Differences in attachment security between African-American and white children: ethnicity or socio-economic status?

https://doi.org/10.1016/j.infbeh.2004.02.002Get rights and content

Abstract

The NICHD Early Childcare Research Network data set was used to examine differences in attachment security between African-American children (n = 142) and white children (n = 1002). African-American children's mean score on the Attachment Q-sort (AQS) [Monographs of the Society for Research in Child Development 60 (1995) 234] was substantially lower (.20) than that of white children's (.30). The pattern of covariation between attachment security and predictor variables was similar in the African-American and white subgroups. In both groups, maternal sensitivity was the strongest predictor of attachment security. A mediational model explaining the difference in attachment security included income and sensitivity: African-American ethnicity was related to low income which through (in-)sensitivity affected the quality of the infant–mother attachment relationship (family stress model). Our findings on African-American mother–infant dyads support one of the basic tenets of attachment theory: the association between maternal sensitivity and attachment security. Children of African-American and white families in the USA may be exposed to culturally specific experiences, but these do not alter the relation between attachment security and pertinent predictor variables. Poverty may, however, seriously hamper maternal sensitivity.

Introduction

Already in the first few years after birth, large differences between African-American and white children have emerged in various domains of development (Garcia Coll, 1990, McLoyd, 1990a, McLoyd, 1998, Spencer, 1990). In this paper we focus on differences in attachment security. Attachment is conceptualized as the emotional bond or tie of infants to their parents, and attachment theory has become a major source of hypotheses for research on the socio-emotional development of young children (Ainsworth, Blehar, Waters, & Wall, 1978; Bowlby, 1969). Central questions in this paper are whether African-American and white children differ in attachment security; and – when such differences exist – how they can be explained. Are differences in attachment security the result of factors that are associated with ethnicity, or should they be ascribed to socio-economic circumstances?

Differences in attachment security between African-American and white children may result from three separate processes. First, attachment measures may be culturally biased and yield unwarrantedly low scores for security in cultural groups different from the culture in which the measures were developed. Second, associations between attachment quality and assumed precursors of secure attachment may be divergent in different cultural contexts, i.e. in African-American and white families. And lastly, a third variable, related to ethnicity and attachment, may be responsible for differences in attachment security between the two groups.

Considering the first hypothesis, it must be noted that attachment researchers have been blamed of having an ethnocentric bias. Their measures of sensitivity and attachment security would be biased toward Western ways of thinking, and would result in lower rates of children that are considered secure in other ethnic or cultural groups (Jackson, 1993, Rothbaum et al., 2000). Evidence supporting this argument is, however, weak. First, as Van IJzendoorn and Sagi (1999) have shown in their review, cross-cultural attachment research has demonstrated the normativity of secure attachment in diverse cultures. Intra-cultural differences in the development of attachment appear larger than cross-cultural differences, and departures from the numerical normativity of secure attachment may be the result of incorrect application of the assessment procedure. For instance, when Strange Situation separations are not curtailed when infants are distressed for more than 20 s, the procedure is more than mildly stressful; and that may constitute a plausible explanation of the overrepresentation of insecure-resistant infants in one Japanese sample (Miyake, Chen, & Campos, 1985; see Van IJzendoorn & Sagi, 2001). Using the Attachment Q-sort (AQS; Waters, 1995) in a cross-cultural study, Posada et al. (1995) did not find greater similarity of Q-sort profiles within each of seven socio-cultural groups than between the groups. Moreover, mothers’ description of their most ideal child in different cultures turned out to be similar to the expert's description of a prototypical securely attached child (Posada et al., 1995, Vereijken, 1996). In sum, there is insufficient evidence to ascribe differences in attachment security between African-American and white children to biased measures of attachment.

The second explanation of differences in developmental outcomes between African-American and white children pertains to the existence of diverging developmental processes in different cultural groups. Two competing hypotheses can be distinguished (Rowe, Vazsony, & Flannery, 1994). The no group difference hypothesis is that there are few, if any, differences among ethnic groups in most developmental processes (where evidence for such processes can be found in the patterns of covariation among relevant variables). According to this hypothesis, members of a society of different ethnic origins are exposed to variables common to all ethnic groups in that society. Culturally specific experiences do not alter the associations among developmental variables. This hypothesis allows for main effect differences in variables but states that the correlations among variables do not differ between groups. In contrast, the group differences hypothesis presupposes culturally relative models of socialization (Garcia Coll, 1990, Ogbu, 1981, Ogbu, 1993) and it states that correlations among developmental variables are different between groups. According to this model, development occurs within cultural contexts that are associated with qualitatively different processes. Any observed differences in developmental processes are assumed to be adaptive responses to the demands of the cultural environment (Ogbu, 1981). Some studies on the effects of parental discipline have provided evidence for the group differences hypothesis: Authoritarian parenting (a restrictive, often physical parenting style) was found to be associated with negative socio-emotional outcomes for European-American children in preschool or school age, but not for African-American children (Baumrind, 1972, Baumrind, 1993; Deater-Deckard, Dodge, Bates, & Pettit, 1996; Dornbusch, Ritter, Leiderman, Roberts, & Fraleigh, 1987).

Cross-cultural studies on attachment have presented some evidence for the universality of the hypothesized association between parental sensitivity and attachment security in different cultures (the no group difference hypothesis), but not in all studies this association was found (Van IJzendoorn & Sagi, 1999). Because the meta-analytic association between attachment and sensitivity is only modest (r = .24, De Wolff & Van IJzendoorn, 1997), lack of statistical power may have prevented unequivocal decisions about observed associations between sensitivity and attachment in small samples (the majority of the studies across different cultures included less than 40 families). Alternatively, correlations among the pertinent variables may differ between groups (the group differences hypothesis), and divergent associations between attachment and sensitivity (or other precursors) may be found for African-American and white families. For instance, Jackson (1993) considered multiple caregiving as instantiated in African-American families as distinctive and without parallel in the cross-cultural literature on attachment. Child care for African-American children is generally conducted within a relatively large social network of friends and acquaintances of the family. There are often several adult caregivers with designated responsibility for infant care and a larger set of children and adults who also provide care. In Jackson's (1991) study,the number of households providing daily care for any one child ranged from 1 to 4; and the infants encountered on average 15 familiar adults on a recurring weekly basis. This seems to imply a context for the infant–mother attachment formation that is rather different from the experiences of most white American infants.

The third explanation of different developmental outcomes assumes that the patterns of covariation among relevant variables are similar in the different ethnic or cultural groups, but that a ‘third variable’ on which the groups differ is responsible for differences in outcome. The first candidate for such an explanatory factor distinguishing African-American and white children in the USA may be socio-economic status. African-American children generally come from families with lower incomes and a lower educational level of the parents than white children. Thus, diverging child outcomes among families with different ethnicities may as a matter of fact be caused by differences in family income (Jencks & Philips, 1998). Family income is associated with the development of children and youth (Huston, McLoyd, & Garcia Coll, 1994). Studies based on the National Longitudinal Survey of Youth-Child Supplement (NLSY-CS), Infant Health Development Program, and the Panel Study of Income Dynamics (PSID) have shown that income effects are strongest during the preschool and early school years (Duncan, Brooks-Gunn, & Klebanov, 1994; Duncan, Yeung, Brooks-Gunn, & Smith, 1998). This early childhood effect has been found to be particularly strong when low income is persistent.

Income may have a differential effect on distinct child outcomes. Stronger effects of adverse economic conditions have been found on children's school and cognitive achievement than on children's socio-emotional development (Duncan & Brooks-Gunn, 1997; Haveman & Wolfe, 1994). Nevertheless, Linver, Brooks-Gunn, and Kohen (2002) demonstrated that child behavior problems were associated with low family income. Two processes were responsible this association. First, low income was related to child behavior problems via maternal emotional distress, which was related to observed parenting practices. This mediational model is known as the family stress model. The family stress model postulates that low income influences children's development (measured in terms of school achievement, school engagement, or behavior problems) because of its impact on parent mental health, which then influences parenting practices, which in turn are associated with children and youth outcomes (Conger et al., 1992, Conger et al., 1993; Elder & Caspi, 1988; McLoyd, 1989). Empirical studies have demonstrated that economic hardship diminishes parental abilities to provide warm, responsive parenting (Dodge, Pettit, & Bates, 1994) and contributes to an increase in the use of harsh punishment (Dodge et al., 1994; McLoyd, Jayaratne, Ceballo, & Borquez, 1994; Smith & Brook-Gunn, 1997). A second model, known as the investment model, hypothesizes that income is associated with child development because it enables parents to purchase materials, experiences, and services that are beneficial to children's well-being and development (Becker & Thomes, 1986; Haveman & Wolfe, 1994; Mayer, 1997).

Studies on attachment have shown a high incidence of insecurity in poor families where poverty is combined with other social risks such as social isolation, maternal depression, or inadequate caretaking (e.g., Barnard et al., 1988; Lyons-Ruth, Connell, and Grunebaum, 1990). In a meta-analysis of infant attachment (Van IJzendoorn, Goldberg, Kroonenberg, & Frenkel, 1992) the distribution of classifications in low-SES samples was not contrasted with the distribution of middle-class samples, but in a meta-analysis of adult attachment, secure representations of attachment were underrepresented in low-SES samples (Van IJzendoorn & Bakermans-Kranenburg, 1996). Research on attachment in African-American children involved, apparently without exception, only low-income families (Sims-Stanford, 1997, Nelson, 1991), which is not surprising as African-American ethnicity and low income tend to go together.

In sum, the focus of our research is whether African-American and white children differ in attachment security; and – when such a difference exist – whether differences in attachment security can be explained by patterns of covariation that are either associated with ethnicity, or with socio-economic differences. Our database is the NICHD Study of Early Child Care (see NICHD, 1996, NICHD, 1997, NICHD, 2000, NICHD, 2002, NICHD, 2003b). The NICHD study covers a large, more or less representative sample of US children (N = 1364 children) who were recruited at ten different sites in the USA (see Section 2). About 13% of these children had mothers who were African-American, and about 83% had white mothers. The children were followed across the first seven years of their lives, and their development was monitored at regular intervals.

Attachment was assessed with the Attachment Q-sort (Waters, 1995). The AQS consists of 90 specific behavioral descriptions of 12–48-month-old children in the natural home-setting, with special emphasis on secure-base behavior (Vaughn & Waters, 1990). Waters and Deane (1985) introduced the AQS for assessing attachment security in infants and toddlers as an alternative to the Strange Situation laboratory procedure (SSP, Ainsworth et al., 1978), and it appears to have some advantages over the SSP (see Van IJzendoorn, Vereijken, Bakermans-Kranenburg, & Riksen-Walraven, in press). For the current study, it is of relevance that AQS observations are conducted in the home, and they may therefore have higher ecological validity. The SSP has been criticized for its lack of ecological validity (Bronfenbrenner, 1979). Furthermore, because the application of the AQS does not require the stressful separations used in the SSP, the method can be applied in cultures and populations in which parent–infant separations may be less common (e.g., Kazui, Endo, Tanaka, Sakagami, & Suganuma, 2000). We therefore use the AQS (at 24 months of the children's age) to explore differences in attachment between the two ethnic groups. We include ratings of maternal sensitivity during the first two years as well as relevant background variables.

Section snippets

Participants

Participants in the NICHD Study of Early Child Care were recruited from 10 sites around the USA throughout 1991. From among 8986 mothers giving birth during selected 24-h sampling periods, potential participants were screened to determine their eligibility for the study. Subjects were excluded from the sample if (a) the mother was under 18, (b) the mother was not conversant in English, (c) the family was planning to move, (d) the child was hospitalized for more than 7 days following birth or

What are the similarities and differences between African-American and white families?

In Table 1 means and standard deviations of socio-economic, child rearing, and developmental variables for the African-American and white families are presented. From this table it can be derived that in almost all respects the African-American and white families differ considerably. African-American mothers (M = 24.68, S.D. = 5.38) were on average almost four years younger than the white mothers (M = 28.98, S.D. = 5.38) in our sample, and the African-American children included in the sample

Discussion

The extensive data set of the NICHD Early Childcare Research Network was used to examine differences in attachment security between African-American and white children at two years of age. African-American children's attachment security was substantially lower than in white children. In fact, their mean score of .20 was similar to the meta-analytical average of .21 for children in clinical samples (Van IJzendoorn et al., in press). The African-American children did not only show less secure

Acknowledgements

This study was conducted by the NICHD Early Child Care Research Network supported by NICHD through a cooperative agreement that calls for scientific collaboration between the grantees and NICHD staff. The authors acknowledge the generous way in which the NICHD Study on Early Child Care has made this unique data set available for further secondary analyses. They are grateful to Mariëlle Linting and Patrick Groenen for their assistance in the data preparation and their contributions to our

References (72)

  • O. Aviezer et al.

    Ecological constraints on the formation of infant–mother attachment relations: When maternal sensitivity becomes ineffective

    Infant Behavior and Development

    (2003)
  • M.D. Ainsworth

    Infancy in Uganda

    (1967)
  • M.D.S. Ainsworth et al.

    Patterns of attachment

    (1978)
  • K.E. Barnard et al.

    Prevention of parenting alterations for women with low social support

    Psychiatry

    (1988)
  • R.M. Baron et al.

    The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations

    Journal of Personality and Social Psychology

    (1986)
  • D. Baumrind

    An exploratory study of socialization effects of black children: Some black–white comparisons

    Child Development

    (1972)
  • D. Baumrind

    The average expectable environment is not good enough: A response to Scarr

    Child Development

    (1993)
  • G.S. Becker et al.

    Human capital and the rise and fall of families

    Journal of Labor Economics

    (1986)
  • J. Belsky

    The determinants of parenting: A process model

    Child Development

    (1984)
  • J. Belsky

    Modern evolutionary theory and patterns of attachment

  • J. Bowlby

    Attachment and loss. Vol. I. Attachment

    (1969)
  • U. Bronfenbrenner

    The ecology of human development

    (1979)
  • R.D. Conger et al.

    A family process model of economic hardship and adjustment of early adolescent boys

    Child Development

    (1992)
  • R.D. Conger et al.

    Family economic stress and adjustment of early adolescent girls

    Developmental Psychology

    (1993)
  • K. Deater-Deckard et al.

    Physical discipline among African American and European American mothers: Links to children's externalizing behaviours

    Developmental Psychology

    (1996)
  • M.S. De Wolff et al.

    Sensitivity and attachment: A meta-analysis on parental antecedents of infant attachment

    Child Development

    (1997)
  • K.H. Dodge et al.

    Socialization mediators of the relation between socioeconomic status and child conduct problems

    Child Development

    (1994)
  • S.M. Dornbusch et al.

    The relation of parenting style to adolescence performance

    Child Development

    (1987)
  • G.J. Duncan et al.

    Economic deprivation and early childhood development

    Child Development

    (1994)
  • G.J. Duncan et al.

    How much does childhood poverty affect the life chances of children?

    American Sociological Review

    (1998)
  • G.H. Elder et al.

    Economic stress in lives: Developmental perspectives

    Journal of Social Issues

    (1988)
  • C.T. Garcia Coll

    Developmental outcome of minority infants: A process-oriented look into our beginnings

    Child Development

    (1990)
  • R. Haveman et al.

    Succeeding generations: On the effects of investments in children

    (1994)
  • R. Haveman et al.

    Child events and circumstances influencing high school completion

    Demography

    (1991)
  • Huston A. C., McLoyd, V. C., & Garcia Coll, C. T. (Eds.). (1994). Children and poverty [Special issue], Child...
  • Jackson, J. F. (1991). Multiple caregiving among African-Americans and infant attachments: Issues and an exploratory...
  • J.F. Jackson

    Multiple caregiving among African Americans and infant attachment: The need for an emic approach

    Human Development

    (1993)
  • M. Kazui et al.

    Intergenerational transmission of attachment Japanese mother–child dyads

    Japanese Journal of Educational Psychology

    (2000)
  • M.R. Linver et al.

    Family processes as pathways from income to young children's development

    Developmental Psychology

    (2002)
  • K. Lyons-Ruth et al.

    Infants at social risk: Maternal depression and family support services as mediators of infant development and security of attachment

    Child Development

    (1990)
  • S. Londerville et al.

    Security of attachment, compliance, and maternal training methods in the second year of life

    Developmental Psychology

    (1981)
  • S.E. Mayer

    What money can’t buy: Family income and children's life chances.

    (1997)
  • V.C. McLoyd

    Socialization and development in a changing economy: The effect of paternal job and income loss on children

    American Psychologist

    (1989)
  • Cited by (0)

    View full text