing component of Canada=s child population. Their successful adjustment and development will have profound impacts on Canadian society.

Our study results also indicated that between 1994/95 and 1996/97, immigrant families did not enjoy the same degree of economic improvement, and faced a more volatile job market, even though they were more likely to participate in the labour force than non-immigrant families. This finding may be not applicable to all immigrant families since the survey only covered those with young children, and thus were likely to be new immigrant families. Therefore, the finding may reflect the initial resettlement difficulties that often confront new immigrants, and again calls for more comprehensive resettlement policies to help new immigrants fully realize their potentials.

Aber, J.L., N.G. Bennett, D.C. Conley, J. Li. 1997. The effect of poverty on child health and development. Annual Review of Public Health 18:463-483.

Anderson, E.R. Greene, S.M., Hetherington, M.E., and Clingempeel, W.G. 1999. The dynamics of parental remarriage: Adolescent, parent, and sibling influences. In M.E. Hetherington, et al. Coping with divorce, single parenting, and remarriage: A risk and resiliency perspective, pp. 295-319. Mahwah, NJ, USA: Lawrence Erlbaum Associates, Inc.

Basavarajappa, K. G., R.P. Beaujot and T.J. Samuel. 1993. Impact of Migration in the Receiving Countries - Canada. Geneva: International Organization for Migration.

Beaujot, R. 1991. Population Change in Canada: The Challenges of Policy Adaptation. Toronto: McClelland and Stewart Inc.

Beiser, M., Devins, G., Dion, R., Hyman, I., and Lin, E. 1997. Immigration, Acculturation and Health. Report to the National Health Research and Development Program (NHRDP), Project No. 6606-6414-NPHS, Ottawa.

Beiser, M., F. Hou, I. Hyman, and M. Tousignant, 1999. Poverty, family processes and mental health of new immigrant children. Paper submitted for publication. Clarke Institute of Psychiatry, University of Toronto.

Beiser M., Dion R., Gotowiec A., Hyman I., and Vu N. 1995. Immigrant and refugee children in Canada. Canadian Journal of Psychiatry, 40: 67-72.

Beiser, M., P. J. Johnson and R. J. Turner. 1993. Unemployment, underemployment and depressive effect among Southeast Asian refugees. Psychological Medicine 23: 731-743.

Blau, F. 1999. The effect of income on child development. (forthcoming) Review of Economics and Statistics.

Brooks-Gunn, J., 1995. Strategies for altering the outcomes of poor children and their families. In P.L. Chase-Lansdale and J. Brooks-Gunn (Eds.) Escape From Poverty: What Makes a Difference for Children, pp 87-117. New York: Cambridge University Press.

Chang L., Morrisey R.F., and Koplewicz H.S. 1995. Prevalence of psychiatric symptoms and their relation to adjustment among Chinese-American youth. Journal of the American Academy of Child and Adolescent Psychiatry, 34:91-99.

DeVoretz, D. 1995 Diminishing Returns: The Economics of Immigration Policy. Toronto, University of Toronto Press.

Elder,G. H., Nguyen, T. V., and Caspi, A. 1985. Linking Family Hardship to Children=s Lives. Child Development 56:361-75.

Epstein, N. B., L. M. Baldwin, D.S. Bishop. 1983. The McMaster Family Assessment Device. Journal of Marital and Family Therapy, 9: 171-180.

Finkel, S. 1995. Causal Analysis With Panel Data (Sage University Paper series on Quantitative Applications in the Social Sciences, 07-105). Thousand Oaks, CA: Sage.

Guo, G. 1998. The timing of the influences of cumulative poverty on children=s cognitive ability and achievement. Social Forces, 77: 257-88.

Hetherington, M.E. and Stanley-Hagan, M. 1999. The adjustment of children with divorced parents: A risk and resiliency perspective. Journal of Child Psychology and Psychiatry and Allied Disciplines, 40(1):129-140.

Human Resources Development Canada. 1996. Special Surveys: National Longitudinal Survey of Children and Youth, User=s Handbook and Microdata Guide.

Huston, A. C. 1991. Children in Poverty: Development and Policy Issues. In A.C. Huston (Ed.) Children in Poverty: Child Development and Public Policy. Pp 1-22. Cambridge: Cambridge University Press.

Kessler, R. C. and D. F. Greenberg. 1981. Linear Panel Analysis: Models of Quantitative Change. New York: Academic Press.

Klimidis S., Stuart G., Minas I.H. 1994. Immigrant status and gender effects on psychopathology and self-concept in adolescents: A test of the migration-morbidity hypothesis. Comparative Psychiatry, 35:393-404.

Levebvre, P. and Merrigan, P. 1998. Family background, family income, maternal work and child development. The Working Paper Series, Applied Research Branch, Human Resources Development Canada.

Lipman, E. L. and Offord, D. R. 1995. Psychosocial morbidity among poor children in Ontario. In G. Duncan, J. Brooks-Gunn (Eds.), Consequences of Growing Up Poor. Russell Sage Foundation.

Mayer, S. E. 1997. What Money Can=t Buy: Family Income and Children=s Life Chances. Cambridge, MA: Harvard University Press.

McLeod, J. D. and M. J. Shanahan. 1996. Trajectories of poverty and children=s mental health. Journal of Health and Social Behavior, 37: 207-220.

McLoyd, V. C. 1989. Socialization and Development in a Changing Economy: The Effect of Parental Job and Income Loss On Children. American Psychologist 44:293-302.

National Council of Welfare 1998. Poverty Profile 1996. Minister of Public Works and Government Services Canada. Catalogue # H67-1/4-1996E.

Picot, G., M. Zyblock, and W. Pyper. 1999. Why do children move into and out of low income: changing labour market conditions or marriage and divorce? Working paper 11F0019MPE No. 132, Statistics Canada.

Radloff, L, 1977. The CES-D Scale: a self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 388-401.

Statistics Canada. 1998. National Longitudinal Survey of Children and Youth: Cycle 2, 1996. The Daily. October 28.

Strayhorn, J.M. and Weidman, C. S. (1988). A parent practices scale and its relation to parent and child mental health. Journal of the American Academy of

Child and Adolescent Psychiatry, 27, 613-618.

Zhou, M. 1997. Growing Up American: The challenge confronting immigrant children and children of immigrants. Annual Review of Sociology, 23:63-95.

Note: this chart presents emotional disorder scores corresponding to absolute income changes calculated as percentage. The five percent points on the x-axis represent the mean absolute income change (middle point), 1 and 2 standard deviations below the mean change (from middle to left), and 1 and 2 standard deviations above the mean change (from middle to right). These percent points were calculated from the original variable which is the logarithm of the ratio of cycle 2 to cycle 1 average household income. The five points from the original variable were evenly spaced on the x-axis, although the transformed percent changes were not. Among persistently poor families the mean of the logarithmic absolute income change is -.011. The corresponding ratio of cycle 2 to cycle 1 average income is .99 which represents a 1% decrease between cycle 1 and cycle 2. Thus, on average absolute income for persistently poor non-immigrant families remained virtually unchanged. One standard deviation (.327) above the mean is .316, corresponding to a ratio of 1.37 which represents a 37% increase in absolute income. Two standard deviations above the mean is .643, corresponding to a ratio of 1.90 which represents a 90% increase in absolute income.

Table 1 Economic changes in immigrant and non-immigrant families

Table 2 Changes in Poverty Status and Developmental Behaviours

Note: (1) numbers in parentheses are standard deviations. (2) * difference between 1994/95 and 1996/97 is significant at p<.05, ** p<.01.