Poverty and Early Child Adjustment


Elizabeth B. Owens

University of California at Berkeley





Daniel S. Shaw

University of Pittsburgh





In Positive adaptation in children at risk (pp. 267-292), S. Luthar (Ed.).  Boston, MA: Cambridge University Press.

Childhood poverty appears to be an enduring and entrenched problem, resistant to most social and economic policies intended to lift families above the poverty line.  Although rates of  poverty among families of preschool-aged children initially declined during the 1960s when anti-poverty programs directed at children and families were initiated, rates rose throughout the 1970s and 80s, and leveled off in the 90s, with the consequence that young children continue to experience poverty at alarmingly high rates.  In point of fact, in 1999 about one in five infants and preschool-aged children in the U. S. lived in families whose incomes fell below the poverty threshold (U. S. Census Bureau, 2000a).  Poverty is considered a pervasive and nonspecific stressor, rather than a bounded one, because it negatively affects many aspects of individual and family functioning; yet at the same time, many impoverished children are positively adjusted (Luthar, 1999; Garmezy, 1991; Werner & Smith, 1992).  How is it that some children are vulnerable to the effects of poverty while others demonstrate positive adjustment (i. e., resilience)?  Attempts to answer this question are at the core of this chapter.  Our primary objective is two-fold:  to summarize findings from relevant literatures regarding factors associated with better or worse adjustment among young impoverished children, and to showcase one effort toward the identification of such factors using data from the Pitt Mother and Child Project, a longitudinal study of adjustment and psychopathology among young boys from low-income families.  In addition to identifying predictive factors, we will consider whether they are associated with outcomes among all children in the sample or just those at highest risk. We will also test the limits of resilient adjustment across time, and will discuss implications our findings have for both basic developmental science and early intervention programs.

Poverty as a risk factor for early childhood development

            To be established as a risk factor, a variable must temporally precede the negative outcome with which it is associated and must have a meaningful association with that outcome (Kramer et al., 1997).  Early childhood poverty, by which we mean living in a family prior to the age of 5 whose income falls below a specified level necessary for minimum coverage of basic expenses (usually the federal poverty thresholds set by U. S. Census Bureau), clearly meets these criteria.  Numerous researchers throughout recent decades have documented the negative child outcomes associated with living under the poverty threshold (see Brooks-Gunn & Duncan, 1997 for review), especially when poverty is experienced during the first 5 years of life (Duncan, Yeung, Brooks-Gunn, & Smith, 1998).  The detrimental outcomes are many, but can be grouped into three categories:  poor physical health (e.g. Jason & Jarvis, 1997; Pollitt, 1994), lower intellectual attainment and poor school performance (e.g. Dubow & Ippolito, 1994; Guo, 1998), and increased likelihood of social, emotional, and behavioral problems (e.g. Dubow & Ippolito, 1994; Duncan, Brooks-Gunn, & Klebanov, 1994).  In terms of effect sizes, Brooks-Gunn and Duncan (1997) document the following odds ratios for these outcomes associated with childhood poverty compared to not living in poverty:  in the health domain, the risk for low birth weight is 1.7, for child mortality 1.7, and for lead poisoning is 3.5; in the achievement domain, the risk for grade retention is 2.0; in the social/emotional/behavior domain, the risk for parent-reported behavior problems is 1.3. 

Of note too is that children who live in extreme or enduring poverty have the worst outcomes (Duncan et al., 1994; Guo, 1998; Korenman, Miller, & Sjaastad, 1995).  This is an important finding as it suggests that although income poverty is typically operationalized as a static, dichotomous construct, there may be linear effects of the level of family income or duration of impoverishment.  Explaining variations in child adjustment may have more to do with the severity or chronicity of poverty than with its presence or absence.  

Conceptualizing positive adjustment among young impoverished children

Developmentalists have adopted various methods of operationalizing child positive adjustment, one of which involves the creation of indices reflecting the achievement of stage-salient developmental tasks.  For example, Farber and Egeland (1987), in their study of resilient outcomes among maltreated young children, developed composite measures of positive adjustment or resilience at different ages that depended on achievement of developmentally appropriate tasks.  At 18 months, secure attachment was considered resilient; at 24 months, autonomy and problem solving were; at 42 months measures of self-regulation, success with peers, and levels of self-awareness contributed to their measure of resilience.  The appeal of this approach is its developmental sensitivity and inclusion of the concept of fit.  Adjustment is conceptualized in different ways across development that are defined according to both the changing capacities of the child and changing environmental demands and expectations.  

Operationalizing positive adjustment among impoverished children could also be guided by consideration of those domains in which they are at greatest risk for poor outcome.  This approach suggests that assessing physical health, intellectual attainment and school performance, and/or social/emotional/behavioral adjustment (the three primary domains of impairment for impoverished children) would be the best way to determine successful adjustment among these children. Children experiencing poverty might be considered positively adjusted if they are physically healthy, demonstrate the achievement of age-appropriate cognitive skills or are receiving good grades at school, or do not display social, emotional, or behavioral problems.

In assessing adjustment, some investigators have considered outcome domains individually, whereas others have considered them collectively.  In their study of maltreated children, Cicchetti, Rogosch, Lynch, and Holt (1993) measured seven indicators of competent adjustment in the following domains:  aspects of interpersonal behavior with peers, indicators of psychopathology, and school difficulties.  Children were defined as resilient when four or more of the indicators were present.  Similarly, Radke-Yarrow and Sherman (1990) provide an example of this cross-domain approach to defining positive adjustment.  In their study of children at-risk due to serious psychopathology in parents, they chose a group of "survivors" (those functioning adequately) who had no psychiatric diagnoses, were performing at grade level, related well to peers and to adults at school and home, and had a positive self-concept.

We have conceptualized positive adjustment (also called resilience) as a dynamic process that is reflected in a measurable outcome and indicates competent functioning despite the experience of significant adversity (Luthar, Cicchetti, & Becker, 2000).  Our focus is on social and behavioral adjustment.  In our own research, impoverished children were considered positively adjusted if they displayed levels of total behavior problems at age 8 within the normal range (within .5 standard deviation above the sample median), and displayed levels of social skills greater than the sample median.  In this way, positive adjustment was considered average or better adaptation in the domains of both psychopathology and social competence, but not necessarily exceptional functioning.  Pragmatically too, although our sample is large by some standards, we felt that it might be difficult to find many at-risk children with exceptional outcomes and thus identified children as positively adjusted if they demonstrated at least average functioning at a particular time point. 

Factors associated with positive adjustment

We turn next to a brief review of factors potentially associated with positive social and behavioral adjustment among young children in poverty.  As there have been relatively few studies specifically focused on adjustment patterns of this group, we also draw upon developmental theory and findings regarding adjustment of both preschool-aged and school-aged children at risk due to exposure to other chronic psychosocial stressors.

Within the resilience framework, factors associated with positive adjustment among at-risk children can be organized according to child, family, and community domains (Luthar & Zigler, 1991; Masten, Best, & Garmezy, 1990).  In this chapter we will consider factors in the child and family domains.  We do not mean to de-emphasize the importance of the community domain, by which we mean extrafamilial factors such as peer relationships, relationships with adults outside of the family, and community resources.  Theoretically, a number of these factors might be related to positive adjustment in young at-risk children.  We chose to focus only on child and family factors because we believed that the child's own characteristics and characteristics of the parents and family environment would likely be more influential than extrafamilial factors during the early childhood period. 

Child.  Among the potentially most important factors affecting the adjustment of all children is the quality of attachment to their primary caregivers. 1 Bowlby’s (1969) attachment theory suggests that attachment security should be related to later social and behavioral adjustment.  It is thought that primarily through its impact on a child's internal working model, attachment security influences thoughts and feelings about one's self in relationship with others, and about others in relationship with one's self (Bretherton & Munholland, 1999).  These thoughts and feelings are hypothesized to influence interpersonal behavior, including social competence and behavior problems that occur within relationships (e. g., noncompliance and aggression). 

Infant attachment security has, in fact, been shown to predict various developmental outcomes among young children from low-income families and families with high levels of psychosocial problems (see Greenberg, 1999 for review).  In particular, insecurely attached infants from such families tend to show higher levels of emotional and behavioral problems (particularly aggression) as children (Erickson, Sroufe, & Egeland, 1985; Lyons-Ruth, Alpern, & Repacholi, 1993; Renken, Egeland, Marvinney, Mangelsdorf, & Sroufe, 1989; Shaw, Owens, Vondra, Keenan, & Winslow, 1996).  In one large sample of low SES children, infant attachment security was predictive of psychopathology at age 10-11 (Urban, Carlson, Egeland, & Sroufe, 1991). Urban et al. (1991) also reported more dependence, less ego resilience, and less social competence at follow-up among their insecurely attached infants.

            A second potentially important predictive factor is child temperament.  It is possible that temperament and later adjustment are linked because an underlying biological process contributes to both.  It is also possible that temperamentally difficult children may elicit certain caregiver responses, such as inconsistent or harsh discipline, that then lead to later maladjustment (Bates, 1980).  As predicted, certain temperamental attributes reflecting whether a child is "difficult" or "easy-going" have been shown to predict adjustment among at-risk young children. Werner and Smith (1982) described their resilient children as agreeable, cheerful, friendly, relaxed, and sociable as toddlers.  Tschann, Kaiser, Chesney, Alkon, and Boyce (1996) found teacher-rated "easy" temperament to protect preschoolers in high-conflict families from developing teacher-rated externalizing problems.  Radke-Yarrow and Sherman (1990), in their sample exposed to parental psychopathology, described the resilient young children as socially engaging and charming.  Alertness, high activity, and curiosity have been found to predict resilient outcome in infants and young children exposed to multiple psychosocial stressors (Egeland & Sroufe, 1981; Farber & Egeland, 1987).  Similar findings are noted among school-aged children at risk, in which positive temperamental attributes have been shown to predict resilience (O'Keefe, 1994; Radke-Yarrow & Brown, 1993; Smith & Prior, 1995).

Intelligence is another attribute widely documented to protect against adversity (Luthar & Zigler, 1991; Masten & Coatsworth, 1998).  Intelligent children may have an easier time learning social skills or responding to certain parental socialization efforts, such as reasoning.  Perhaps IQ is associated with aspects of behavioral adjustment such as compliance because intelligent children may better understand, and therefore be able to follow, rules and procedures.  It is also possible that children with higher IQs are better able to use internal verbal mediation strategies in order to regulate negative emotions, or are better able to use verbal strategies in conflict situations in order to avoid resorting to aggressive or disruptive behavior.

Among at-risk preschoolers, developmental quotients and language test scores have been found to predict positive adjustment (Erickson et al., 1985; Pianta, Egeland, & Sroufe, 1990; Werner & Smith, 1982). A number of researchers have also found child IQ to be negatively associated with disruptive behavior and school failure in their at-risk samples (Masten et al., 1990; Smith & Prior, 1995).  Radke-Yarrow and Brown (1993) found high IQ to be associated with the lack of a psychiatric diagnosis.  However, childhood intelligence has not consistently been found to be associated with positive adjustment (Cicchetti et al., 1993; Egeland, Kalkoske, Gottesman, & Erickson, 1990).  Furthermore, among older children it has been argued that while intelligent but stressed youth may be relatively well-adjusted, they can lose their overall  advantage because of greater sensitivity to both negative forces in their lives (Luthar, 1999).

            Given this accumulation of theory and evidence, in our own work we have explored whether infant attachment security, level of negative emotionality, and child IQ might be predictive of positive adjustment among young impoverished children.

Family.  Developmental theories suggest diverse ways in which family factors might be predictive of adjustment.  For example, Patterson’s (1982) coercion theory carefully details patterns of reinforcement among certain negative and ineffective disciplinary strategies (e. g. giving into a child's disruptive behavior or hitting the child) and child noncompliance and aggression. Negative and rejecting parenting may also be a model for disruptive and angry child behavior, and/or could contribute to a child's low self-esteem, which may also produce behavioral maladjustment in children.  Rohner (1986) argues for a biologically-based need for positive interactions with caretakers, and that maladjustment is a response to rejection from caregivers.  There is evidence to support such posited associations between negative parental behaviors, including criticism, punitiveness, neglect, and rejection, and different dimensions of child maladjustment in low-SES samples (Dodge, Pettit, & Bates, 1994; Shaw et al., 1998) and impoverished samples (Conger, Ge, Elder, Lorenz, & Simons, 1994; McLoyd, Jayaratne, Ceballo, & Borquez, 1994; Sampson & Laub, 1994), although much of the work with impoverished samples has involved adolescents rather than young children.

Positive parental behaviors, particularly responsiveness and acceptance, may also be associated with child adjustment (Rohner, 1986; Shaw & Bell, 1993).  Such behaviors should have the opposite effect of negative and rejecting behavior in that responsiveness and acceptance may enhance the parent-child relationship and increase the effectiveness of socialization attempts.  Evidence shows that among at-risk children, positive parental behaviors and parent-child relationship quality are linked with social/behavioral adjustment.  Early maternal sensitivity and responsiveness (Bradley et al., 1994; Egeland & Sroufe, 1981; Erickson et al., 1985; Farber & Egeland, 1987; Shaw, Keenan, & Vondra, 1994), acceptance of child behavior (Bradley et al., 1994), and involvement (Erickson et al., 1985; Werner & Smith, 1982) have been associated with secure attachment and low levels of behavior problems in at-risk samples. Werner and Smith (1982) also reported that “emotional supportiveness” was associated with resilience in young at-risk children. Among school-aged children, Fisher, Kokes, Cole, Perkins, and Wynne (1987) found active, warm, and reciprocal parent-child relationships to predict measures of competence in boys with psychiatrically ill parents.  Mother-child warmth and relationship quality have been associated with a lack of behavior problems in children exposed to stressful life events and family violence (O’Keefe, 1994; Smith & Prior, 1995).  Masten et al. (1990) found interviewer ratings of parenting quality to be protective for girls (not boys) under stress, although “parenting quality” was not precisely defined.  A positive parent-child relationship and parental involvement have been found to be protective among at-risk children (Cowen, Wyman, Work, & Parker, 1990), and Stouthamer-Loeber et al. (1993) found a positive relationship with parents to predict nondelinquency.

Theoretically, parental personality traits and psychopathology should also be related to child adjustment.  It is possible that whatever genetic influences are partly responsible for parental personality or psychopathology are passed onto children who express those genes as behavior problems.  These influences may also be environmentally-mediated in that they have an impact on parental behavior and the parent-child relationship.  Patterson and Capaldi (1991) found mothers with aggressive personality styles to be less effective parents and to monitor their children less well.  Angry mothers might also, through modeling, teach their children to be aggressive.  Mothers who are hostile and/or depressed are more likely to use inconsistent, permissive, harsh and punitive, or coercive discipline (Gelfand & Teti, 1990; Kochanska, Kuczynski, & Maguire, 1989; Peterson, Ewigman, & Vandiver, 1994). Depressed parents may be more likely to disengage from the parenting role, to emotionally withdraw from their children, and to neglect their parenting responsibilities (Gelfand & Teti, 1990; Osofsky & Thompson, 2000).  Harnish, Dodge, and Valente (1995) found the association between maternal depressive symptomatology and child behavior problems for the Caucasians in their sample to be partially mediated by poor mother-child interaction quality (i. e., low enjoyment and sensitivity, high controllingness). 

            However, links between parental personality and dimensions of child adjustment have been somewhat inconsistent, at least among very young children.  Some find predictions in at-risk samples from negative personality traits including aggressivity and hostility (Renken et al., 1989; Shaw, Vondra, Hommerding, Keenan, & Dunn, 1994), but others do not (Egeland & Sroufe, 1981; Egeland et al., 1990), or report relations only for girls (Pianta et al., 1990). On the other hand, evidence clearly exists for relations between maternal psychopathology, especially depression, and child behavioral adjustment (Leadbeter & Bishop, 1994; Osofsky & Thompson, 2000; Pannacione & Wahler, 1986; Rose, Rose, & Feldman, 1989; Shaw & Vondra, 1995).

Furthermore, many have argued that family factors are the primary mechanisms by which poverty influences child social and behavioral development (see Child Development, 1994, vol. 65(2), special issue on Children and Poverty).  It is suggested that the strains of living in poverty negatively affect parental adjustment and behavior, for example, by increasing parental dysphoria and use of rejecting or inconsistent discipline (Brody et al., 1994; Conger et al., 1994; Dodge et al., 1994; McLoyd et al., 1994; Sampson & Laub, 1994), or by decreasing parent-child relationship quality and warmth (Dodge et al., 1994; Sampson & Laub, 1994), which may then contribute to child maladjustment.  

            Given this accumulation of theory and evidence, we tested whether maternal rejection, acceptance, and responsivity, mother-child relationship quality, maternal aggressive/hostile personality, and maternal depressive symptoms showed relations with social/behavioral adjustment among the impoverished children in our sample.

The Pitt Mother and Child Project

Participants and procedures.  The Pitt Mother & Child Project is an on-going study of developmental precursors of antisocial behavior among boys from low-income families in the Pittsburgh, PA metropolitan area.  Three hundred-ten infant boys and their mothers were recruited from Women, Infant, and Children (WIC) nutritional supplement clinics when the boys were approaching 1.5 years.  Sixty percent of the mothers are white and 40% are black.  Almost two-thirds of the mothers had 12 or fewer years of education, and about 60% were married or living with partner.  At 18 months, two-thirds of the families were living under poverty threshold, and by age 6, about half were.  Forty percent of the families were living under poverty threshold at all visits between 18 months and 5 years.

Boys and their mothers have been seen at 1.5, 2, 3.5, 5, 5.5, 6, and 8 years of age in the home and/or in the lab for 1 ˝ - 3 ˝ hours.  They participated in a series of developmentally appropriate unstructured and semi-structured interaction tasks, many of which were videotaped for later coding.  At each visit mothers also completed questionnaires and interviews regarding demographic variables, their own adjustment, family functioning, and child behavior.  Typically 280 to 300 of the families were seen at each of the visits (after 1.5 years) so that attrition at each age was usually less than 10%.


            Early childhood poverty.  Early childhood poverty was indexed using maternal report of monthly family income and the number of people living at home at each of the 1.5-, 2-, 3.5-, and 5-year-visits.  Poverty thresholds were established for households of different sizes in the different years in which data were collected (U. S. Census Bureau, 2000b), and it was determined whether each boy was living in a family that was above or below the poverty threshold at each of the four visits at or before age 5.  A total score ranging from 0 to 4 reflected the number of assessments in which the family was living under the poverty threshold.  Eighteen percent of families never lived in poverty, 11% families lived in poverty at one visit, 12% lived in poverty at two visits, 19% lived in poverty at three visits, and 40% lived under the poverty threshold at all four visits.

            Positive adjustment.  Positive adjustment was indexed using total scores from the Child Behavior Checklist (CBCL; Achenbach, 1991) completed by mother, and the Social Skills Rating System (SSRS; Gresham & Elliot, 1989), completed by teacher.  Both were administered when the child was approximately 8 years old.  Both instruments are widely used and possess satisfactory psychometric properties.  The total score from the CBCL is based on both internalizing and externalizing problems; the total score from the SSRS reflects cooperation, assertiveness, and self-control with peers and adults. 

            In our analyses, a dichotomous score reflecting the presence or absence of positive adjustment was created.  Children were considered positively adjusted if 1) their total behavior problem score from the CBCL was well-within the normal range (i. e., less than a T score of 55, which is within .5 standard deviation above both the normative and sample median of 50), and 2) their total social skills score was higher than average (i. e., higher than the sample median of 37).  Children who did not meet both of these criteria were not considered to be positively adjusted (although, they were not necessarily maladjusted either).  Of the 207 children with CBCL and SSRS scores, 83 (40%) were considered positively adjusted and 124 (60%) were not.

            Child factors tested for association with adjustment.  Infant attachment security was measured at 1.5 years using the Strange Situation procedure (Ainsworth & Wittig, 1969).  Children were classified as avoidant, secure, ambivalent, or disorganized by reliable, trained raters.  For analyses herein, children were categorized as secure or insecure (avoidant, ambivalent, or disorganized).  Child negative emotionality was measured at 1.5 years by reliable, trained raters who viewed videotapes of the hour-long lab visits and rated the frequency, duration, and intensity of child fussing and crying.  Child intelligence was measured at 5.5 years using prorated Verbal and Performance scores derived from a four-scale short form of the Wechsler Preschool and Primary Scale of Intelligence - Revised (Wechsler, 1989). 

Family factors tested for association with adjustment.  Maternal rejection was derived at 1.5, 2, and 3.5 years from videotaped observations of a clean-up task (Winslow, 1995), with scores averaged across age.  Maternal acceptance and responsivity were derived from the Acceptance and Emotional/Verbal Responsivity subscales of the HOME Inventory (Caldwell & Bradley, 1984), an observer-report measure of maternal behavior and quality of the home environment that was completed at the age 2 visit.  Parent-child relationship quality was measured using maternal-report on a parent adaptation of the Teacher-Child Relationship Scale (Pianta & Steinberg, 1991), administered at 5 and 6 years.  Scores from the Openness and Conflicted/Angry (negative) scales at both assessment points were averaged to create a score for mother-child relationship quality.  Maternal depressive symptoms were self-reported using the total score from the Beck Depression Inventory (BDI, Beck, Ward, Mendelson, Mock, & Erbaugh, 1961) administered at 1.5, 2, 3.5, and 5 years.  The average of these four scores was used in analyses.  Maternal aggressive/hostile personality was measured using the Aggression subscale of the Jackson Personality Research Form (Jackson, 1967) administered and 1.5 years, and the Overt Anger subscale from the Handling Anger Questionnaire (Spielberger et al., 1985), administered at 3.5 years.  Scores from these two measures were averaged to create a single score for maternal aggressive/hostile personality.


Analyses were conducted in four stages:  1) testing of associations between each child and family factor and positive adjustment among those children living under the poverty threshold at one or more visits between 1.5 and 5 years of age, 2) exploring conjoint effects of significant child and family factors identified at stage 1, 3) testing of main versus interactive effects of child and family factors, 4) and ascertaining cross-time limits to positive adjustment.

Stage One. Although all children in our sample were from low-income families, some were not impoverished.  Only those who were impoverished at one or more of the four visits between 1.5 and 5 years were considered at-risk (n = 235, 82% of original 310), and only those in this at-risk group with available CBCL and SSRS data at age 8 (n = 167, 54% of original 310) were used in the primary analyses.  Impoverished children with (n = 66, 40% of those analyzed) and without (n = 101, 60% of those analyzed) positive adjustment according to their CBCL and SSRS total scores were compared using a series of t-tests, and results are presented in Table 1.  Because these analyses were largely exploratory (few have investigated factors associated with adjustment among young impoverished children), the only procedure chosen to control Type I error rate was the use of two-tailed tests.  Of note too is that among the 101 without positive adjustment, 61 had either CBCL total or SSRS total scores that met criteria for adjustment, while 40 did not meet criteria in either domain.

Two child factors were associated with positive adjustment.  Impoverished children with secure attachments at 18 months were 2.5 times more likely to be positively adjusted at age 8 compared to those with insecure infant attachments (odds ratio = 2.48; 95% CI = 1.29 to 4.80).  Child Verbal IQ at age 5.5 was positively associated with positive adjustment and the size of the IQ effect in children with versus without positive adjustment was moderate (Cohen's d = .43).  Two family factors also showed significant associations with positive adjustment.  Mothers of impoverished children who were positively adjusted at age 8 were more likely to describe their relationships with their children as open and relatively free of conflict at ages 5 and 6.  These same mothers also self-reported lower aggressive/hostile personality scores when their children were age 1.5 to 3.5, compared to mothers of children who were not positively adjusted.  The size of each of these effects was moderate (Cohen's d = .51 and .48, respectively). 

Stage 2.  In order to test conjoint effects while controlling for collinearity among these four child and family factors, a logistic regression equation was computed in which these factors were entered stepwise, with presence or absence of positive adjustment employed as the dependent variable.  We did not enter all child and family predictors at this point in order to preserve power to detect significant effects and chose a stepwise entry procedure due to our interest in empirically identifying those child and family factors most strongly and uniquely related to outcome.  Maternal aggressive/hostile personality (Wald = 4.43, p = .04; odds ratio = .58, 95% CI = .35 to .96), infant attachment security (Wald = 7.36, p = .01; odds ratio = 2.88, 95% CI = 1.34 to 6.20), and mother-child relationship quality (Wald = 6.25, p = .01; odds ratio = 2.04, 95% CI = 1.17 to 3.59) each emerged as a statistically significant and unique predictor of child positive adjustment.  (Statistics presented here are from the final step once all significant variables were entered into the equation).  Once these factors were in the equation, the relation between child Verbal IQ and positive adjustment was not significant.  Therefore, with the exception of Verbal IQ, the factors individually associated with outcome continued to show unique and significant associations with positive adjustment even when considered conjointly.

Stage 3.  A key question involves whether associations between child and family factors and positive adjustment exist only or to a greater degree in at-risk children compared to children not at risk.  Statistically, the question is whether factors exert main effects (present regardless of level of risk) or interactive effects (present only or to a greater degree in children at-risk). While knowledge regarding factors associated with positive adjustment across all children is valuable, identifying factors that operate specifically under conditions of risk furthers our understanding of the concept of resilience (the process and outcome of positive adaptation despite adversity [Luthar et al., 2000]), and some argue that only under such conditions would a factor associated with good outcome be considered protective (Rutter, 1985).

Our sample is not ideal for testing such a question because of its restricted continuum of risk (i. e., all participants were low-income).  However, we conducted analyses comparing the two extremes of risk levels, that is, children who were not living under the poverty threshold at any visit to those who were living under the poverty threshold at every visit.  Although the proportion of positively adjusted children was not very different in these two groups (37% versus 43%, respectively), at least theoretically chronic poverty between the ages of 1.5 and 5 years should pose greater risks to adaptive development in the long term, as compared to low family income in the absence of poverty.  And despite the comparable rates of positive adjustment in these two groups, some evidence suggests that the chronically impoverished group was in fact worse off.  Within both groups of children (never impoverished and chronically impoverished) who were not positively adjusted, almost all (83% and 84%, respectively) were at or below the median on social skills.  However, while only 26% of these never impoverished children were also high on behavior problems, 53% of these chronically impoverished children showed elevated levels of behavior problems. 

            Hierarchical logistic regressions would typically be used to test main and interactive effects of putative predictive factors on the presence or absence of a positive adjustment.  In these, risk would be entered first and the predictor and its interaction with risk would then be entered either stepwise or hierarchically. However, we believed that this analytic strategy was not optimal for our purposes.  Entering risk on step one allows for tests of the effects of predictive factors once the effect of risk has been statistically controlled, but not for tests of predictor effects within varying risk contexts.  Additionally, given that statistical tests of interactions are generally low powered (Smith & Sechrest, 1991), interactions can be difficult to detect when independent variables are normally distributed (McClelland & Judd, 1993), and not all interaction effects are best represented by multiplicative interaction terms (Rutter, 1983), we instead split the sample according to high- or low-risk status.  Then, relations between predictors and positive adjustment were tested using t-tests and effect sizes within each group. 

Associations between child and family factors and the presence or absence of positive adjustment in the never impoverished and chronically impoverished groups are presented in Table 1.  As can be seen, maternal rejection, maternal depressive symptoms, and mother-child relationship quality were significantly associated with lower rates of positive adjustment only among the chronically impoverished children.  Furthermore, effect sizes were notably larger among the chronically impoverished children (.63, .59, and .85, respectively) compared to effect sizes among the not impoverished children (.00, .08, and .39, respectively).  Maternal aggressive/hostile personality was also significantly associated with lower rates of positive adjustment only among the chronically impoverished children, but the effect size difference between the theoretically most at-risk and low-risk groups was smaller (.78 versus .49).

Because of collinearity among predictors, relations between family factors and positive adjustment in the chronically impoverished group were probably not unique.  In this group, the largest correlation was .38 between maternal depressive symptoms and aggressive personality.  Absolute values of all other correlations among these four predictors were in the range of .20 to .30 (i.e., shared variance of 4% to 9%), with the exception of the .06 correlation between maternal rejection and depressive symptoms.  Given the relatively small number of subjects in the never impoverished and chronically impoverished groups, logistic regressions accounting for collinearity among variables by including all predictors were notably underpowered to detect significant effects and were not computed.  Therefore, our findings suggesting that less maternal rejection, depression, and better quality mother-child relationships were associated with positive adjustment primarily among those presumably at highest risk must be interpreted with the collinearity among these predictors in mind.

Stage 4.  In order to explore the cross-time stability of positive adjustment among our participants, children who were impoverished and positively adjusted at age 8 were selected (n = 66).  Forty of these children had CBCL and SSRS ratings available at age 6.  Of these 40, 25 met criteria for positive adjustment at age 6.  In other words, 62% of the impoverished children who showed positive adjustment at age 8 and had data available at age 6 showed positive adjustment at age 6.  Furthermore, of the entire 112 impoverished children who had CBCL and SSRS data available at ages 6 and 8, 73% showed stability of adjustment across time (positively adjusted at both times or not positively adjusted at both times), while 27% were positively adjusted at one time and not the other. Although the amount of missing data prevents drawing firm conclusions, these descriptive statistics suggest stability of adjustment ratings among a majority of children in our sample. 

We also investigated which child and family factors were associated with cross-time stability versus instability of positive adjustment in the 112 impoverished children with CBCL and SSRS data at ages 6 and 8.  Using a logistic regression equation, we entered all ten child and family factors stepwise to see which best predicted positive adjustment at ages 6 and 8 (n = 25) versus positive adjustment at 6 or 8 (n = 30) among the impoverished children in our sample.  Only child Verbal IQ emerged as a statistically significant predictor of stable positive adjustment versus unstable positive adjustment (Wald = 5.98, p = .01; odds ratio = 1.09, 95% CI = 1.02 to 1.16).  Although given 10 putative predictors this test was underpowered to detect more than one significant effect, results of a series of 10 t-tests revealed similar results (only Verbal and Performance IQ were associated with the presence or absence of stable positive adjustment).

Summary and Conclusions

We aimed to identify factors associated with positive adjustment at age 8, defined by a lack of behavior problems and above average social skills, among the impoverished young children participating in the PMCP.  Our search was guided by developmental theory and empirical evidence regarding factors that might discriminate those who were considered positively adjusted from those who were not.  These factors included infant attachment security and temperament, child IQ, mothers’ rejection, depressive symptoms, aggressive/hostile personality, acceptance and responsivity, and the quality of the mother-child relationship.

Results indicated unique and moderately strong effects for secure attachment, a good mother-child relationship, and lack of maternal aggression/hostility.  Among children who lived under the poverty threshold at some point between the ages 1.5 and 5 years, those with secure attachments at 18 months were 2.5 times as likely as others to show positive adjustment 5.5 years later.  Similarly, the average mother-child relationship quality score at age 5 to 6 was twice as large, and the early-measured maternal aggressive/hostile personality score was half as large, for children who were positively adjusted at age 8 compared to those who were not.  Verbal IQ was also associated with later adjustment, but only when considered individually. 

These primary findings suggest predictability across stages of childhood development, from infancy to middle childhood.  Although predictive factors measured during the infant/preschool period may simply be associated with causal factors not identified, it is possible that certain early experiences have lasting effects.  Primary findings also implicate the importance of the child's close relationships and the affective or emotional aspects of the family environment in the prediction of positive adjustment. Exactly how close relationships and the affective nature of a child's immediate environment influence adjustment is not suggested by our findings, but it is possible that the impact is direct (e. g. modeling of negative affectivity resulting in child behavior problems and poor interpersonal skills), or indirect (e. g. maternal aggressive personality negatively impacting disciplinary behaviors, which then lead to behavior problems).

In order to determine whether a factor was associated with outcome regardless of risk or to a greater degree among children most at risk, we compared those who were low-income but never impoverished to those were who were impoverished throughout their first five years.  Maternal rejection, depressive symptoms, and mother-child relationship quality were significantly related to positive adjustment only among those most at risk, and each showed moderate to large effect size differences across the most and least at-risk groups.  This suggests that these family factors interact with risk status in the prediction of positive outcome, rather than exerting primarily main effects regardless of risk status.  However, findings must be interpreted with caution because of our inability to account for collinearity among predictors.

Such interactive effects may be interpreted as protection afforded by the presence of certain positive family factors (e. g., high quality mother-child relationship), or as increased vulnerability due to the presence of the negative family factors (e. g., low quality mother-child relationship).  Stouthamer-Loeber et al. (1993) drew attention to this issue when they identified a number of "double edged" variables that contained both protective and risk effects for delinquency.  A high quality mother-child relationship might be protective because in the context of frequent and intense stressors (often found in the environments of chronically impoverished children) there is a special need for the parent-child relationship to be strong.  However, it could also be argued that a high quality mother-child relationship or low levels of maternal depressive or aggressive symptoms should predict positive adjustment regardless of a child’s level of risk and therefore are not best described as protective.  Rather, such predictive factors should increase vulnerability by having a greater impact in an already risk-laden situation, as is suggested by a cumulative threshold model.  Up to a certain point, family factors appear less influential.  However, in the presence of chronic poverty certain family factors are more important for adjustment.  It also may be that chronically impoverished children have few ways to compensate for a poor relationship with mother or for the experience of maternal rejection (e. g., they may not have another parent at home or a caring teacher at school to whom to turn), which might explain the greater negative effect of certain family factors.

The stronger associations between mother and child adjustment among the chronically poor children may also have been due to the fact that mothers perhaps were globally maladjusted, i. e., when depressed, her own adjustment, her relationship with her child, and her ratings of that child's adjustment (which partially determined who was positively adjusted) all looked bleak.  Alternatively, perhaps high maternal maladjustment operated as a third variable, leading both to maternal inability to work and subsequent chronic poverty (Knitzer, Yoshikawa, Cauthen, & Aber, 2000), and to high levels of child maladjustment, with effects on children being either environmentally or genetically mediated.

Our findings indicated modest consistency of resilient status across domains and over time.  Of the at-risk children who showed age 8 positive adjustment in at least one of the two domains considered (behavior problems and social skills), about half were well adjusted in both whereas half did well in one area but not the other.  Almost two-thirds of the impoverished children who met criteria for positive adjustment at age 8 had also shown positive adjustment at age 6.  Family factors did not predict stable versus unstable positive adjustment, but child verbal intelligence did.  It is possible that because children leave the confines of home and enter school around age 6, their own coping abilities (as opposed to parental characteristics and behaviors) come to play a greater role in determining whether they tend to retain the patterns of adjustment shown previously.

            Without knowledge of causal mechanisms underlying associations between child and family factors and positive adjustment, the following implications for early intervention are tentatively suggested.  When one's goal is to foster positive adjustment among children in poverty, our findings suggest the need for concerted attention to promoting positive mother-child relationships.  This idea is supported from our data on impoverished infants with insecure attachments to their mothers, who were more than twice as likely to manifest subsequent maladjustment compared to those with secure attachments early in life, and our finding that mother-child relationship quality at ages 5 to 6 was associated with the presence of positive adjustment at age 8.  The promotion of positive mother-child relationships may be particularly important for those experiencing chronic poverty.  As noted above , perhaps because there are typically frequent and intense stressors in the environments of chronically impoverished children, there is a greater need for the parent-child relationship to be strong. The promotion of positive mother-child relationship quality could be achieved through parent-infant psychotherapy (Heinicke et al., 1999; Lieberman & Zeanah, 1999) or multisystemic family therapy (Henggeler, 1990), the latter addressing both proximal and more distal risk factors that compromise the quality of caregiving (Shaw, Bell, & Gilliom, 2000).  Aspects of maternal adjustment (aggressive/hostile personality and depressive symptoms) should also be targets of intervention, perhaps especially among chronically impoverished families.  Maternal adjustment might be directly treated with individual therapy or medication.  As a relevant example, Luthar and Suchman (1999) stressed the importance of using supportive psychotherapy to address maternal comorbid psychopathology and adjustment problems among low SES substance abusing mothers when the goal is to improve parenting behaviors and mother-child relationships.

            In summary, in our study of impoverished children we found that positive adjustment at age 8 was associated with secure attachment manifested at 18 months, low maternal aggressive personality at 1.5 to 3.5 years, and positive mother-child relationships at 5 to 6 years.  Comparisons between children who experienced chronic poverty and those who were never impoverished (but were nevertheless low-income) suggested that the presence or absence of age 8 positive adjustment among the former was more strongly linked with prior maternal rejection, maternal depressive symptoms, and mother-child relationship quality than among the latter.  Our findings regarding the limits of positive adjustment point to the dangers of assuming that resilience is either absolute across domains or fixed over time.  At the same time, our findings underscore the potential benefits of preventive early interventions that foster secure infant attachments and positive mother-child relationships, as well as interventions designed to address maternal adjustment problems, when one’s goal is to promote positive social and behavioral adjustment among impoverished children.





Achenbach, T. M.  (1991).  Manual for the Child Behavior Checklist/4-18 and 1991 profileBurlington,

            VT: University of Vermont, Department of Psychiatry.

Ainsworth, M. D. S., & Wittig, B. A.  (1969).  Attachment and exploratory behavior of one-year-olds in a

            Strange Situation.  In B. M. Foss (Ed.), Determinants of infant behavior:  Vol. 4 (pp. 113-136). 


Bates, J. E.  (1980).  The concept of difficult temperament.  Merrill-Palmer Quarterly, 26, 299-319.

Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. E., & Erbaugh, J. K.  (1961).  An inventory for

            measuring depression.  Archives of General Psychiatry, 4, 561-571.

Bowlby, J.  (1969).  Attachment and loss:  Vol. 1:  AttachmentNew York:  Basic Books.

Bradley, R. H., Whiteside, L., Mundfrom, D. J., Casey, P. H., Kelleher, K. J., & Pope, S. K.  (1994). 

Early indications of resilience and their relation to experiences in the home environments of low

birthweight, premature children living in poverty.  Child Development, 65(2), 346-360.

Bretherton, I. & Munholland, K. A.  (1999).  Internal working models in attachment relationships:  A

            construct revisited.  In J. Cassidy and P. R. Shaver (Eds.), Handbook of Attachment:  Theory,

            Research, and Clinical Applications (pp. 89 - 114).  New YorkGuilford.

Brody, G. H., Stoneman, Z., Flor, D., McCrary, C., Hastings, L., & Conyers, O.  (1994).  Financial

            resources, parent psychological functioning, parent co-caregiving, and early adolescent

            competence in rural two-parent African-American families.  Child Development, 65(2), 590-605.

Brooks-Gunn, J., & Duncan, G. J.  (1997).  The effects of poverty on children.  The Future of Children,

7(2), 55-71.

Caldwell, B. M., & Bradley, R. H.  (1984).  Manual for the Home Observation for Measurement of the

EnvironmentLittle RockUniversity of Arkansas.

Cicchetti, D., Rogosch, F. A., Lynch, M., & Holt, K. D.  (1993).  Resilience in maltreated children:

            Processes leading to adaptive outcome.  Development and Psychopathology, 5(4), 629-648.

Conger, R. D., Ge, X., Elder, G. H., Lorenz, F. O., & Simons, R. L.  (1994).  Economic stress, coercive

family process, and development problems of adolescents.  Child Development, 65(2), 541-561.

Cowen, E. L., Wyman, P. A., Work, W. C., & Parker, G. R.  (1990).  The Rochester Child Resilience Project:

Overview and summary of first year findings. Development and Psychopathology, 2, 193-212.

Dodge, K. A., Pettit, G. S., & Bates, J. E.  (1994).  Socialization mediators of the relation between

socioeconomic status and child conduct problems.  Child Development, 65(2), 649-665.

Dubow, E. F., & Ippolito, M. E.  (1994).  Effects of poverty and quality of the home environment on

changes in the academic and behavioral adjustment of elementary school-age children.  Journal of

Clinical Child Psychology, 23(4), 401-412.

Duncan, G. J., Brooks-Gunn, J., & Klebanov, P. K.  (1994).  Economic deprivation and early childhood

development.  Child Development, 65, 296-318.

Duncan, G. J., Yeung, W. J., Brooks-Gunn, J., & Smith, J. R.  (1998).  How much does childhood poverty

affect the life chances of children?  American Sociological Review, 63(3), 406-423.

Egeland, B., & Sroufe, L. A.  (1981). Attachment and early maltreatment.  Child Development, 52, 44-52.

Egeland, B., Kalkoske, M., Gottesman, N., & Erickson, M. F.  (1990).  Preschool behavior problems: 

Stability and factors accounting for change.  Journal of Child Psychology and Psychiatry, 31, 891-909.

Erickson, M.F., Sroufe, L. A., & Egeland, B. (1985).  The relationship between quality of attachment and

behavior problems in preschool in a high-risk sample.  In I. Bretherton & E. Waters (Eds.),

Growing points of attachment theory and research.  Monographs of the Society for Research in Child Development, 50(1-2), 147-167.

Farber, E. A., &  Egeland, B.  (1987).  Invulnerability among abused and neglected children.  In E. J.

Anthony & B. J. Cohler (Eds.), The invulnerable child (pp. 253-288).  New York: Guilford Press.

Fisher, L., Kokes, R. F., Cole, R. E., Perkins, P. M., & Wynne, L. C.  (1987).  Competent children at risk: 

A study of well-functioning offspring of disturbed parents.  In E. J.  Anthony & B. J. Cohler

(Eds.), The invulnerable child (pp. 253-289).  New YorkGuilford Press.

Garmezy, N.  (1991).  Resilience and vulnerability to adverse developmental outcomes associated with

poverty.  American Behavioral Scientist, 34(4), 416-430.

Gelfand, D. M. & Teti, D. M.  (1990).  The effects of maternal depression on children.  Clinical

Psychology Review, 10(3), 329-353.

Greenberg, M. T.  (1999).  Attachment and psychopathology in childhood.  In J. Cassidy and P. R. Shaver

            (Eds.), Handbook of Attachment:  Theory, Research, and Clinical Applications (pp. 469 - 496). 

            New YorkGuilford Press.

Gresham, F. M., & Elliot, S. N.  (1989).  Social Skills Rating System:  Parent, teacher, and child forms

Circle Pines, MN:  American Guidance Systems.

Guo, G.  (1998).  The timing of the influences of cumulative poverty on children's cognitive ability and

            achievement.  Social Forces, 77(1), 257-287.

Harnish, J. D., Dodge, K. A., & Valente, E.  (1995).  Mother-child interaction quality as a partial mediator

of the roles of depressive symptoms and socioeconomic status in the development of child

behavior problems.  Child Development, 66(3), 739-753.

Heinicke, C. M., Fineman, N. R., Ruth, G., Recchia, S. L., Guthrie, D., &  Rodning, C. (1999). 

            Relationship-based intervention with at-risk mothers:  Outcome in the first year of life.  Infant

            Mental Health Journal, 20(4), 349-374.

Henggeler, S. W. (1990).  Family therapy and beyond: A multisystemic approach to treating the behavior

            problems of children and adolescentsPacific Grove, CA:  Brooks/Cole.

Huston, A. C., Garcia Coll, C. T., & McLoyd, V.C.  (1994).  Special issue:  Children in poverty.  Child

            Development, 65(2).

Jackson, D. N. (1967). Personality Research Form manual. New York:  Research Psychologists Press.

Jason, J. M., & Jarvis, W. R.  (1987).  Infectious disease:  Preventable causes of infant mortality. 

            Pediatrics, 80, 335-341.

Knitzer, J., Yoshikawa, H., Cauthen, N. K., & Aber, J. L.  (2000).  Welfare reform, family support, and

            child development:  Perspectives from policy analysis and developmental psychopathology. 

            Development and Psychopathology, 12(4), 619-632.

Kochanska, G., Kuczynski, L., & Maguire, M.  (1989).  Impact of diagnosed depression and self-reported

            mood on mothers' control strategies:  A longitudinal study.  Journal of Abnormal Child

            Psychology, 17(5), 493-511.

Korenman, S., Miller, J. E.,  Sjaastad, J. E.  (1995).  Long-term poverty and child development in

            the United States: Results from the NLSY.  Children and Youth Services Review, 17, 127-155.

Kraemer, H. C., Kazdin, E., Offord, D. R., Kessler, R. C., Jensen, P. S., & Kupfer, D. J. (1997).

            Coming to terms with the terms of risk.  Archives of General Psychiatry, 54, 337-343.

Leadbeater, B. J., & Bishop, S. J.  (1994).  Predictors of behavior problems in preschool children of inner-

city Afro-American and Puerto Rican Adolescent Mothers.  Child Development, 65(2), 638-648.

Lieberman, A. F., & Zeanah, C. H.  (1999).  Contributions of attachment theory to infant-parent

            psychotherapy and other interventions with infants and young children. In J. Cassidy and P. R.

            Shaver (Eds.), Handbook of Attachment:  Theory, Research, and Clinical Applications (pp. 555 -

            574).  New YorkGuilford Press. 

Luthar, S. S.  (1999).  Poverty and Children's AdjustmentThousand Oaks, CA:  Sage Publications, Inc.

Luthar, S. S., Cicchetti, D., & Becker, B.  (2000).  The construct of resilience:  A critical evaluation and

guidelines for future work. Child Development, 71(3), 543-562.

Luthar, S. S., & Suchman, N. E.  (1999).  Relational psychotherapy mothers' group:  A developmentally

informed intervention for at-risk mothers.  Development and Psychopathology, 12, 235-253.

Luthar, S. S., & Zigler, E.  (1991).  Vulnerability and competence:  A review of research on resilience in

childhood.  American Journal of Orthopsychiatry, 61(1), 6-22.

Lyons-Ruth, K., Alpern, L. & Repacholi, B.  (1993).  Disorganized infant attachment classification and

maternal psychosocial problems as predictors of hostile-aggressive behavior in the preschool classroom.  Child Development, 64(2), 572-585. 

Masten, A. S., & Coatsworth, J. D.  (1998).  The development of competence in favorable and

unfavorable environments:  Lessons from research on successful children.  American Psychologist, 53(2), 205-220.

Masten, A. S., Best, K. M., & Garmezy, N.  (1990).  Resilience and development:  Contributions from the

            study of children who overcome adversity.  Development and Psychopathology, 2, 425-444.

McClelland, G. H., & Judd, C. M.  (1993).  Statistical difficulties of detecting interactions and moderator

            effects.  Psychological Bulletin, 114(2), 376-390.

McLoyd, V. C., Jayaratne, T. E., Ceballo, & Borquez, J.  (1994).  Unemployment and work interruption

among African American single mothers:  Effects on parenting and adolescent socioemotional

functioning. Child Development, 65(2), 562-589.

O'Keefe, M.  (1994).  Adjustment of children from maritally violent homes. Families in Society, 75, 403-415.

Osofsky, J. D. & Thompson, M. D.  (2000).  Adaptive and maladaptive parenting:  Perspectives on risk

            and protective factors.  In J. P. Shonkoff & S. J. Meisels, (Eds.), Handbook of early childhood

            intervention, 2nd ed. (pp. 54-75). New York: Cambridge University Press.

Pannacione, V., & Wahler, R.  (1986).  Child behavior, maternal depression, and social coercion as factors

            in the quality of child care.  Journal of Abnormal Child Psychology, 45, 747-758.

Patterson, G.  (1982).  Coercive family processes (Vol. 3)Eugene, OR:  Castalia.

Patterson, G. R., & Capaldi, D.  (1991).  Relation of parental transitions to boys' adjustment problems:  I.

A linear hypothesis.  II. Mothers at risk for transitions and unskilled parenting.  Developmental

Psychology, 27(3), 489-504.

Peterson, L., Ewigman, B., & Vandiver, T.  (1994).  Role of parental anger in low-income women: 

Discipline strategy, perceptions of behavior problems, and the need for control.   Journal of Clinical Child Psychology, 23(4), 435-443.

Pianta, R. C., Egeland, B., & Sroufe, L. A.  (1990).  Maternal stress and children's development: 

Prediction of school outcomes and identification of protective factors.  In J. Rolf, A. S. Masten, D Cicchetti, K. H. Nuechterlein, & S. Weintraub (Eds.), Risk and protective factors in the development of psychopathology (pp. 215-235).  New YorkCambridge University Press.

Pianta, R. C., & Steinberg, M.  (April, 1991).  Relationships between children and kindergarten teachers:

Associations with home and classroom behavior.  Paper presented at the meeting of the Society for Research in Child Development, Seattle, WA

Pollitt, E.  (1994).  Poverty and child development:  Relevance of research in developing countries to the

United StatesChild Development, 65(2), 283-295.

Radke-Yarrow, M., & Brown, E.  (1993).  Resilience and vulnerability in children of multiple-risk

            families.  Development and Psychopathology, 5(4), 581-592.

Radke-Yarrow, M., and Sherman, T. (1990). Hard growing: Children who survive. In J. Rolf, A. S.

            Masten, D. Cicchetti, K. H. Nuechterlein, & S. Weintraub (Eds.), Risk and protective factors in

            the development of psychopathology (pp. 97-119).  New YorkCambridge University Press.

Renken, B., Egeland, B., Marvinney, D., Mangelsdorf, S., & Sroufe, L. A.  (1989).  Early childhood

antecedents of aggression and passive-withdrawal in early elementary school.  Journal of

Personality, 57(2), 257-281.

Rohner, R. P.  (1986).  The warmth dimension:  Foundations of parental acceptance-rejection theory

Beverly Hills, CA:  Sage Publications.   

Rose, S. L., Rose, S. A., Feldman, J. F.  (1989).  Stability of behavior problems in very young children. 

Development and Psychopathology, 1(1), 5-19.

Rutter, M. (1983).  Statistical and personal interactions:  Facets and perspectives.  In D. Magnusson & V.

            L. Allen (Eds.), Human development:  An interactional perspectiveNew York:  Academic Press.

Rutter, M. (1985). Resilience in the face of adversity: Protective factors and resistance to psychiatric

            disorder.  British Journal of Psychiatry, 147, 598-611.

Sampson, R. J., & Laub, J. H.  (1994).  Urban poverty and the family context of delinquency:  A new look

at structure and process in a classic study.  Child Development, 65(2), 523-540.

Shaw, D. S., & Bell, R. Q.  (1993).  Developmental theories of parental contributors to antisocial

behavior.  Journal of Abnormal Child Psychology, 21(5), 493-518.

Shaw, D. S., Keenan, K., & Vondra, J. I.  (1994).  Developmental precursors of externalizing behavior:

Ages 1 to 3.  Developmental Psychology, 30, 355-364.    

Shaw, D. S., Bell, R. Q., & Gilliom, M. (2000).  A truly early starter model of antisocial behavior

revisited.  Clinical Child and Family Psychology Review, 3, 155-172.

Shaw, D. S., Owens, E. B., Vondra, J. I., Keenan, K., & Winslow, E. B.  (1996).  Early risk factors and

pathways in the development of early disruptive behavior problems.  Development and Psychopathology, 8, 679-699.

Shaw, D. S., & Vondra, J. I.  (1995).  Infant attachment security and maternal predictors of early behavior

problems:  A longitudinal study of low-income families.  Journal of Abnormal Child Psychology, 23(3), 335-357.

Shaw, D. S., Vondra, J. I., Hommerding, K. D., Keenan, K. & Dunn, M.  (1994).  Chronic family

adversity and early child behavior problems:  A longitudinal study of low income families. 

Journal of Child Psychology and Psychiatry, 35(6), 1109-1122.

Shaw, D. S., Winslow, E. B., Owens, E.  B., Vondra, J. I., Cohn, J. F., & Bell, R. Q.  (1998).  The

development of early externalizing problems among children from low-income families:  A

transformational perspective.  Journal of Abnormal Child Psychology, 26, 95-107.

Smith, J., & Prior, M. (1995). Temperament and stress resilience in school-age children:  A within-

families study.  Journal of the American Academy of Child and Adolescent Psychiatry, 34(2),


Smith, B., & Sechrest, L.  (1991).  Treatment of aptitude X treatment interactions.  Journal of Consulting

and Clinical Psychology, 59(2), 233-244.

Spielberger, C. D., Johnson, E. H., Russell, S. F., Crane, R. J., Jacobs, G. A., & Worden, T. J.  (1985).

            The experience and expression of anger:  Construction and validation of an anger expression

            scale.  In M. A. Chesney & R. H. Rosemann (Eds.), Anger and hostility in cardiovascular and

            behavioral disorders (pp. 5-30).  New York:  Hemisphere/McGraw-Hill.

Stouthamer-Loeber, M., Loeber, R., Farrington, D. P., Zhang, Q., van Kammen, W., & Maguin, E.

(1993).  The double edge of protective and risk factors for delinquency: Interrelations and

            developmental patterns. Development and Psychopathology, 5, 683-702.

Tschann, J. M., Kaiser, P., Chesney, M. A., Alkon, A., Boyce, W. T.  (1996).  Resilience and vulnerability

            among preschool children:  Family functioning, temperament, and behavior problems.  Journal of

            the American Academy of Child and Adolescent Psychiatry, 35(2), 184-192. 

U.S. Census Bureau. (2000a).  Poverty in the United States:  1999 [On-line].  Available: 


U.S. Census Bureau (2000b).  Historical poverty tables [On-line].  Available: 


Urban, J. Carlson, E., Egeland, B., & Sroufe, L. A.  (1991).  Patterns of individual adaptation across

childhood.  Development and Psychopathology, 3(4), 445-460.

Werner, E., & Smith, R. (1982).  Vulnerable but invincible:  A longitudinal study of resilient children and

            youth. New York:  McGraw Hill.

Weschler, D.  (1989).  Weschler Preschool and Primary Scale of Intelligence - Revised.  San Antonio,

            TX:  The Psychological Corporation.

Winslow, E. B., Shaw, D. S., Bruns, H., & Kiebler, K.  (March, 1995).  Parenting as a mediator of child

            behavior problems and maternal stress, support, and adjustment.  Paper presented at the meeting

            of the Society for Research in Child Development, Indianapolis, IN.







            1.  We have chosen to place attachment in the child domain because, as measured by the Strange Situation, it is fundamentally a measure of the child’s behavior, although we acknowledge that theoretically it is presumed to in part reflect the quality of the caregiving environment.