Understanding Parenting

Parenting has been a centerpiece of developmental inquiry from the beginning of the field, reflecting the firm belief that childrearing makes the child. Only in the 1990s has this belief come under intense scrutiny, in debates over the influence of parenting relative to that of genetics and peers (Borkowski et al., in press; Harris, 1995, 1998; Rowe, 1994). While these debates have focused on children of school age and older (few dispute the significant role of parents during the earliest years of life), they have implications for the understanding of the more enduring influences of parenting in adolescence and adulthood. The controversy, moreover, highlights important shifts in studying and understanding the role of parents in early development.

The classic, early studies of childrearing sought to identify styles of parenting that promoted competent behavior in preschoolers (i.e., a child who is happy, self-reliant, self-controlled, friendly, and cooperative as distinct from withdrawn or immature) (see Baldwin et al., 1945; Baumrind, 1967, 1971, 1973; Emmerich, 1977). The answers yielded by this research highlighted the combined influence of clear standards of conduct, firm control, and ample warmth. The heritage of this research is an abiding interest in both the management or control function and the emotional quality of parent-child relationships. In the ensuing 25 years, however, static characteristics of parents as restrictive or warm have been challenged by substantial evidence that parenting changes over time, varies from one child to another, and is not just received, but is also shaped, by the child's own behavior (Grusec and Goodnow, 1994; Holden and Miller, 1999).

The result is a concept of parenting and parental influence that is more complex and conditional than that which emerged from prior eras of research (Collins et al., 2000; Maccoby, 2000). Succinct formulas for good parenting have been replaced by an appreciation for the many ways in which parents adjust what they do in response to the needs and characteristics of their children, the conditions in which they live, and the circumstances of their own lives (Cowan and Cowan, 1992; Elder, 1991; Holden and O'Dell, 1995). Whereas scientists used to study parenting in isolation, contemporary researchers take into account the network of contexts in which parenting is embedded. These contexts include the child care programs that children attend, the peers they associate with, the stability and socioeconomic strata of their families, their parents' marital relationship, the neighborhoods they live in, and the times they live in.

Challenges to the notion that children are relatively passive players in the socialization process (Bell, 1968; Bell and Chapman, 1986; Engfer et al., 1994) have led to a substantial body of evidence showing the many ways in children contribute to their rearing environments, including influences on the parenting they receive. Researchers now realize that they need to consider the ways in which parents affect children and children affect parents. The growing reliance on research designs that address the interplay of genetics and socialization has both confirmed the substantial influence of parenting on child development and increased awareness of the complex ways in which parenting intersects with the child's inherited strengths and vulnerabilities to affect the pathways that are followed en route to adulthood (see Collins et al., 2000; Rutter et al., in press). The methodological challenges involved are considerable, as discussed in more detail in connection with the effects of impaired parenting. Accordingly, research on parenting has become a very complex endeavor.

Bringing Fathers into the Picture

The research discussed in this chapter on the multifaceted dimensions of parenting is primarily focused on mothering. Fathering, in contrast, has received less attention, and this literature has tended to focus on men's economic contributions to their families, the developmental consequences of father absence, and distinctions between the roles of fathers and mothers (Lamb, 1999; Parke, 1996; Pleck and Pleck, 1997). This began to change in the 1970s with a growing emphasis on family dynamics, including the marital relationship, as they affect child development (see Belsky, 1984; Cummings and O'Reilly, 1997; Dunst, 1985) and is continuing to change as an increasingly interdisciplinary group of scholars is exploring how fathers affect developmental trajectories (see Cabrera et al., 2000).

Contemporary research on fatherhood has highlighted several themes. First, fathers seem to be both more and less involved in their children's lives today than was true even a decade ago, revealing a growing dichotomy in children's experiences of fathering. On one hand, there is evidence that paternal involvement has increased over the past three decades, as seen in higher rates of single fathers raising children, greater involvement of fathers in child care while their wives are in the labor force, and more self-reported time spent by fathers with their children (Pleck, 1997; Yeung et al., 1998). At the same time, unprecedented numbers of children are now spending part or all of their childhoods in families headed by single mothers (see Chapter 10). While there is a growing appreciation of the extensive father involvement that can characterize children in single-mother families and extensive policy interest in the issues involved, scientists are only beginning to explore the factors that predict this involvement (Coley and Chase-Lansdale, 1999; W. Johnson, 1998) and its effects on children (see Garfinkel et al., 1994). It certainly appears to be the case that fathers' economic situation is closely involved. The consequences of these markedly different experiences of fathering for children's development, for their conceptions of parenting, and for their assumption of parenting roles remain largely unexamined.

Second, fathering is increasingly viewed as involving multiple functions that go well beyond the role of breadwinner (Amato and Rivera, 1999; Lamb et al., 1985a; Parke, 1996; Parke and Buriel, 1998). Indeed, there is growing acceptance that father involvement per se is not necessarily linked to positive outcomes for children (see, for example, Hoffman et al., 1999). Rather, it is the variety of ways in which fathers take responsibility for their children that many now believe to be the most important component of fathering (Lamb, in press). This entails not only financial responsibility, but also functions that have typically been thought of as the purview of mothers—tasks such as taking children to the doctor or to lessons, arranging for and transporting the child to and from child care, monitoring the child's safety and whereabouts, and scheduling play dates. As with mothers, the emotional quality of the father-child relationship also appears to be extremely important to children's adjustment and well-being and may, in fact, influence the benefits of increased involvement and responsibility.

Finally, despite the rapid changes affecting the ethnic and racial composition of the nation's families, there is almost no research on how the roles of fathers and other men in young children's lives are evolving in the context of diverse values and family structures (Cabrera et al., 2000). Both theoretical models and empirical work need to incorporate culturally diverse conceptions of fatherhood related to race and ethnicity, as well as other dimensions of culture such as immigrant status and religion.

In the following discussion of parenting, we note the research that specifically addresses fathers. The examples are far from plentiful, but we are hopeful that the renewed attention to fathering that is now characterizing developmental science will foster expanded efforts to study parenting in all of its varieties and certainly as it involves both mothers and fathers.

Security of Attachment

Although virtually all infants become attached to their caregivers, attachment relationships differ in how much security they provide. Assessments of attachment security, whether conducted in a laboratory or at home, focus on the child's exploratory behavior in the presence of the caregiver, responses to separation, and reactions upon reunion. A secure attachment is assumed to exist when the infant or toddler explores comfortably in the presence of his or her caregiver, keeps track of and seeks proximity with the caregiver, happily and eagerly seeks contact after having been separated, and, in general, shows signs of trust and delight in the caregiver's presence. In contrast, children whose exploratory play is disrupted because they are preoccupied with the caregiver, who avoid or resist contact after separation, display distress and anger upon reunion, and are not easily comforted are considered insecurely attached. An insecure attachment is not, however, equivalent to no attachment at all. Even a young child who is insecure about the caregiver's nurturance derives important emotional support from her presence that is not derived from the company of someone to whom the child has no attachment at all. After all, even a resistant infant turns to the caregiver for help.

Recent research on children who have experienced highly disruptive, sometimes abusive care has led to important refinements in views of insecure attachment. Studies of physically abused infants and toddlers have noted significantly elevated proportions of insecure attachments (Crittenden, 1988; Lyons-Ruth et al., 1987), as have studies of children of clinically depressed mothers (DeMulder and Radke-Yarrow, 1991; Lyons-Ruth et al., 1990, 1991). Studies of neglected children, such as those reared in orphanages or removed from their homes because of severe neglect, have shown that some, but certainly not all, of these children do not seem to organize their behavior in meaningful ways around one or a few adults. They do not fit typical patterns of insecurity, but rather display inconsistent and disorganized responses to their caregivers. The field is just beginning to document these unusual patterns of attachment behavior and to explore their clinical roots and implications for children's development. We are far from being able to say anything definitive about these disordered patterns of attachment behavior, but they form one of the cores of the nascent field of infant mental health (Osofsky and Fitzgerald, 2000; Zeanah, 2000). This field of clinical research, albeit new, highlights what has become increasingly evident: infants and young children have rich emotional/psychological lives and can suffer in ways that heretofore had never been realized.

The interest in documenting unusual patterns of attachment behavior in search of a better understanding of infant mental health and disorder increases the need to broaden cultural understanding of attachment and the assessment of its security. It is important to realize that the laboratory assessments that have formed the basis for much of the research on attachment security have been designed to produce only mild challenges for the infant. The meaning of the laboratory assessment depends on whether the challenges are both mild and meaningful within the child's culture and life history.

For example, the standard laboratory assessment (termed the “strange situation”) relies on brief (e.g., 3 minute) separations from the parent. In cultures in which infants are frequently separated from parents for brief periods, the child's reactions are presumably influenced by expectations, based on previous history, that the parent will return and be helpful. In cultures in which separations rarely occur, it is presumed that these experimentally imposed separations may take on a very different meaning for the infant. Indeed, research on Japanese infants who are rarely separated from parents during their first year initially demonstrated high rates of presumably insecure attachment (Takahashi, 1986, 1990). Later reinterpretations of the results, however, emphasized the vast difference in the strange situation between the Japanese and European-American cultures (van IJzendoorn and Sagi, 1999).

A study of desirable and undesirable attachment behavior among white and Hispanic (Puerto Rican) mothers provides a compelling illustration of these differences (Harwood et al., 1995). The white mothers preferred that toddlers balance autonomy and relatedness (playing at a distance and involving the mother prior to separation and greeting the mother happily during the reunion), and they disliked clinginess (clinging to the mother prior to separation, crying continuously during separation, and being unhappy during the reunion). In contrast, Puerto Rican mothers preferred that toddlers display respectfulness (sitting near the mother and waiting for a signal before playing with the toys prior to separation, waiting quietly for the mother to return during separation), and they disliked highly active or avoidant (ignoring the mother before, during, and after separation) behavior.

This and other cross-cultural evidence on attachment raises significant issues regarding the ways in which parents and young children form expectations about each other and, in turn, behave and react in each other's presence (and in the strange situation). We strongly suspect that, across all cultures, children form attachments and use parents as sources of security and comfort. Even when the relationship is somewhat insecure, children seek comfort and maintain proximity to parents, deriving important emotional support from the caregiver's presence that other adults cannot provide; however, they do not derive the same developmental benefits that accrue from a secure attachment.

Specific attachment patterns result from an intricate interplay among characteristics of the child, the capacities of the parent, and the broader context of their relationship (see Isabella, 1995; Lamb et al., 1985b; Thompson, 1999a, for reviews). Secure attachments are seen more often in the context of parenting that is dependable and sensitive to the child's intentions and needs, enabling the child to count on the caregiver's future availability and assistance (Ainsworth et al., 1978; Belsky, 1999; De Wolff and van IJzendoorn, 1997; Isabella, 1995; Thompson, 1997, 1998a). Infants and toddlers are less likely to establish secure attachments with caregivers who are generally detached, intrusive, erratic, or rejecting. The important role of sensitive caregiving in the establishment of secure attachments is compellingly illustrated by a recent intervention that randomly assigned low-income mothers of infants who were observed in the first two weeks of life to be irritable to a program designed to enhance maternal sensitivity and responsiveness or to a control group. After 9 months, the mothers who had received the programs were significantly more responsive and stimulating than the control group mothers and their infants engaged in more sophisticated exploratory behavior and were significantly more likely to be securely attached (van den Boom, 1994, 1995). A follow-up when the children were 3½ years old documented sustained effects in maternal sensitivity, attachment security, and the children's observed cooperation with the mother. Interestingly, the husbands of mothers who participated in the intervention were also more responsive to their preschoolers. Adoption studies add to the evidence regarding the importance of sensitive, responsive care. Toddlers fostered or adopted from conditions of extreme neglect have been found to reorganize their attachment behavior over time, exhibiting behavior reflective of secure expectations of support and comfort as they adapt to foster or adoptive parents who provide sensitive, responsive, and consistent care (Chisholm, 1998; Dozier, in press a, in press b; O'Connor et al., 1999).

Providing sensitive, responsive, and consistent parenting of infants and toddlers is challenging work. Both characteristics of the child and of the parent can make this type of parenting difficult to achieve. For example, newborns who continue to react to repeated stimuli after other newborns have tuned out or habituated to the repeated stimulation are somewhat more likely to form insecure attachment relationships to caregivers (Warren et al., 1997). Babies who become disorganized when stressed and those who get very upset when limits are placed on their actions are also somewhat more likely to develop insecure attachments (Fox, 1985; Gunnar et al., 1996; Izard et al., 1991). These infant characteristics don't predestine children to insecure attachment; rather they shift the odds. This may be because it is harder for people to provide the sensitive parenting such children need. It may be because it is not clear what the baby needs or because the needs of such infants exceed the time, attention, and sensitivity that the parents can provide given all of the demands on them. Indeed, when parents can manage to maintain high degrees of sensitivity and responsiveness, even temperamentally difficult infants develop secure and trusting relationships (Goldberg, 1990; Mangelsdorf et al., 1990; van IJzendoorn et al., 1995). Challenging parental life circumstances can also result in an imbalance between the infant's needs and what the parent can provide. Emotional problems such as depression, economic stress, and marital conflict can interfere with sensitive and responsive parenting, be disruptive of secure attachments (see Belsky and Isabella, 1988; Thompson, 1999b; Waters, 1978), and constitute a significant source of instability over time in attachment security.

It also appears to be the case that atypical attachments are more common among atypical samples, including premature infants, children with Down syndrome, and children with autism (Atkinson et al., 1999; Capps et al., 1994). In particular, a significantly larger share of children at the extremes of reproductive risk or who have an identifiable developmental disability display disorganized or unclassifiable patterns of attachment to their mothers. Much remains to be understood about the meaning and consequences of atypical attachments. They may arise from problems parents experience in being sensitive to their child (i.e., difficulty of reading the infant's cues), from these children's cognitive limitations, from the added stress that can accompany raising a child with special needs, or from limitations of the typical model for studying attachment when applied to these special populations. There is a tremendous need for research in this area, given its role in elucidating child factors and surrounding conditions that impinge on early attachments, as well as the developmental significance of behavioral differences in patterns of relating to important others among both atypically and typically developing children (see Vondra and Barnett, 1999).

Mothers and Others

The large majority of research on early attachments has focused on the parent-child relationship and, specifically, on the mother-infant relationship, despite the fact that young children establish close relationships with a surprising variety of people, including relatives, child care providers, and friends. Children certainly develop secure attachments to their fathers that do not depend on the security they derive from their attachments to mothers (Thompson et al., 1985). Grandmothers are also important attachment figures, and their support of the mother can facilitate secure attachment in infants (Crockenberg, 1987; Myers et al., 1987). Grandmothers are an especially important source of child care during the earliest months and years of life, as we discuss in Chapter 11.

Howes (1999) proposed the following three criteria for identification of attachment figures other than the mother: provision of physical and emotional care, continuity or consistency in the child's life, and emotional investment in the child. We do not know whether there is a specific limit to the number of people with whom very close emotional connections can be established at different ages. Regardless of their number and variety, from the child's perspective, close personal relationships are not interchangeable. The child may turn to a substitute attachment figure or even a relatively unfamiliar but friendly person when the preferred one is not available, but the distress about the loss of one beloved person is not easily alleviated by the ongoing availability of the others, as many parents faced with their children's reactions when a beloved babysitter or child care teacher goes out of their lives can attest.

Young children clearly benefit from opportunities to develop close relationships with different caregivers. As with the mother, the security of these relationships is based primarily on the trust and confidence that each adult has inspired in the child. It is also clear that the child's primary caregiver (usually the mother) remains central in this constellation of attachment relationships (Berlin and Cassidy, 1999; Howes, 1999; NICHD Early Child Care Research Network, 1997a). The security of attachment between a mother and her child is more influential on early psychosocial growth than are the relationships a child has with other caregivers at home or outside the home, and even children in extensive child care continue to show an overwhelming preference for their mothers (Easterbrooks and Goldberg, 1990; NICHD Early Child Care Research Network, 1997a, 1998a, 1998b, 1999a, in press(a)). Although there once was concern that spending many hours in nonparental care might undermine the child's primary relationships, recent studies now reassure parents that this is not the case (Berlin and Cassidy, 1999; Easterbrooks and Goldberg, 1990; Howes, 1999; NICHD Early Child Care Research Network, 1997a).

Moreover, attachment relationships are specific to each adult, so that an insecure attachment to one caregiver may develop at the same time that a secure relationship grows with another (Howes et al., 1988; Suess et al., 1992). For example, children often exhibit secure attachment behavior with one parent but not the other (Belsky et al., 1996a). Infants and toddlers who develop secure attachments either to their mothers or their child care providers are observed to be more mature and positive in their interactions with adults and peers than are children who lack a secure attachment. However, the most socially skilled children are those who have established secure attachments with both their mothers and care providers (Howes et al., 1995a). In this context, it is important to recognize that child care can be used effectively to provide respite for highly stressed parents who may be prone to child abuse or at risk of having their children placed in foster care (Crittenden, 1983; Kempe, 1987; Roditti, 1995; Subramanian, 1985). Unfortunately, as a result of pervasively high turnover in child care providers and frequent changes in arrangements, children are more often insecurely than securely attached to their child care providers (Galinsky et al., 1994; Howes, 1999; Seltenheim et al., 1997; Whitebook et al., 1990).

At a minimum, this knowledge indicates that the significance of a young child's attachments to caregivers other than their mothers and fathers merit respect and consideration. It also raises very significant issues for intervention efforts focused exclusively on parent-child relationships and suggests the value of extending these efforts to embrace other significant adults in the young child's life.

Links to Development

Early attachments are important not only as an indicator of the parent-child relationship, but also for their significant effects on other aspects of the child's functioning. They appear to have their most consistent and enduring influence on young children's social and emotional development, although they also foster the exploratory behavior that is so vital to early learning and seem to bolster parents' efforts to support learning (Matas et al., 1978; also, see Thompson, 1998b and 1999a, for reviews). We are, however, only beginning to understand the mechanisms that underlie these connections between parent-child attachment and developmental outcomes.

Longitudinal studies suggest that early attachments set the stage for other relationships, as children move into the broader world beyond the immediate family (Bretherton and Munholland, 1999; Sroufe and Fleeson, 1986, 1988; Thompson, 1998a). This occurs as young children acquire the ability to encode their early attachment relationships at the level of mental representations, which, in turn, guide their expectations about the availability and responsiveness of other partners. In this sense, research on internal representations of relationships converges with cognitive research on the child's unfolding theory of mind, as reviewed in Chapter 6. Securely attached young children compared with their insecurely attached peers have an easier time developing positive, supportive relationships with teachers, friends, and others whom they encounter as they grow up (Sroufe and Egeland, 1991; Sroufe et al., 1993; Thompson, 1998a, 1999a). Securely attached children may respond more positively to unfamiliar people (such as new classmates, a family acquaintance, or a substitute teacher) as well. It appears, however, that the positive expectations for close relationships that are inspired by a secure parent-child relationship—or, in the case of insecurely attached children, their distrust or ambivalence—are most apparent in their encounters with familiar partners.

There is also emerging evidence that securely attached young children are found, for example, to have a more balanced self-concept (Cassidy, 1988; Verschueren et al., 1996), more advanced memory processes (Belsky et al., 1996b; Kirsh and Cassidy, 1997), a more sophisticated grasp of emotion (Laible and Thompson, 1998), a more positive understanding of friendship (Cassidy et al., 1996; Kerns, 1996; Park and Waters, 1989) and they show greater conscience development (Kochanska, 1995, 1997; Laible and Thompson, in press) than insecurely attached children. These associations are especially evident when attachment security and other behaviors are measured at the same point in time, thus displaying a dense web of associated outcomes.

Beyond these specific developmental outcomes, secure attachments seem to play a very important role in shaping the systems that underlie children's reactivity to stressful situations. As discussed in Chapter 8, experiments with animals have yielded similar findings (Suomi, 1997) and further suggest that early mothering can affect the neural circuitry that governs behavioral stress responses in the offspring (Caldji et al., 1998; Liu et al., 1997). The development of noninvasive means of studying the activation of the stress-hormone system that produces cortisol has allowed the study of stress physiology in the everyday lives of infants and young children. The results of these studies indicate that, as in the work on nonhuman primates (Gunnar et al., 1981; Levine and Wiener, 1988) and human adults (Cohen and Wills, 1985), stress in young children is intimately linked with social experiences. About the time that infants begin to form specific attachments to adults, the presence of caregivers who are warm and responsive begins to buffer or prevent elevations in stress hormones, even in situations that elicit behavioral indicators of distress in the infant (Gunnar et al., 1996; Nachmias et al., 1996; Spangler and Schieche, 1994). In contrast, insecure attachment relationships are associated with higher cortisol levels in potentially threatening situations (Gunnar et al., 1992; Hertsgaard et al., 1995; Nachmias et al., 1996; Tout et al., 1998).

For example, in one study, toddlers were exposed to a live clown who entered the room and invited them to “come over and play.” Toddlers who were securely attached to the parent who accompanied them showed no rise in stress hormones to this strange event, even if they were frightened and wary of the clown and were generally described by the parent as more temperamentally fearful and anxious. In contrast, toddlers who showed the same behavioral signs of fear and wariness and were described as having a similarly fearful and anxious temperament, but who had an insecure attachment to the parent who was with them, showed significant elevations in this stress hormone. This was true despite the fact that the security of the attachment relationship was assessed separately, on a different day, and in a different context (i.e., the strange situation; Nachmias et al., 1996). These studies with infants and young children seem to be saying that secure emotional relationships with adults appear to be at least as critical as individual differences in temperament in determining stress system reactivity and regulation.

Beyond this emerging evidence regarding physiological reactions to stress, there is much to learn about how secure attachments function to promote and protect early development. Some have proposed that secure attachments enhance the child's receptivity to other facets of parents' socialization efforts. From this standpoint, a secure attachment inducts the child into what has been characterized as a “mutual orientation of positive reciprocity” between parent and child (Kochanska, 1997; Maccoby, 1983, 1992; Maccoby and Martin, 1983). This mutuality, in turn, heightens the child's receptivity to the many ways in which parents socialize their children to get along with others, deal effectively with conflict, and become motivated early learners (Kochanska, 1997; Kochanska and Thompson, 1997; Waters et al., 1991). The children, in effect, are more receptive to the parent's instruction, guidance, and teaching, which then reinforces the parent's sensitive parenting and, in all likelihood, further binds their secure attachment.

This evidence of the developmental significance of secure attachments supports the focus on relationship building in early intervention studies with high-risk populations of children. It is also important to recognize, however, that the effects of early attachment relationships are provisional and contingent on many other influences on psychosocial growth, as well as on continuity or change in the parent-child relationship itself (Sroufe et al., 1990, 1999). The security or insecurity of attachment relationships can change in the early years of life. A child who begins with an insecure relationship may, for example, later have opportunities to develop a sense of secure confidence in the same caregiver. Changes in attachment may arise from changing family circumstances, such as the birth of a sibling or periods of family stress (Cummings and Davies, 1994a; Teti et al., 1996a; Vaughn et al., 1979). There is therefore no guarantee that the influence of early attachment security will endure, unless that security is maintained for the child in the years that follow. The instability of early attachments renders efforts to trace long-term consequences very difficult. At best, we can conclude that the effects of early secure attachments are conditional. They shift the odds toward more adaptive development, but subsequent experiences and relationships can modify their longer-term impacts, sometimes substantially.

Fostering Cooperation and the Development of a Conscience

The growth of cooperation in the context of close relationships has been studied much less intensively in young children than has the growth of love in the context of attachment. Yet at the same time that attachment security is taking shape late in the first year through the sensitivity and warmth of the caregiver, another dimension of the relationship is being forged by the negotiation of conflict between parent and child. Developmental scientists are showing renewed attention to this aspect of the parent-child relationship because of its relevance to the early origins of psychosocial problems in young children, including defiance, withdrawal, and conduct problems (Caspi et al., 1995; Moffitt, 1990; White et al., 1990), as well as its contribution to the emergence of conscience, moral values, and consideration for others (Kochanska and Thompson, 1997).

Young children can experience conflict with virtually every family member, as well as with the peers with whom they play. As noted earlier, for example, getting along with peers is one of the central developmental tasks of early childhood. Sibling relationships are also a potent arena for conflict between young children, as well as for empathy, cooperation, and social comparison (Dunn, 1993; Dunn and Kendrick, 1982). How parents manage these episodes of conflict can be significant for how young children learn about the feelings of others, the skills of competent sociability, and how to negotiate and cooperate. Even more important, however, is conflict between a young child and a parent because of the significance of their attachment relationship and the adult's capacity to guide the child in learning how to manage disagreement and defiance.

Young children's conflicts with caregivers who are skilled at helping them learn to manage experiences of disagreement and defiance early in life can provide a foundation for the growth of empathy and prosocial motivation, as well as the development of skills for negotiating and successfully resolving conflicts with others (Eisenberg and Murphy, 1995; Goodnow, 1996, 1997). These conflicts do not need to be momentous; in fact, they often arise in the context of everyday exchanges in which, for example, parents say “no” to a child's request (e.g., for a snack, for a particular dress) or attempt to gain the child's cooperation (e.g., picking up toys or taking a bath). In this light, how young children experience conflict with their caregivers provides a forum for learning how to address conflict in their encounters with others throughout life. Conflicts and the negotiations they entail also provide essential practice as children learn acceptable ways to elicit help and to be assertive about their own needs and interests. They also provide opportunities for parents to learn how best to issue directives and make requests of their child. Little is currently known with assurance about how these experiences become catalysts for the growth of prosocial behavior and the rudiments of conscience, or the development of dysfunctional social behavior. It is clear, however, that nothing focuses a young child's attention on what others are thinking, feeling, and expecting better than the realization that conflict with that person must be resolved.

Research in this area has moved away from static characterizations of parenting style (e.g., authoritarian, permissive) toward a more dynamic understanding of what parents are doing when they set limits, create incentives, and administer punishments. As a result, researchers are now trying to understand how parents and others work with young children to foster capacities for safe, socially acceptable, self-regulated behavior in the context of conflict. This, in turn, shifts attention from whether parents are doing the right things or the wrong things to limit unacceptable behaviors, to how they encourage the joint resolution of conflict and the social understanding and skills that come with it. The focus of inquiry is thus less on the moment of conflict, anger, or frustration and more on what happens next.

This perspective also focuses attention on moments of “negotiable disagreement” (Goodnow, 1996), in which children try out a variety of strategies and in which parents—North American parents at least—teach children about the more or less acceptable ways to phrase a dissent or to negotiate a compromise (Kuczynski, 1993; Kuczynski and Kochanska, 1990). Among the strategies that children learn are deferral (“later, ok?”), offering a compromise, offering a reason for not complying, and acting as if no directive or request had been made (Kuczynski and Kochanska, 1990; Leonard, 1993). The phenomenon of interest then becomes the particular areas on which negotiation or divergence in values are more or less acceptable, and the particular ways in which differences are accepted, negotiated, or encouraged (Goodnow, 1997). The second and third years of life appear to be pivotal for the child's emerging capacities and inclination to be cooperative and considerate toward others. Toddlers are developing the cognitive skills to understand parental standards and apply them to their own behavior and achieving capacities for self-regulation that enable them increasingly to comply with internalized standards of conduct (Kopp, 1982, 1987; Kopp and Wyer, 1994). They are also becoming increasingly aware of the feelings and perspectives of others, which provides a resource for empathic responding to another in distress (Zahn-Waxler and Radke-Yarrow, 1990; Zahn-Waxler et al., 1992).

At the same time, the parent-child relationship is changing, as the child's growing assertiveness and the parents' growing use of prohibitions and sanctions lead to what can sometimes seem like endless conflicts of will (Biringen et al., 1995; Campos et al., 1992b). Parents now use emotional signals to convey approval or disapproval, sometimes before the obviously contemplated act of misbehavior even occurs (Emde and Buchsbaum, 1990; Emde et al., 1987). All young children internalize messages from these interactions; what is of interest is what they internalize.

These experiences during the toddler years provide the first opportunities for the child to answer the question, “What must I do to maintain good relations with others?” In the years that follow, the child's understanding of how to manage conflict becomes elaborated as the behavioral expectations of parents expand to incorporate consideration for others, basic skills at self-care, safety concerns, and compliance with family routines and manners (Gralinski and Kopp, 1993). The strategies used by parents to elicit cooperation also change to build on the child's maturing capacities for self-regulation. Specifically, they begin to rely more on explanations, bargaining, indirect guidance, and other nonassertive strategies (Belsky et al., 1996c; Crockenberg and Litman, 1990; Kuczynski et al., 1987). At the same time, however, children are also asserting their own independent judgment, making the preschool years ones of greater cooperation and greater conflict between parents and their offspring (Kuczynski and Kochanska, 1990; Kuczynski et al., 1987). Young children tend to comply more with behavioral standards as they reach the preschool years, but they also show a greater tendency to refuse before they comply and to negotiate, compromise, and display other indicators of self-assertion (Gralinski and Kopp, 1993; Vaughn et al., 1984).

Complicating this process is the fact that young children want to feel that they are in control of their lives. Long before babies understand that they are the ones making things happen, the controllability and predictability of stimulation affects their attention, emotions, and behavioral reactions (Sullivan et al., 1991). In studying face-to-face interactions between young infants and their caregivers, for example, researchers have noted that after a period of back and forth smiles and vocalization that often build in intensity, babies will look away. Skilled caregivers react by remaining quiet for a moment. The baby then looks back and the two begin to interact again. Unskilled caregivers or ones who are depressed sometimes ignore the cue or try even harder to get the baby's attention when he looks away. This often makes the baby fussy and irritable and increases the time he looks away. Overall, in pairs in which the baby controls the action (by looking toward and away), the caregiver keeps the infant's attention longer and elicits more smiles, coos, and active infant participation. Social interaction with a baby, however, is somewhat of a one-way street. Let the adult be the one to turn away and ignore the baby (e.g., to answer the phone), and the baby often doesn't wait patiently for the adult to look back; he gets more demanding.

By 1 year of age, being able to control the action can actually alleviate fearful responses to potentially frightening events. In one study discussed earlier (Gunnar, 1980), 12-month-olds were presented with a toy monkey that clangs symbols and flashes its eyes and can be quite frightening to children this young. The infants who were able to turn the toy on for a few seconds at a time did so repeatedly and often smiled and laughed. They liked it. In contrast, the children who could not control its actions were often upset, cried, and tried to get away from it. For older children, issues of control have been studied in the context of more subtle situations in which, for example, adults offer rewards if children engage in certain activities or are highly directive and intrusive while children are at work on a task (Fagot, 1973; Hamilton and Gordon, 1978; Lepper et al., 1973). These circumstances presumably undermine children's sense of autonomy and feelings that they are engaged in an activity because they want to do it. In fact, following these manipulations, children's levels of interest and persistence decline significantly. These situations are not unlike those in which a parent insists that a child clean up his room before he can play outside or finish her dinner before she can have dessert. The challenge for parents is one of encouraging cooperation while also fostering feelings of control and self-determination that lead the child to cooperate because he or she wants to. This also entails learning when to push for cooperation and when not to; when to go along with the child's requests and when to say “no.” As we saw earlier (Parpal and Maccoby, 1985), children are more likely to cooperate with parents who have on a previous occasion shown that they are willing to follow a child's interests.

Beginning in early childhood, as these examples illustrate, cooperation is not primarily a matter of whether parents consistently and firmly enforce their intentions on offspring, but is rather an interactional process in which a child's capacities to understand, agree with, and be motivated to comply by a positive parent-child relationship are also important (Grusec and Goodnow, 1994; Kuczynski et al., 1997). Interactions that, at one extreme, become highly coercive and engage parents and children in escalating battles of will can contribute to the mix of factors that place children on a path toward dysfunctional social behavior (Dodge, 1990; Patterson et al., 1992). Alternatively, when these interactions are characterized by clear and consistently enforced limits, low levels of emotional arousal, ample affection, and a deemphasis on the use of power, threats, and criticism (Campbell, 1997; Herrera and Dunn, 1997; Lepper, 1981; Maccoby, 1992; Zahn-Waxler et al., 1979), children learn to observe and ultimately internalize their parents' standards of conduct.

Caregivers who are warm and provide clear expectations for child behavior that are consistently enforced also encourage early conscience development (Eisenberg and Murphy, 1995; Kochanska, 1991, 1993, 1995). At times, this can involve directly focusing the child's attention on the consequences of misbehavior (especially when those consequences involve harm to others) or their responsibility for harm, explaining why certain actions are inappropriate or harmful, or drawing attention to the needs of another person whom the child can assist (Hoffman, 1983, 1988; Zahn-Waxler and Kochanska, 1990; Zahn-Waxler et al., 1979). The benefits of these activities for the child are enhanced when parents themselves model morally responsible behavior and respond prosocially to others. In short, when parents are clear about their expectations (e.g., “you need to put your action figures back in the box before we eat lunch”) and direct but noncoercive in enforcing them (e.g., “we'll eat lunch as soon as you are done”), rather than threatening (e.g., “if you don't hurry up, you won't see those action figures for a week”) in the context of a warm relationship with the child, early moral understanding is most likely to thrive (Eisenberg and Murphy, 1995).

Early conscience also grows significantly in contexts other than direct conflict over misbehavior. When parents and offspring converse about the day's events, for example, moral lessons are often implicit in what the adult conveys and what the child learns from their conversation (Dunn, 1987, 1988; Dunn et al., 1995). In these situations, moreover, children can reflect on what they hear—whether the conversation concerns the reasons for a sibling's outburst, the parent's response to being wronged, or a recounting of the child's own previous misbehavior—without the heightened emotion that may make it difficult for a young child to learn the same lessons in the context of a discipline encounter. Moreover, everyday family life is characterized by routines that enlist the young child's cooperation in rituals like bedtime, storyreading, waking, mealtimes, bathing, and other recurrent, predictable events (Fiese et al., 1999; Miller and Goodnow, 1995). The presence of these routines is one way of making expectations known and of avoiding constant confrontations. Children thus learn cooperation not only in the context of conflictual encounters and occasions for mutual give and take, but also in the predictable flow of daily life.

Interactions with siblings as well as parents are also important catalysts to early moral understanding, especially in relation to disputes over rights, possessions, and territory (Dunn and Munn, 1987; Slomkowski and Dunn, 1992; Tesla and Dunn, 1992). And at times, parents foster early conscience development when they devise alternative control strategies, such as avoiding a discipline encounter by proactively structuring circumstances or providing anticipatory guidance, each of which succeed in enlisting the child's cooperation in a nonconfrontational manner (Belsky et al., 1996c; Holden, 1983). Although these influences have been studied almost exclusively in the context of parent-child relationships (especially mother-child interactions), there is reason to believe that they are also important in the child's relationships with other caregivers, including fathers, grandparents, child care providers, and teachers.

The ways that caregivers can best support early conscience development also depend on the young child's temperamental characteristics (Kochanska, 1991, 1993, 1995). Depending on the extent to which a child is dispositionally more inhibited and fearful, for example, the parent's disciplinary efforts may either provoke cooperation or distressed withdrawal. Relatively gentle discipline characterized by suggestions and reasoning appears to be especially important for these children, for whom power assertive techniques are neither necessary nor effective. It is important for caregivers to calibrate their response to misbehavior according to the child's personality attributes, as well as the child's tolerance for stress and capacities for understanding. Regardless of temperament, however, developmental researchers have found that a secure, positive relationship with the parent is the best predictor of early moral growth. In a sense, a relationship of warmth and mutual responsiveness provides a context in which the parent's values and standards are most likely to be believed, accepted, and adopted by the young child (Kochanska, 1991, 1993, 1995).

Encouraging Exploration and Learning

Our prior discussion of children's emerging capacities for communication and learning (Chapter 6) documented the many ways in which parents support young children's linguistic and cognitive development. Much of early learning, in short, requires environmental supports, and children are dependent on their parents for providing them.

Starting in infancy, researchers have sought to identify the facets of parenting that are associated with higher scores on various tests of developmental status and cognitive abilities. The contingency and sensitivity with which parents respond to their baby's cues emerge consistently as important correlates of early cognitive outcomes (Beckwith and Cohen, 1989; Beckwith and Parmelee, 1986; Donovan and Leavitt, 1978; Landry et al., 1997; Vietze and Anderson, 1980). Infants whose parents can interpret, adjust their own behavior, and respond appropriately to their bids for attention, moods and states, expressions of interest, and efforts to communicate their needs are more advanced on virtually all assessments of developmental and cognitive status. Sensitive give and take between parent and infant appears to get children off to a good start on early markers of cognitive growth, just as it facilitates secure attachments. Other aspects of parenting that have shown positive associations with these outcomes include encouragement of exploration (in contrast to highly restrictive parenting), provision of a rich verbal environment, and ample amounts of nurturance and warmth (Clarke-Stewart et al., 1979; Cowan et al., 1991; Olson et al., 1986; Pettit et al., 1997). These features point to parents' role in creating an environment that is playful and nurturing, is rich in conversation, strikes a balance between safety and freedom to explore, and, in general, builds a belief in the child that the world is a receptive and responsive place.

A related literature has focused more directly on the interplay between the child's emerging capacities and the parents' ability to structure learning opportunities to both bolster and challenge these capacities. Originally proposed by the Russian psychologist Vygotsky (e.g., Vygotsky, 1978), children's learning is assumed to proceed optimally when parents (and teachers) present material in the child's “zone of proximal development.” This zone is defined in terms of tasks that are difficult for the child to perform independently but have components that can be accomplished with assistance. These kinds of processes, which have been portrayed as scaffolding (Wood, 1986; Wood et al., 1976), emphasize the ways in which parents orchestrate children's experiences to provide them with the most effective levels of support. Although most of the studies in this area have been concerned with cognitive development, parents have been observed to engage in the same kinds of supportive activities as they facilitate their children's entry into peer groups, with demonstrated benefits for the child's later social skills (Finnie and Russell, 1988). They undoubtedly apply, as well, to other situations in which parents attempt to manage or shape children's experiences—from making play dates to arranging child care—so that they remain within the child's tolerances for stimulation and challenge, while also fostering new capabilities (see, for example, Parke and Buriel, 1998).

These processes have also been examined in the cross-cultural literature on the teaching and learning roles of children and parents. This research has directed attention to the culturally organized ways in which adults involve children in routine activities and interactions, supportively structure their activities, and gradually transfer responsibility for specific tasks as the children acquire understanding and expertise (Goodnow, 1996; Ochs, 1992; Rogoff, 1990). In this sense, early learning is portrayed as a form of apprenticeship that is enacted in different ways in different cultures (Rogoff et al., 1991). For example, in some cultural communities, parents directly instruct children, play with them, and engage in conversations with them that are structured around materials and activities geared to the children's interests and abilities. In other communities, children are expected to learn through observation and participation in adult activities and through play with siblings and peers.

The cross-cultural literature has also called attention to the role that parents' expectations about the importance of various forms of achievement play in children's early learning—their familiarity with particular task strategies, their investments of effort in some tasks and not others, and their readiness to interpret various instructional or learning situations in particular ways (Goodnow, 1998). There are, for example, differences across cultures (e.g., China, Japan, and the United States) in the importance placed on reading compared with mathematics, and in the importance placed on being a well-rounded and happy child compared with being a good student (Stevenson and Lee, 1990). Subgroups within cultures—boys and girls, for example—also encounter different expectations, and children's own assessments of importance can influence what parents and others in their community view as important, as any nonsports-minded parent with a child who excels at baseball can attest.

As children reach the preschool years, researchers have turned their attention to the ways in which parents foster skills and abilities that are considered basic elements of school readiness, namely, literacy and number skills. For example, as mentioned earlier, maternal speech patterns predict vocabulary growth during the first three years of life (Hart and Risley, 1995; Huttenlocher et al., 1991), as well as prekindergarten measures of emergent literacy and print-related skills (De Temple and Snow, 1992). Parents encourage learning very explicitly through frequent visits to the library, routines that include regular reading to the child, and involvement in activities that allow children to play with notions of quantity. These behaviors show strong associations with early literacy and numeracy skills and later academic achievement (Ginsberg et al., 1998; Griffin and Morrison, 1997). Children generally benefit from parenting practices that expose them to high amounts of rich discourse and lots of print-related experiences (Beals et al., 1994; Gallimore and Goldenberg, 1993), as well as opportunities to experiment with numerical concepts (Saxe et al., 1987; Starkey and Klein, 1992). Of particular importance for the early acquisition of literacy and numeracy skills are the language and social interactions that surround such activities as storybook reading and board games that involve number concepts (Case and Griffin, 1990; Snow, 1993).

This work on parent-child interactions per se has been extended to encompass the next broader level of influence, namely the quality and quantity of stimulation and support that the overall home environment provides to a child. The home environment is most commonly assessed with the Home Observation for Measurement of the Environment (HOME) Inventory (Caldwell and Bradley, 1984), which assesses the materials, activities, and transactions that occur within the family setting and are supportive of early learning, defined largely in terms of IQ and traditional academic skills. Literally hundreds of studies have reported significant associations between HOME scores and children's IQ, cognitive and language development, and school performance (Bradley, 1985; Bradley et al., 1989; Clarke-Stewart, 1979; Gottfried, 1984; Linver et al., 1999; Smith et al., 1997; Wachs and Gruen, 1982). These relations hold for white, black, and Hispanic children from low and middle socioeconomic groups, although the patterns of relations may vary somewhat across ethnic groups (Bradley et al., 1989). Virtually every item on the HOME inventory distinguishes poor from nonpoor families both within and across white, black, and Hispanic families.

Efforts to assess the home learning environment raise the question of resources more broadly. Parents play an instrumental role in providing both social (i.e., opportunities for peer interactions, access to other supportive adults) and material (i.e., food, books, vacations, lessons) resources for their children. Economists and sociologists, in particular, have been interested in how the resources that families provide for their children influence their life trajectories (Becker and Tomes, 1986; Haveman and Wolfe, 1994). These family decisions are, of course, constrained by the decisions of government policy makers and employers, much more for some families than for others. Unfortunately, with the major exceptions of research on child care and family income (reviewed in the next two chapters), the influence on early development of the investments that parents make remains unexamined.

Parenting Practices and the Transmission of Cultural Values

Efforts to understand the importance of cultural practices in the rearing of young children, as described in Chapter 3, emphasize the extent to which culture is both reproduced and transformed within each child (Miller and Goodnow, 1995). That is to say, the socialization process that is embedded in the development of early relationships is influenced by the transmission of values and behaviors from one generation to the next, the “transformation” of those values and behaviors by the contemporary social context, and individual differences among caregivers (García Coll and Magnuson, 2000; Harkness and Super, 1992, 1996; Miller and Goodnow, 1995; Super and Harkness, 1997). These processes are of particular importance with respect to immigrant families (Portes, 1996; Rumbaut, 1994; Waters, 1997; Zhou and Bankston III, 1998). There is also a growing appreciation for the ways in which children themselves are not simply passive recipients of cultural influences, but rather active agents who bring both the ability and the willingness to accept, modify, or resist those influences.

Current research on differences in childrearing beliefs and practices is directing increased attention toward the scripts that characterize the daily routines of children and their primary caregivers (Farver, 1993; Farver and Wimbarti, 1995; Farver et al., 1995; Greenfield and Suzuki, 1998; Weisner, 1999). These routines are imbued with expectations that are designed to establish a moral order within which the child is expected to adapt (Schieffelin and Ochs, 1986; Shweder et al., 1995). They also create learning environments that vary dramatically across different cultural settings and groups (Rogoff, 1990). In this context, parents and other important caregivers introduce children to both informal routines and formal institutions that reinforce their cultural values and goals. When confronted by outside influences that they perceive to be undermining these efforts, caregivers can become highly threatened.

Parent belief systems and modes of parent-child interaction provide some of the most important ways in which culture is embedded in the process of child rearing during the early years of life (Levine, 1989; Super and Harkness, 1986). A wide range of cultural practices are salient in this regard, ranging from those related to sleeping, crying, and breast-feeding to those that affect the way parents talk with their children, the way emotion is acknowledged and expressed, and the way a child is expected to respond to praise for an individual achievement (see Chapter 3, Chapter 5, Chapter 6, and Chapter 7 for more detailed discussion). Inuit children are led, through adults' repeated teasing, to learn not to display anger (Briggs, 1992). Gusii mothers seldom gaze into the eyes of their infants, and their children are discouraged from looking adults in the eye (which is seen as an act of disrespect), yet they readily establish secure attachments (Levine, 1990). The widely varying views about sleep arrangements are discussed in Chapter 5.

The ways in which parents talk with children have been portrayed as one of culture's most powerful symbol systems (Harwood et al., 1995; Schieffelin and Ochs, 1986). As we discussed in Chapter 5 and Chapter 6, ways of storytelling have emerged as a prime site for exploring the socialization of values (Goodnow, 1997). Parents encourage some forms of storytelling and not others, sometimes to the point of labeling as “lies” stories that teachers portray as “imaginative” (Heath, 1983). In one study of black families, parents often recounted their everyday experiences in ways that presented themselves in a feisty light—a style that was repeated by their children in their own stories (Miller and Sperry, 1988). The impact, it is argued, extends beyond the kinds of stories that are transmitted from adult to child, to the child's emerging views of the self (e.g., If I'm feisty, that's good; if I'm reticent, that's bad), built up by way of the everyday events that are recounted and endorsed among family members (Miller et al., 1990). One very interesting issue that this raises is how children react when they are confronted with differing or even competing messages between parents or among the various important adults in their lives (Goodnow, 1997).

Individual differences in parents' previous and current interactions with the larger sociocultural context also affect how they raise their children (Parke and Buriel, 1998). The socialization of ethnic minority children, for example, has been conceptualized as a highly complex process that is influenced by the socioeconomic resources (i.e., housing, employment, health care, education, jobs) available to the family and the childrearing goals and adaptive strategies that the parents adopt in this context, among other factors (Harrison et al., 1984, 1990; Hughes and Chen, 1997, 1999). Extended family members, notably grandmothers, play a particularly important role in these processes, contributing, for example, material support, income, child care, and social regulation (Wilson, 1986, 1989).

This model highlights the need for greater attention to the efforts that parents make to help their children meet the challenges they will face as members of a minority group in a race-conscious society. In some cases, this leads parents to encourage their children to adapt to two cultural contexts—that of the child's primary ethnic group and that of the larger “mainstream” society (Boykin and Toms, 1985). This strategy is observed commonly in the lives of many ethnic minority groups in the United States (see, for example, LaFromboise et al., 1993).

Most empirical data on biculturalism and on the use of alternative strategies by ethnic minority parents to facilitate their children's adaptation have been collected from adult and adolescent samples (Parke and Buriel, 1998). It is therefore essential that this kind of research be conducted with samples of younger children, be extended to encompass nonminority parents, and be used to investigate distinctive strategies that are used by specific groups for raising children in an increasingly multicultural society.

An appreciation of the broad range of circumstances in which parents rear young children brings with it tremendous admiration for those who do it well. Conversely, conditions that pose significant challenges to the efforts of parents to get their children off to a good start in life warrant serious concern. We next turn to examine circumstances that seriously threaten and often undermine the parenting process.

Early Development and Maternal Depression

A variety of methods have been used to assess maternal depression, ranging from self-reports to clinical diagnoses. They typically gather information about the mother's mood as well as other symptoms of depression, such as sleep disturbances, difficulties with concentration, loss of motivation, and appetite changes (Campbell et al., 1995). Approximately 1 in 10 women with young children experience depression (Dickstein et al., 1998b; Gelfand et al., 1996), with prevalence rates often reaching two times these levels among mothers living in poverty.¹ Descriptive data from several recent studies of welfare samples have identified rates of moderate to severe depression in the 13 to 28 percent range (Danziger et al., in press; Lennon et al., 1998; Moore et al., 1995; Olson and Pavetti, 1996). These high prevalence rates are cause for concern about the effects of maternal depression on young children. This focus for this discussion, however, is not meant to minimize the need for societal attention to other forms of mental illness that can disrupt parenting (e.g., anxiety disorders, bipolar disorders, alcoholism).

Compared with children of nondepressed mothers, those with depressed mothers show greater risk of developing socioemo-tional and behavior problems, which translate into difficulties in school, poor peer relationships, reduced ability for self-control, and aggression (Campbell et al., 1995, Cummings and Davies, 1994b, Dawson and Ashman, in press; Zeanah et al., 1997). Children of depressed parents are also at heightened risk of serious psychopathology (Cummings and Davies, 1994b; Downey and Coyne, 1990; Zeanah et al., 1997). For example, children of clinically depressed parents are several times more likely to develop major depression then children of parents without such symptoms (Downey and Coyne, 1990).

Depression is not a static state. Mothers with major depressive disorders have periods when their symptoms abate, along with periods of greater suffering. Examining changes in child functioning in conjunction with changes in maternal functioning can help to disentangle genetic and experiential contributions to child outcomes. Work of this sort is being conducted in some cases using physiological measures.

In adults, depression is associated with changes in neural activity measured over the frontal areas of the brain that control emotion regulation (Davidson, 1994), as well as with altered day-night patterns that are evident in disordered sleeping, eating, and cortisol production. When EEG measures are taken, nondepressed adults typically show evidence of greater activity over the left compared with the right frontal region. But negative emotions evoked using movie clips produce a shift in this asymmetry, resulting in greater right than left frontal activity (Davidson et al., 1990). Depressed adults, in contrast, routinely exhibit greater right than left frontal activity, thus resembling the pattern evoked by negative emotionality in healthy adults.

Frontal asymmetry has been studied in infants and toddlers as well as adults (Dawson et al., 1992). Young children, like adults, show increased activity in the right frontal region relative to the left frontal region when negative emotions are evoked (Fox and Davidson, 1987). However, children of depressed mothers have been observed to display this asymmetry even when they are at rest or engaged in an activity with someone other than their mother (Dawson et al., 1992). Furthermore, the magnitude of this asymmetry was related to the timing of the mother's depression. Frontal EEG asymmetry in 3-year-olds was more highly related to maternal depression in the child's second and third years of life than to maternal depression in the child's first year of life. Interestingly, among three-year-olds, cortisol levels in the children were more highly related to maternal depression in the children's first year than to maternal depression in the year prior to testing. Thus, the neurobiology and neuroendocrinology associated with adult depression is observed in young children of depressed mothers in ways that may be somewhat specific to when, during the young child's life, the mother (and presumably the child) suffered most severely from the mother's depressive disorder (Dawson and Ashman, in press; Dawson et al., 1994; Zeanah et al., 1997).

This evidence has led some to suggest that infants and toddlers who are acutely dependent on their mothers, whose frontal lobes are experiencing rapid growth, and whose attachment, social-emotional, and regulatory capacities are developing, are particularly vulnerable to the negative effects of maternal depression (Dawson and Ashman, in press; Dawson et al., 1994; Goodman and Gotlib, 1999; Murray and Cooper, 1997; NICHD Early Child Care Research Network, 1998a; Weinberg and Tronick, 1998; Zeanah et al., 1997). While the severity and chronicity of maternal depression are clearly predictive of disturbances in child development (Campbell et al., 1995; Cummings and Davies, 1999; Frankel and Harmon, 1996; NICHD Early Child Care Research Network, 1998a), there is no definitive evidence regarding how the timing of maternal depression in the course of a child's life affects the child (Cummings and Davies, 1994b; Gelfand and Teti, 1990; Seifer et al., 1996). The neurobiological data suggest, however, that this is an area in which careful study of the timing question is warranted.

Efforts to understand the processes that underlie the developmental consequences of maternal depression have, not surprisingly, focused on mother-child interactions. Many depressed mothers show disrupted patterns of interaction with their infants. They also express self-doubts about their ability to parent well and are more likely than nondepressed mothers to perceive their children as being difficult (Teti et al., 1996b). Maternal depression affects both the emotional availability of the mother and the emotional tenor of her interactions with her child. Specifically, depressed mothers are more likely either to withdraw from their children and respond with little emotion or energy, or to become intrusive and hostile toward them (Frankel and Harmon, 1996; Tronick and Weinberg, 1997; Zeanah et al., 1997). These behaviors undoubtedly contribute to the higher rates of insecure attachment, as well as the withdrawal, reduced activity, and dysphoria that are observed in infants of depressed mothers (Cummings and Davies, 1994b, 1999; Dawson et al., 1992; Frankel and Harmon, 1996; Murray and Cooper, 1997; Seifer et al., 1996; van IJzendoorn et al., 1992).

These disrupted interactions and, more generally, the adverse effects of maternal depression are not seen uniformly. Many depressed women are very good mothers who raise children who are securely attached, do well in school, and do not misbehave (Cummings and Davies, 1994b, 1999; Frankel and Harmon, 1996). This raises important questions about the conditions that either prevent the damaging processes from occurring or protect children from their effects. Parenting by depressed mothers tends to be disrupted primarily when it occurs in conjunction with other sources of stress or adversity. Accordingly, a child of a depressed mother who also experiences poverty, marital discord, or maltreatment, or whose mother is also abusing substances or is an adolescent, is much more likely to exhibit some form of compromised development than is the child of a mother whose depression occurs in the context of an otherwise supportive environment (Cummings and Davies, 1994b; Seifer et al., 1996; Zeanah et al., 1997). Unfortunately, depression often goes hand in hand with poverty, substance abuse, and other factors that place early development at risk (Campbell et al., 1995; Seifer, 1995; Zeanah et al., 1997).

The role of marital discord and of fathers is especially noteworthy in this regard. When maternal depression occurs in a family experiencing marital harmony, mothers are better able to sustain healthy interactions with their children and the children are less likely to display adverse consequences (Cummings and Davies, 1999; Teti, 1999; Teti et al., 1996b). In fact, the occurrence of marital discord in a child's family may predict certain developmental problems more accurately than maternal depression. Relatedly, involved, psychologically healthy, and supportive fathers can buffer children from the detrimental effects of maternal depression, whereas absent or psychologically unhealthy fathers can amplify the effects (Goodman and Gotlib, 1999).

Beyond its direct effects on children, maternal depression can be a major barrier to the effectiveness of early interventions. The high rates of depression among low-income mothers combined with emerging evidence that depression can be a major deterrent to enrollment and full participation in intervention programs, such as home visiting, highlights the critical importance of this relatively hidden issue for those who design, implement, and evaluate early childhood programs (Teti, 1999). Maternal depression can also undermine the intended benefits of early intervention, as illustrated by the New Chance Demonstration. Mothers who participated in this comprehensive program for poorly educated teenage mothers on welfare not only felt more stressed than mothers who did not participate in the program, but the program actually had negative effects on the children of the depressed participants (Quint et al., 1997). It appears that New Chance overwhelmed the capacity of depressed mothers to cope with their situations, with detrimental consequences for their children (Lennon et al., 1998). Early intervention is clearly a complex undertaking for depressed mothers who are also experiencing other sources of stress and for whom mental health services may be a more appropriate first step (see Teti, 1999).

Children Exposed to Abuse and Neglect

Child abuse and neglect can be devastating for children's development (Cicchetti and Carlson, 1989; Cicchetti and Toth, 2000; Goodman et al., 1998; Kolko, 1996; National Research Council, 1993). In 1996, nearly 1 million children were involved in substantiated reports to child protective services agencies (National Center on Child Abuse and Neglect, 1997), and, based on reports from just 21 states (National Committee to Prevent Child Abuse, 1997), over 64,000 children were removed from their homes and placed in alternate care. Extrapolating from data from 40 states that provided figures by age of the child, over one-third of these victims were 5 years old or younger and, of the child victims killed in 1995, 77 percent were under age 3 (National Center on Child Abuse and Neglect, 1997). These official figures do not include community violence or the incidence of children who witness domestic violence, which also have pervasive detrimental impacts on young children (Osofsky, 1999). Moreover, many children are assumed to suffer multiple forms of maltreatment (Thompson and Wyatt, 1999).

Child maltreatment is associated with both short- and long-term adverse consequences for children (see reviews by Kolko, 1996; Malinosky-Rummell and Hansen, 1993; Pianta et al., 1989; Thompson and Wyatt, 1999). Physical abuse extracts a substantial toll on young children's social adjustment, as seen in elevated rates of aggression that are apparent even in toddlers (George and Main, 1979) and appear to derive, in part, from biases in social information processing that lead victims of physical harm to overattribute hostile intent to others (Dodge et al., 1990). On average, children who have experienced physical abuse also have lower social competence, show less empathy for others, have difficulty recognizing others' emotions, and are more likely to be insecurely attached to their parents. Deficits have also been noted in IQ scores, language ability, and school performance, even when the effects of social class are controlled.

These adverse effects are not short-lived. Although the vast majority of abused children do not become abusive adults, abused children are overrepresented among adults involved in both nonfamilial and familial violence (Malinosky-Rummell and Hansen, 1993). Among females, long-term effects manifest themselves as depression, anxiety disorders, and suicidal and self-injurious behaviors. Both men and women who were maltreated as children are at heightened risk for multiple forms of psychopathology (Cicchetti and Lynch, 1995; Kaufman, 1996; National Research Council, 1993). There is less research on physical or emotional neglect, although similar patterns across the same spectrum of outcomes have been reported (see Erickson and Egeland, 1996; Gaudin, 1993; Hoffman-Plotkin and Twentyman, 1984; Maxfield and Widom, 1996, for reviews).

As with maternal depression, abuse that occurs in the context of other adverse circumstances, such as multiple out-of-home placements, additional life stressors, and parental depression, reaps worse consequences. Coping and adaptation in the face of abuse are more likely when abuse is relatively isolated from other sources of adversity and, in particular, when the child receives emotional support from another important adult in his or her life (Garmezy, 1983; Rutter, 2000; Werner, 2000). As we discuss in the section below on institutionalization, the presence of a consistently available and emotionally invested adult appears to have a remarkably restorative influence on children who might otherwise be expected to succumb to the deleterious circumstances of their early lives.

Recent work has explored the presence of posttraumatic stress symptoms in maltreated children, with a special concern for potential alterations in fear-stress physiology that have been found to accompany reactions to trauma among adults and older children. Work is currently ongoing with children 3 years and under to discover how trauma manifests itself in preverbal children (Scheeringa and Zeanah, 1995). By 4 and 5 years of age, however, trauma symptoms typically assessed in older children and adults can be observed (Pynoos and Eth, 1985; Scheeringa et al., 1995). Many children who show these symptoms following traumatic experiences appear to recover when their circumstances improve. In others, however, there is evidence of fairly long-term alterations in the physiology of the fear-stress system, seen in higher levels and atypical daily patterns of cortisol and adrenaline production, that correspond to the duration of maltreatment (De Bellis et al., 1999b; Hart et al., 1996; Kaufman, 1996; Kaufman and Charney, 1999; Pynoos et al., 1996a, 1996b). More severe physiological changes are noted when children suffered for longer periods before rescue (De Bellis and Putnam, 1994; De Bellis et al., 1999b).

Do these changes in the physiology of fear-stress in children have developmental consequences? Certainly chronic abuse in childhood is associated with problems in emotion and behavioral regulation (Pynoos et al., 1995). In addition, maltreatment in childhood is a risk factor for multiple forms of psychopathology that are often seen to co-occur with post trauma symptoms (Cicchetti and Lynch, 1995; Kaufman, 1996; National Research Council, 1993). Many suspect that chronic activation of the physiology of stress during periods of rapid brain development may be producing pathology because of the effects of stress physiology on brain development (e.g., Cicchetti, 1994). However, there has been only one peer-reviewed scientific study that imaged the brains of maltreated children.

The 44 children in this study had all been sexually abused, typically beginning between ages 2 and 6, and most had also been physically abused beginning between ages 1 and 3 and had witnessed violence in the home. The duration of abuse varied but averaged around 3 or more years. These children, who were all of school age at the time of testing and had been living in stable, presumably nonabusive circumstances for several years, were all selected to meet clinical criteria for posttraumatic stress disorder, and many also met criteria for depression and other clinical syndromes (De Bellis et al., 1999a). The imaging data showed that, compared with physically and mentally healthy children matched for age and sex, these children had smaller brain volumes, larger lateral ventricles (i.e., the fluid-filled cavities of the brain), and smaller areas of connection (i.e., the corpus callosum) between the left and right sides of the brain. Most importantly, these findings were correlated with the duration of trauma (although not with age of onset), with children who had been abused longer showing greater differences from their matched controls than children who were rescued after shorter periods of abuse.

These results are preliminary and require replication, but they suggest that a history of chronic and severe trauma in early childhood can be associated with alterations in fear-stress physiology and in brain development. At the same time, there is no reason to interpret these results as indicating permanent impairment. Indeed, there is no evidence on whether or how therapeutic interventions may affect the fear-stress system or the neurological development of children who suffer maltreatment early in life. There is, however, compelling evidence from research on children reared in orphanages and then later adopted into loving families of the remarkable capacity of the young child to recover from aberrant early care.

Orphanage Rearing and Later Adoption

Children growing up in institutions have been the focus of a long-standing literature on early privation (Rutter, 1981a; Skeels, 1966). Studies of orphanage-reared children are now focusing on the wave of Romanian children adopted into families during the early 1990s (Benoit et al., 1996; Groze and Ileana, 1996; Morison et al., 1995; Rutter and the English and Romanian Adoptees [ERA] Study Team, 1998). This literature tells a compelling story about the severe developmental consequences of institutional care that affords neither stimulation nor consistent relationships with caregivers, which often confronts children with other physical adversities, including malnutrition, exposure to pathogens, and untreated chronic illness. It also reveals the remarkable recovery that characterizes many children exposed to these environments once they are adopted into loving homes, as well as the long-term impairments that continue to plague some of them many years after their life circumstances have improved. On both accounts, the fundamental message concerns the vital importance of consistent and committed adults in young children's lives.

Orphanage-reared infants and toddlers who have received adequate medical care and nutrition, but virtually no social or cognitive stimulation and few opportunities to establish a relationship with a consistent caregiver, show striking delays in motor and cognitive growth over the period of institutionalization (Provence and Lipton, 1962). They become extremely unresponsive, showing minimal crying, cooing, babbling, or motor activity. When tested in the orphanage, the deficits increase over time (Ames, 1997; Dennis, 1973; Provence and Lipton, 1962; Rutter and the English and Romanian Adoptees [ERA] Study Team, 1998).

When adequate physical care and developmentally appropriate stimulation are provided and only the availability of stable caregivers is lacking, development is substantially less delayed (Hodges and Tizard, 1989a, 1989b; Tizard and Hodges, 1978; Tizard and Joseph, 1970; Tizard and Rees, 1974). This suggests that stable relationships, as distinct from social-cognitive stimulation, are not required to ensure adequate physical, sensorimotor, cognitive, and language development. Children who have lacked stable and consistent caregiving, however, are not free of problems. Using parent and teacher reports, these children have been found in several studies to show impairments in regulatory aspects of thinking that involve concentration, attention regulation, and inhibitory control, generalizing problem solutions, and excessive concreteness of thought (Hodges and Tizard, 1989b; Tizard and Hodges, 1978, as reviewed in Gunnar, in press).

Importantly, removing children from institutions and placing them in stable families with adequate resources can produce remarkable catch-up growth on developmental milestones and in general cognitive (i.e., measures of IQ) and language development (Ames, 1997; Rutter, 1972, 1981b; Rutter and the English and Romanian Adoptees [ERA] Study Team, 1998). Even children delayed a year or more in behavioral and physical development can achieve normal levels of functioning once they are given the opportunity to live with a loving family.

Nonetheless, a persistent minority of institutionalized children across all studies and samples fails to show this dramatic recovery. They continue to exhibit multiple, debilitating problems in cognition and behavior years after entering their new families. Multiple, often unknown factors are likely to constrain developmental outcomes for this persistent minority. Case by case, these factors may include varying mixtures of genetic, prenatal, and postnatal conditions. If institutional rearing is involved, the continuing deficits found in some children should show a consistent dose-response relationship with the duration of privation experienced. Duration of orphanage exposure is highly confounded with illness, maltreatment, repeated changes in caregiving, and so on, making it exceedingly difficult to isolate duration as the causal factor. However, recent findings based on the Romanian children offer a reasonable test of this hypothesis (O'Connor et al., 2000; Rutter and the English and Romanian Adoptees [ERA] Study Team, 1998). In fact, a strong dose-response relationship was found (with a 24-point IQ difference between the extremes), in which children who spent more time in the institutions showed more persistently poor outcomes. The study controlled for a series of alternative explanations, including the possibilities that later-adopted children might have had greater impairments at birth and that the differing outcomes may have been due to the children's degree of malnutrition, which adds to the credibility of the conclusion.

The outcomes of institutionalized children may also be constrained by the capacity of the adoptive family to provide for the special needs of the child and the availability of appropriate remedial treatment. Surprisingly little is known about post-adoption effects on the recovery of children from such neglectful early environments. As with any family of a child with special needs, however, the challenge for such families may tax their emotional and financial resources. This may be one reason why the recent work on Romanian adoptees has shown that persistent problems were more common for children when two or more were adopted simultaneously from the orphanage by the same family (Ames, 1997).

In contrast to the rapid recovery from gross cognitive and language deficits that is often seen in formerly institutionalized children, social and emotional development appears to be more compromised or more susceptible to long-term impacts (Ames, 1997; Hoksbergen, 1981; Rutter and the English and Romanian Adoptees [ERA] Study Team, 1998; Tizard and Rees, 1974; Verhulst et al., 1990, 1992). Specifically, children from the most depriving conditions, which often included illness, malnutrition, frequent relocations, and thus disruptions in care, appear to be at increased risk for enduring behavioral problems and difficulties in peer relationships that have been seen well into adolescence (Ames, 1997; Fisher et al., 1997; Hodges and Tizard, 1989b; Tizard and Hodges, 1978; Verhulst et al., 1992). Once again, however, many of the children, including some with the most depriving and adverse early backgrounds, do remarkably well when taken out of those circumstances and given the opportunity to develop in stable, loving, and economically resourceful families.

The natural experiment provided by orphanage-reared children also affords the opportunity to ask if children deprived of stable, consistent attachment relationships early in life remain capable of forming such relationships when opportunities arise later in childhood (see Thompson, in press(a)). In contrast to expectations from early attachment theory, the window for forming attachments appears to remain open for a rather long period, at least through the early childhood years, and possibly later. Children seem to be capable of forming their first attachments to parents even when adopted or fostered later in development, although it is unknown how long this adaptive capability endures. Nevertheless, the quality of these children's attachment relationships varies considerably. A substantial minority have difficulties establishing secure attachments with adoptive or foster parents (Chisholm, 1998; Hodges and Tizard, 1989b; O'Connor et al., 1999). Some also display quite shallow relations with others and reduced likelihood of forming intimate relationships with peers later in development (Hodges and Tizard, 1989b). This implicates the importance of stable relationships during infancy for organizing competencies that support the development of attachments that entail strong emotional commitments to specific partners (Gunnar, in press).

A final look at whether the problems seen in institutionalized children derive from their institutionalized rearing or from other factors is provided by another study of Romanian children (Roy et al., 2000). These investigators compared children reared from infancy in residential group homes and those brought up in individual foster homes. Almost all of the children had biological parents with overt psychopathology. The rearing provided by the foster parents differed significantly and in positive ways from the rearing received by the institutionalized children. The findings showed that the institutional children had a much higher level of overactivity and inattention than the foster family children, suggesting that the institutional rearing was the more probable culprit. It is important to note that this finding applies to the children who were studied, namely, those who also came from very high-risk backgrounds.

In sum, the literature on orphanage rearing and later adoption illustrates the dual nature of early experiences. On one hand, highly disrupted or aberrant experiences of early care can have devastating effects on development. On the other hand, the impacts of even these extreme environments are modified by subsequent experiences. Indeed, the rebound shown by orphanage-reared children when they are adopted into loving homes highlights the vital role that caregiving in all of its facets plays in the lives of young children. The literature attests that early deprivation does not doom children to lives of compromised development. It also reveals, however, that not all children escape unharmed. Some children institutionalized as infants continue to show lingering, persistent, and even severe problems, and a substantial minority show traces of their conditions of early care in what appear to be problems in their capacity or willingness to form intimate, secure, and deep relationships with others. Improving care can certainly improve child outcomes, but given the massiveness of the change entailed in adoption (most families involved in international adoptions are middle to upper income, given the expense involved) and the lingering problems of some children, questions are raised about what can realistically be expected from far more modest attempts to alter the consequences for children of adverse early life experiences.