Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER 5 Middle Childhood! in the Context of the Family Eleanor E. Maccoby Between the time when children enter school and the time they reach adolescence, the family plays a crucial role in socialization, although its role is not so predominant as in the early childhood years. In middle childhood, teachers, peers, coaches, and others outside the family have more contact with the child than in early childhood, and they exercise varying degrees of influence. During this time, parents negotiate on behalf of the child with these other socialization agents, but their parenting functions are still exercised mainly through interaction with the child. Psychological research on parent-child interactions has heavily empha- sized infancy and the preschool period rather than later periods. In compiling a recent review of research on this topic, Maccoby and Martin ~ 1983) located more than three times as many studies on children under age 6 as on school- age children. Although the research on certain aspects of family-based so- cialization in the school-age period is thin, studies on family characteristics and their relationship to children's current or future deviance e.g., alco- holism, drug addiction, aggression, delinquency, depression present sub- stantial information on the school-age years Barrington and West, 1981; McCord, 1979; McCord and McCord, 1960; Pulkkinen, 1982; Robins, 1974; Robins and Ratctiff, 1980; Rutter, 1982~. This chapter reviews some of the major themes of existing work, identifies promising themes that appear in more recent writings, and suggests gaps that additional research might help to fill. 184
CON17EXT OF THE FAMILY 185 Most research on socialization within the family has been concerned with individual differences, explaining why children vary in their personal attri- butes. The major hypothesis has been that such variation stems at least in part from differences in parental socialization of children. Many studies have looked for dimensions in which parents differ and have then examined the relationship between these variations among parents and the characteristics of their children. The early work was essentially linear in concept, having its origins in a view of socialization wherein parents, by means of reinforcement and dis- cipline, trained their children to carry out certain behaviors and avoid others. In addition, some researchers studied the development of children's moti- vation to pattern themselves after their parents' values and behaviors via the process of identification. While the concept of identification gave chil- dren a somewhat more active role in their own socialization than did simple reinforcement theories, socialization was still conceptualized primarily as a flow of influence from parents to children, with children acquiring sets of behavioral tendencies in the form of habits or motives. The characteristics of parents and children that were studied varied from narrow and specific to broad and abstract. The child outcome measures ranged from highly specific responses (e.g., the frequency of smiling) to global characteristics, such as intelligence or competence. On the parental side, specific characteristics such as the frequency with which parents rewarded or punished a given behavior were studied; at the opposite ex- treme, such general characteristics as warmth or permissiveness were as- sessed. Although the work did not deny that individuals' behavior varies from one situation to another the focus was to identify individual charac- teristics of both parent and child that had some stability across time and situations and to look for functional relationships between the parental characteristics as antecedents and the child characteristics as outcomes. A major refinement of this point of view has been the study of behavioral patterns, in both parents and children. The assumption underlying the study of parental patterns or clusters has been that the effect of a parental practice depends on the context of other parenting characteristics with which it occurs (see Baumrind, 1967, 1971; Becker, 1964~. Investigation of child behavioral patterns occurs in studies of attachment, in which it is argued that children's attachment is adequately characterized not by counting the frequency of specific behaviors, but only by studying clusters of behaviors taken in context (Sroufe and Waters, 1977~. This point of view has been applied primarily to infants and toddlers, while the trait approach remains predominant in studies that examine how parental practices influence the characteristics of preschool children.
186 DEVELOPMENT DURING MIDDLE CHILDHOOD In recent years, a more interactive, less linear point of view has emerged. Researchers increasingly examine the effects of children on parents, the effects of parents on children, and cyclical processes. Little of this work, however, is developmental in concept. Researchers are only beginning to ask: How does parent-child interaction change with the developmental level of the child? Does a particular child-rearing style have a different impact on children of various ages ? How much does the impact of parental treatment of a school-age child depend on the relationship established with that child at earlier periods of development? How much are parents limited or facilitated in the relationship they can have and the child-rearing methods they can use with a school-age child by the characteristics that the child has developed during the first 6 years of life? The major contention of this chapter is that the process of child-rearing undergoes important changes as children develop, and that middle child- hood, with regard to parent-child relationships, has its own distinctive fea- tures. The chapter presents developmental changes that normally occur as children enter the middle childhood years, some of the concomitant changes in parents' child-rearing roles, and some of the more traditional socialization findings those concerned primarily with individual differences among school- age children and their parents. The chapter goes on to describe how parents differ from one another, then summarizes some of the major findings for this age period concerning the way parental variations relate to the variations among children in their personalities and social behavior. It takes up some of the differences among groups of parents (e.g., social class, ethnic, and family-structure groups) and considers some of the conditions that might bring these differences about. CHILD REARING AND DEVELOPMENTAL CHANGE Issues and Processes of Socialization Amount of Parent-Child Interaction As children enter the school-age years, there is a great decline in the amount of time they spend in their parents' presence and in the total amount of time their parents devote to them. Hill and Stafford (1980) reported in a time-use study that parents spend less than half as much time in caretaking, teaching, reading, talking, and playing with children ages 5-12 as they do with children of preschool age. The drop in interaction time is more pre- cipitous and occurs earlier with a lower parental education level. Other studies concur in finding this strong decline in interaction rates with the increasing age of the child (e.g., Baldwin, 1955~.
CONTEXT OF THE FAMILY 187 Parent'Child issues There are important changes, too, in the kind of issues parents deal with in their day-to-day interactions with their children. Interactions with pre- schoolers focus on modesty, bedtime routines, control of temper tantrums, fighting with siblings or other children, eating and table manners, getting dressed by themselves, and attention seeking (Newson and Newson, 1968; Sears et al., 1957~. Some of these issues carry over into the school-age years (e.g., fighting, children's reactions to discipline). New issues emerge by age 7 (Newson and Newson, 19761: whether to require chores and how to enforce standards of performance; whether children should be paid for household work; how they can be encouraged to entertain themselves rather than relying on parents for activity planning; how to support them in their re- lationships with peers and whether to monitor their friendships and dis- courage or encourage contact with specific children. An important parental issue in middle childhood is how to keep track of children's whereabouts and activities now that they are spending more time away from home. With varying degrees of success, parents teach their chil- dren to inform them of their whereabouts at all times; parents often require either that children come directly home from school to discuss what they propose to do or that they check in by phone. An issue entirely new to middle childhood is how parents deal with the child's problems at school e.g., a child's unwillingness to go to school or a child's report of an encounter with the teacher (Newson and Newson, 1977~. Parents also become concemed about how much to become involved in the child's schoolwork. Roberts and colleagues ( 1981), working with an American longitudinal sample of children studied at ages 3 and 12, reported increasing parental emphasis on children's achievement during this period, a trend related primarily to their performance in school. Although there is reason to believe that the issues arising between parents and children change significantly as children enter and progress through the middle childhood period, data on these changes are limited, and many issues have been studied only minimally. We know little, for example, about what moral or ethical matters come up in family exchanges or how they are dealt with. We should be aware that the above discussion of issues proceeds from a rather culture-bound perspective. Weisner's chapter in this volume shows that in non-Western societies, in which most of the worId's children live, the 6-12 age period is when children enter the work world, contributing to the necessary survival functions of families-e.g., care of younger chil- dren, agricultural work, care of animals. In such societies, the issues that preoccupy Western parents may have little relevance.
188 DEVELOPMENT DURING MIDDLE CHILDHOOD Techniques of Discipline Parents report that interactions with school-age children are in some respects easier than with preschoolers; for example, it may be possible to use reasoning rather than discipline. The children's increasing ability to discuss issues with their parents, however, is a mixed blessing; parents often weary of extended arguments and regret their children's increasing skill in catching their parents' inconsistencies. Studies consistently show a decline in the use of physical punishment as children get order (Clifford, 1959; Fawl, reported in Patterson, 1982; Newson and Newson, 1976~. Parents also decreasingly use distraction and physically moving the child away from forbidden or dangerous activities. At the same time, there is evidence of an increase, as children move from the preschool into the school-age years, in parents' using deprivation of privileges, appeals to the child's self-esteem or sense of humor, arousal of the child's sense of guilt, and reminders that children are responsible for what happens to them (Clifford, 1959; Roberts et al., 1981~. Threatening, ignoring, and isolating appear to decline from the time of school entrance to the later middle childhood years. Changes in Affectional Relationships If we assume that the issues and techniques of discipline change with children's development, as well as the nature of the parent-child relationship itself, we are on thin ground as far as research evidence is concerned. The work that is available does not reveal a consistent picture. Consider for example the role of affect in the relations between parents and children, taking positive affect first. Newson and Newson (1968) reported that, at age 4, the question of open displays of affection did not appear to be prob- lematic. In a majority of cases, both parents and children seemed to accept and value cuddling. By age 7 (Newson and Newson, 1976), many of the children were becoming circumspect, avoiding affectionate displays with their parents whenever a peer was present, although they still sought physical affection at private moments such as bedtime. From the parents' standpoint, mothers expressed continued readiness for physical contact with their chil- dren. The implication of this descriptive account is that it is primarily the children, rather than the parents, who pull away from physical affection and provide the impetus for whatever decline in such displays occurs with age. Shows of physical affection, of course, are only one aspect-and a fairly narrow one of the affectional relationship between parent and child. The
CONTEXT OF THE FAMILY 189 detailed work on attachment that is now available in infancy and the early childhood years has no counterpart in research on middle childhood. While a few writers (Bow~by, 1969; Marvin, 1977) have studied the transition that occurs at about age 3 from proximity-seeking to more distal forms of at- tachment, few have examined transitions that occur beyond this point. Some years ago, Baldwin (1946), working with families in the Fels lon- gitudinal study, compared the relationships between parents and 3-year-olds with those between parents and 9-year-olds. At the younger age, higher levels of parental warmth were recorded on all the relevant measures: child centeredness, approval, acceptance, affection, and rapport. Parents of 9- year-olds were less warm in all these ways and more severe and critical. Lasko (1954), working with the same data but reporting on more age levels and differentiating the sample by birth order, reported that the decline in parental warmth was found primarily for firstborn children. She indicates that later-looms have an affectional relationship with their parents that is more stable over time. A recent report from another longitudinal study presents a different pic- ture. Roberts et al. (1981) reported a decrease, between the ages of 3 and 12, in displays of physical affection, but there was little change in the reported mean levels of enjoying parenting, having positive regard for the child, and having respect for the child's opinions and preferences. Thus it appears that although parental warmth is shown in different ways with older children, it may not shift downward, as Baldwin claimed. A third study (a cross-sectional one by Armentrout and Burger, 1972) asked children in grades 4 through 8 to describe certain aspects of their parents' child-rearing practices. The parental acceptance-rejection balance changed with the ages studied. For boys, acceptance by parents increased from the fourth to fifth grade, then declined; for girls, the peak was reached in the sixth grade and declined thereafter. The trends in displays of negative affect between parents and children are somewhat clearer than the trends for displays of positive affect. Generally, anger between parents and children declines as children move into the school years. Goodenough (1931) reported that the frequency of angry outbursts by children declined after the age of 18-24 months. Patterson (1982) re- ported a steady decline, between ages 2 and 15, in the frequency of coercive behavior directed by children toward other family members (whining, yell- ing, hitting, ignoring others' overtures), and Newson and Newson showed a decrease in the frequency of temper tantrums between the ages of 4 and 7. Concomitantly, a steady decrease occurred, from age 3 to age 6 to age 9, in the frequency of disciplinary encounters (Clifford, 1959~; that is, in conflictual incidents in which the parent directed the child to do something
190 DEVELOPMENT DURING MIDDLE CHILDHOOD or to stop doing something and the child failed to comply. Such conflicts, especially when they culminate in physical punishment, often involve angry emotions for both parents and children. The finding that physical punish ment is much less frequent with school-age than preschool children indicates that parents and school-age children either have fewer conflicts or have reamed to deal with them without letting them escalate into highly emo- tional encounters. When school-age children do become angry, however, they do not recover from it as quickly as they did when younger (Clifford, 1959), and this means their parents must deal with the aftereffects of emo- tional outbursts sulkiness, depressed mood, avoidance of parents, or passive noncooperation. Information on the displays of positive and negative affect between parents and children gives us only a partial picture of their affective relationship. Attachment bonds within families presumably have more generalized man- ifestations and, with respect to such bonds, a number of important questions remain unanswered. What is the capacity of children in middle childhood to form strong bonds with new caretakers' What is the impact of disruption at this age of earlier-formed bonds' To what extent do the reactions of children at this age to disruption of bonds depend on the nature of inter- personal ties formed at earlier ages' When disruptions of bonds occur at this age, what are the implications for the child's functioning at subsequent age periods? ~ discuss these issues briefly below but must note that they are still open questions. Changes in Control Processes It is a usual assumption that a major accompaniment of children's de- velopmental change is the gradual shift of controlling functions from parent to child. In fact, this assumption is largely unverified. We have little in- formation concerning changes that occur in the degree and kind of parental control between the preschool and school-age periods. Baldwin's (1946) reports from the Fels longitudinal study in some ways run counter to a transfer-of-power trend. Comparing the behavior of a group of parents toward their 3-year-olds with the behavior of the same parents when their children were 9, Baldwin reported that the parents were more restrictive, more coercive, stricter, and even somewhat less democratic with the children when older. The parental attitude toward 3-year-olds was pre- dominantly indulgent and protective, while much more was expected of older children, who were thought to be capable of conforming to nearly adult norms of behavior. In a cross-sectional study with more age groups (each half-year or year from ages 6 to 10), Emmerich (1962) questioned parents concerning their
CONTEXT OF THE FAMILY 191 child-rearing. He extracted two dimensions: one running from nurturant (positive, facilitating reactions) to restrictive (negative, interfering reac- tions), and the other, a power scale, reflecting the amount of active control exerted by the parent, which included rewards as well as punishment. Neither factor showed any consistent increase or decrease over the 6-10 age period. The more recent study by Armentrout and Burger ( 1972) covered an older age range (10-14~. These children reported a considerable decline in the amount of psychological control (possessiveness, intrusiveness, arousal of guilt) exercised by their parents. On the firm-lax dimension, however, there was an increase in parental firmness up to the age of 12, then a decline. In a study of preadolescents and adolescents, Dombusch et al. (1983) analyzed the responses of youth to questions concerning how decisions were made regarding: how they should spend their money, how they should dress, how late they might stay out, and what peers they should associate with. Over the age range 12-17, an increasing proportion of youth reported that they make such decisions alone, and a decreasing proportion reported that their parents made them alone. The proportion reporting that such decisions were made jointly remained constant. The findings would be consistent with a shift in the locus of decision making from parent-alone, to joint, to youth- alone. Taken together, these studies suggest that, at least in some respects, the transfer of power from parent to child occurs somewhat more slowly than had been supposed, with the major shift to genuine child autonomy begin- ning to occur at about age 12. The information to support this suggestion is meager, however, and it is obvious that children ages 6-12 already are participating in the controlling and managing processes. This participation is a simple necessity that stems in part from the decrease in time that parent and child are together. ~ suggest that the original conception of parents transferring control directly to their children may be an oversimplified one- that a better conceptualization involves an intermediate process that may be called coregu~non. That is, before they relinquish control of a given aspect of their children's lives, parents continue to exercise general super- visory control, while children begin to exercise moment-to-moment self- regulation. This phase is paramount for many aspects of children's behavioral devel- opment during the 6-12 age period. The process of coregulation, if it is to be successful, must be a cooperative one, with clarity of communication between parent and child of paramount importance. The parental tasks during this period are threefold: First, they must monitor, guide, and support their children at a distance-that is, when the children are out of their presence; second, they must effectively use the times when direct contact does occur; and third, they must strengthen in their children the abilities
192 DEVELOPMENT DURING MIDDLE CHILDHOOD that will allow them to monitor their own behavior, to adopt acceptable standards of good and bad behavior, to avoid undue risks, and to know when they need parental support or guidance. Children must be willing to inform parents of their whereabouts, activities, and problems so that parents can mediate and guide when necessary; parents must keep informed about events occurring outside their presence and must coordinate agendas that link the daily activities of parent and child. Parents seem to have some, although not explicit, knowledge that different methods are needed for out-of~sight control than for face-to~face control of children. Grusec and Kuczynski ( 1980) found that most parents use a variety of methods in attempting to influence their children's behavior, and that the method chosen depends in part on whether infractions occur in or out of the parents' presence. For in-presence infractions (e.g., quarreling among siblings, excessively noisy behavior, or throwing a ball in the living room), parents tend to use power-assertive methods. For infractions not directly observed, however, (e.g., stealing money, teasing an old man, running into the street) parents are more likely to use reasoning and explanation. In an unpublished study, Kuczynski compared methods of parental influence when the parents either (1) knew they would be absent or (2) did not know they would be absent at the time of the child's compliance with parental instruc- tions.When the child was out of parental sight, parents were more likely to use inductive reasoning and character attribution e.g., "You are certainly nice and helpful." With boys, they also used less power assertion if they needed out-of-sight compliance. With the child's increasing age, we should expect an increase in parents' fostering out-of-sight compliance. These changes in child-rearing occur concurrently with a variety of norm- ative developmental changes that occur in all children during middle child- hood, albeit at somewhat different rates. We turn now to a consideration of these changes-what they are and how they might be linked to the ontogeny of the parent-child relationship. Because there is little research focusing on how changes in child-rearing are related to the developmental level of the child, this section is necessarily speculative. It will point to some gaps in our theorizing and knowledge and suggest some promising approaches. Normative Developmental Changes in Children Developmental changes occurring in early childhood, such as the shift from crawling to walking and running or the acquisition of language, are dramatic and universally acknowledged. The developmental advances that occur in middle childhood are less obvious but nonetheless important. A
CONTEXT OF THE FAMILY 193 review of the developmental changes in middle childhood follows (for a fuller exposition, see Maccoby, 1984~. Social Cognition and Social Competence Between ages 6 and ~ 2 children's ability to adopt the perspectives of others and to recognize other people's purposes and probable reactions substantially increases (Selman and Byme, 19741. Partially as a consequence of this improved understanding, school-age children also improve considerably in their referential communication skills (Krause and Glucksberg, 1969~. That is, they are more able to select and convey the crucial information necessary for a partner to understand the message that the child wants to convey. A second social-cognitive gain has to do with the child's increasing un- derstanding of social roles their requirements and how they intersect. The acquisition of sex-role knowledge has been more extensively studied than most other aspects of role concepts, but it is reasonable to believe that in middle childhood there is considerable expansion of children's understanding of other roles as well e.g., teacher and pupil, leader and follower. Some aspects of role reaming that affect peer interaction are discussed by Hartup (in this volume). Some changes in children's conceptions of parent-child roles may affect family interactions. Specifically, there is a shift in children's conceptions about authority and the basis for parents' rights to exercise it (Damon, 1977~. While preschoolers tend to think that parental authority rests on the power to punish or reward, the beginning of an exchange relationship can be seen with the onset of the middle childhood period. Children of this age begin to say that they ought to obey because of all the things their parents do for them. Some time after the age of 8, children also begin to give weight to their parents' expert knowledge and skill as a reasonable basis for parental authority. We may assume that parental appeals based on fairness, the return of favors, or reminders of the parents' greater experience and knowledge would be increasingly persuasive as the child progresses through this period, so that parents would less often fee} compelled to resort to promises of reward or threats of punishment. There are many other components of the social competence acquired in middle childhood. Some (e.g., entry skills that facilitate a child's joining peer group activities) have been studied primarily in the context of peer interaction. We know little about the role of other family members in children's acquisition of the social skills that they employ outside the family, although there is reason to believe that their behavior with siblings does generalize to peers (Dunn, 1983~. It would be useful to know more about
194 DEVELOPMENT DURING MIDDLE CHILDHOOD the transfer of social skills, not only from the family to other contexts, but also in the reverse direction; it is reasonable to believe that social skills acquired during peer interaction are sometimes brought to bear in a child's negotiations with family members. Self-Concepts About the time of entrance into school, children begin to acquire the ability to view the self from an outside (other-person) perspective. At about the same time or a little later, children also begin to define themselves more in terms of such attributes as appearance, possessions, or activities (see Markus and Nurius, in this volume). Perhaps in part as a consequence of these changes, children become more susceptible to attributional appeals. Grusec and Redler (1980) found, for example, that the generalized altruism of 5-year-olds was not enhanced when the experimenter responded to their generosity by an attribution such as "I can see that you are the kind of person who likes to help." Eight-year-olds, however, showed increased help- fuIness in an unrelated situation a week or two later as a result of such treatment. Although to our knowledge the point has not been tested, it seems likely that during the school-age years children would become more responsive to parental reminders that other people will not think well of them if they behave in certain ways. That children are beginning to see themselves as others see them, however, does not necessarily mean that they will become more tractable. When children realize that they can tailor their behavior and emotional expressiveness to what they think are the expectations and values of a given audience, they become more self-conscious and less open, even with their parents. Some parents (see Newson and Newson, 1976) complain that when their children have entered the school-age period, it is no longer so easy to know what their children are thinking and feeling; this renders the task of monitoring and guidance more difficult. Impulsivity ImpuIsivity declines fairly steadily from early childhood into the school- age years. As we have already noted, the frequency of angry outbursts de- clines, and children are more able to endure frustration and accept delays in gratification. Although not clearly demonstrated in research, it is probably also true that children improve in their ability to regulate their bodily activity according to the demands of the situation they are in, so that they exhibit less restless motion and wild running about as they grow older.
CONTEXT OF THE FAMILY 195 What are some of the implications of these changes for parents? Probably less parental power assertion is required. Halverson and Waldrop (1970) reported that mothers of impulsive preschool children used more negative, controlling statements than did the mothers of less impulsive children. (Of course, we do not know from this correlation whether the mothers' autocratic treatment caused the children's impulsiveness or vice versa. Perhaps both effects occurred. ~ An elegant set of studies by Barkley and colleagues (Barkley and Cunningham, 1979; Barkley, 1981) provides a clearer picture of the causal directions. These studies show that when the hyperactive behavior of diagnosed children is brought under control through drug therapy, ma- ternal directiveness decreases and supportive responses increase. Whalen et al. (1980, 1981) have shown similar decreases in controlling behavior by teachers following drug therapy with hyperactive boys. Informal experience in dealing with young children suggests that when children are angry, dis- tracted, or highly excited they cannot listen to reason; in order to get a message through to them at such times, parents must raise their voices, become peremptory, and perhaps resort to physical restraints. The decline in the use of physical punishment as the child gets older is well documented, and we may assume that one of the factors underlying this change is the decrease in the number of occasions on which the child has lost control. With children's increasing control over their emotionality and hyperac- tivity that they achieve with age, parents become progressively more able to engage their children in problem-so~ving dialogues that may require time and sustained attention from both parent and child. Clifford (1959) noted with children ages 3 and 6 disciplinary encounters were more common when the children were tired or hungry; by the age of 9, this correlation had disappeared. This finding suggests that older children are more able to main- tain self-control under the stresses of moderate hunger or fatigue; they do not become so restless or demanding that they require parental power- assertive controls. Another aspect of impulse control is children's increasing ability to focus their attention on task-relevant information and carry out exhaustive searches for needed information without being distracted by irrelevant information and events (see Lane and Pearson, 1982, for a review). There is some indication that when a parent is trying to get a child to do an onerous task and the child becomes distracted, the parent becomes more peremptory- uses shorter sentences and, at least with boys, more often resorts to giving direct commands (Chapman, 1979~. Perhaps older children's greater ability to sustain their attention during conversation with parents permits parents to use longer sentences and fewer imperatives than they use for younger children (Chapman, 1979~.
196 DEVELOPMENT DURING MIDDLE CHILDHOOD Cognitive Executive Processes Many developmental changes occur in cognitive processes as children enter and traverse the school-age period. These are reviewed elsewhere in this volume (see Fischer and Bullock). For purposes of this chapter, the changes that appear most relevant to parent-child interaction are those reflecting growth in the so-called executive functions (Steinberg and Powell, 1983~. When children are asked to solve a problem, there is a regular increase with age in the extent to which they adopt a plan or goal for their activity. They become more efficient in the use of their memory capacity, in part because many subroutines are automated (Brown and DeLoache, 1978~. Thus children can make more and better use of previously stored information. In addition, they become progressively more able to monitor their own knowledge and their progress toward goals; they can determine what infor- mation is necessary in order to carry out a plan. School-age children are increasingly able to coordinate goals and subgoals into action hierarchies- deciding which of a set of subroutines is to be run off first, which second, and so on. When an initial set has not been productive, they are increasingly able to break off the subroutine and adopt a new problem~solving strategy. The sheer accumulation of knowledge is important too, particularly knowI- edge concerning social scripts. That is, children store prototypic represen- tations of the sequences in which events usually occur and the roles that various actors play in these sequences. Clearly, these cognitive changes among school-age children must have profound implications for parent-child interaction. Children's increasing knowledge of social scripts and their ability to govern their own actions must mean that they need fewer cues from their parents to keep them on a planfu! track; furthermore, they should be more able to coordinate their activities with those of their parents. Consequently, parents can drop some of their monitoring activities. Such activities as bathing and brushing teeth will become fully automated and will require no monitoring. For other activities the child can assume some portion of the monitoring; however, how the checking processes are divided between parent and child is something we know little about. To our knowledge, no studies have focused on changes in the locus and content of monitoring with children's increasing age, and this is a serious gap. We can only assume that there is a transition stage, during which moment-to-moment monitoring by parents is replaced by a joint monitoring process, so that with older children parents intervene only at the higher modes of action hierarchies, while children self-monitor the subroutines. A
CONTEXT OF THE FAMILY 197 second serious gap in our knowledge is the degree to which parents and children share social scripts and what the consequences of discrepancies are. Vulnerability to Stressors There is no reason to believe that children in the mic32le childhood period are any more or any less vulnerable to stress than they are at any other age (Maccoby, 1983~. The particular stressors to which they are vulnerable, however, are different from those at earlier or later ages, and the nature and duration of the stress effects are probably also different, although evidence . . . . on t ills point IS lnCOnC. .USlVe. The complexity of this issue is shown in the research on attachment, separation, and bereavement. Much of the work has been done with infants, toddlers, and preschoolers, but some studies include older children and yield interesting comparisons. Although children ages 6-12 are clearly less disc tressed by relatively short~term separations from their parents than are younger children, they grieve more intensely and for a longer time over the death of a parent (see Rutter, 1983, for a review of evidence). It is not known, however, whether the long-term effects of bereavement and sepa- ration are greater or less when they occur during middle childhood than at other times (Garmezy, 1983:63; Rutter, 1983:18-19~. Part of the problem in making strong inferences about a child's degree of stress at the time of a separation or bereavement is that the effects of a child's age are usually confounded by the length of separation, the nature of alternative care during separation or bereavement, and the nature and duration of living conditions prior to the stressful event. Wallerstein and Kelley (1980) report that children's reactions to their parent's divorce varies with the children's age (see section on single parents below). In Wallerstein and Kelly's study, however, the numbers of children in each age group are low and there is no comparison group of children from nondivorced families; therefore, replication is needed. Important questions remain: the effects of separation and bereavement depend not only on how much distress children experience at the time these events occur, but also on how quickly the children form secure bonds with new caretakers. Do children in middle childhood not only grieve longer but also find it more difficult to form new relationships with a stepparent or adoptive parent? Do school-age children benefit more than younger children from maintaining a relationship with the noncustodial parent after divorce? Important though the issues of separation and bereavement are, other vuInerabilities to stress occur during middle childhood. School~age children
198 DEVELOPMENT DURING MIDDLE CHILDHOOD show greater sensitivity to insults or other attacks on their rights or com- petence; to disruptions in intimate friendships with peers; and to unfavorable comparisons with the academic or athletic achievements of other children. At the same time, there is decreased vulnerability to many stressors that plague younger children, such as having to wait for an expected treat or having to adapt to household demands and routines. Thus far in this section, developmental change has been presented as a relatively autonomous process one that is largely independent of the history of the parent-child relationship but that has an impact on the relationship. For such changes as physical growth, this is a reasonable picture. For other aspects of development, age-related changes are a joint product of organismic development and the cumulative effects of earlier experience, with prior parent-child interaction being a significant portion of that experience. How the two forces are weighted in bringing about developmental change is an important research issue in itself, one about which we cannot generalize at present. Clearly, however, the sequence of developmental events is more complex than would be implied by either a simple parent-shapes-child formulation or one that views developmental change as inherent in the child and changes in parental behavior as mainly a reflection of children's development. Chil- dren's developmental changes emerge, in part, from prior socialization efforts by their parents. Presumably, a child who is developing new skills in self- regulation will employ these skills quite differently, depending on whether a relationship of trust, open communication, and mutual willingness to accept influences from the partner was previously established between parent and child. And, presumably, some children develop self-regulatory skills faster than others, depending in part on their parents' prior effectiveness. To date, however, research on the family almost totally lacks work on how parental effectiveness in socializing school-age children is related to the earlier phases of interaction. What is needed is longitudinal research that looks for sequential phases in the parent-child relationship, each phase being a joint product of the relationship that existed earlier and the new capabilities of the child. The Role of Mutual Cognitions Another gap in our knowledge of parent-child relationships concerns the way parents and children label one another and the attributions they make concerning one another's motives (see Hess, 1981; Maccoby and Martin, 1983~. Presumably such cognitions play a larger role in middle childhood than they did when the child was younger. Parents and children do not
CONTEXT OF THE FAMILY 199 react to one another solely on the basis of the partner's immediately prior behavior. They react on the basis of how this behavior is interpreted and what they think the behavior signals in the way of a sequence of actions that will probably follow. As children grow older, parents and children accumulate experience with one another and develop expectations of the other's probable reactions. It appears inevitable that each must begin to label the other in terms of broad stereotypes and that these codified expectations must influence the inter' action between parent and child. It seems likely that parents with more than one child would label their children differentially, selecting a given child as more moody or more cooperative or more argumentative or less bright than other members of the family. Parents' demands of and responses to this child's behavior would be tailored accordingly e.g., a child Cent tiffed as especially bright might get more explanations. The labels given to a specific child, then, would presumably reflect not only that child's own behavior, but how it compared with that of siblings. To our knowledge, there has been no research that directly studies (~) the degree of agreement among family members in the labels applied to individual family members or (2) the way in which parental stereotypes or labels affect their interaction with each child. Several studies, however, do suggest a role for cumulative expectations. Halverson and Waldrop (1970) observed adults' reactions to boys who varied in hyperactivity. Women interacting with unfamiliar boys did not tailor their controlling behaviors to the level of the child's hyperactivity, even though hyperactive boys did display hyperactive behavior during the observational session. When mothers were observed interacting with their own sons, however, matemal behavior was related to the child's hyperactivity as independently assessed. Thus, the mothers must have been reacting to their accumulated knowledge of their child's hyperactive characteristics, rather than to the details of the behavior that occurred during the observational session. This interpretation is consistent with a report by Paton (cited in Bell, 1977) that when the impulsive behavior of hyperactive children is brought under control by medication, their mothers do not immediately change their style of dealing with the child. Instead, in a period as tong as several months, mothers reorganize their reactions toward, and presumably their perceptions of, their changed child. In a similar vein, Chapman (1979) noted that when a distracting stimulus was introduced during the task~oriented interaction of a mother and son, mothers increased their use of commands; the commands, however, were not contingent on specific instances of the child's loss of attention. Rather, the mothers seemed to be reacting to their anticipation of attention lapses. That this matemal reaction occurred with boys but not
200 DEVELOPMENT DURING MIDDLE CHILDHOOD with girls may reflect either an expectation that boys are more likely to be distracted or a cumulative experience that commands are more effective for controlling distraction with boys than with girls. The role of mutual cognitions has recently become the focus of some interesting research on peer interactions. Dodge (1980) showed that an ambiguous action by a school-age chilc! is interpreted differently by other children depending on whether the actor has a reputation for being aggres- sive. Furthermore, aggressive children are more likely than other children to interpret peers' behavior toward them as having aggressive intent. The two sets of perceptions conjoin to increase the likelihood of agonistic en- counters when interactions involve children with an aggressive history. We can only assume that mutual cognitions play a similar role between parents and children, and that their history of experience with one another leads each to make assumptions concerning the intentions that lie behind the other's overt acts. This is only an assumption, however, and many questions remain unanswered: How flexible are parents in changing their expectations and conceptions about their children as they undergo the kind of developmental changes outlined earlier? How can one interrupt cycles of self-fulfi~ling prophecies once they gather momentum? As children change with age, what are the changes in labels they give to their parents, in their differentiation between the two parents, and in the way their labels affect their reactions to parental inputs? These questions deserve a place on the agenda of future research on parent-child interaction in the school-age years. Thus far we have considered primarily the developmental changes that almost all children undergo during the 6-12 age period and how these changes may affect parent-child interaction. Parents also undergo life-span developmental changes. For example, they become more experienced child- rearers, especially if more children are bom. They may undergo occupational changes. As child-rearing demands are reduced, many mothers return to work, and one or both parents may take on more responsibilities outside the home. In some families, new responsibilities for the care of elderly grand- parents are assumed. The spousal relationship will almost inevitably change, and as their children reach the middle childhood years, parents must re- negotiate their respective roles in the child-rearing process. Furthermore, it is likely that parents' child-rearing values change as well, especially if the parents change occupation, residence, or marital status. We discuss below the impact of several changes that often occur in the lives of parents of school-age children marital disruption, mothers working but note here that the experiences of children in the middle childhood years have almost never been conceptualized in terms of the life-cycle stage that their parents have entered at the same time.
CONTEXT OF THE FAMILY 201 We move away from a description of normative developmental change and consider the fact that children change at different rates and along somewhat different developmental trajectories. Parents of school-age chil- dren also differ considerably from one another in the ways they deal with the issues of middle childhood and the degree to which they can enter effectively into the process of coregulation. VARIATIONS AMONG FAMILIES Dimensions of Parental Variation Most parents use a variety of child-rearing techniques, depending on their goals and the situation in which the child's behavior is an issue. As we noted earlier, parents control infractions occurring in their presence in other ways than they control actual or potential infractions occurring away from home. Parents also vary from day to day in the way they react to their children's behavior and in the demands they make on children. Thus when parents are differentiated as to their permissiveness or warmth or strictness, these designations reflect only average tendencies. Some parents are biased in one direction, others in another, but for any parent, considerable variation in behavior occurs from time to time. A parent may also interact quite differently with different children in the same family. It is known that children within a family differ from one another nearly as much as they differ from children in other families (Rowe and Plomin, 1981; Scarr et al., 1981), and this fact suggests considerable within-family variability in pa- rental as well as child behavior. It should be understood, therefore, that the research described below deals with between-family variability that may be large or small, relative to the amount of variation that occurs in a given parent from time to time and when dealing with different children in the same family. Parents differ in terms of the sheer amount of interaction they have with their children; in the kinds of behavior they regard as acceptable or unac' ceptable; in the aspirations they have for their children and the demands they make for accomplishment; in the kinds of disciplinary and motivating techniques they use to influence their children; in the amount of control they attempt to exercise; and in the affective quality of their relationships with their children. Early factor analysis of parents' reports concerning their child-rearing yielded two major dimensions of variation (Schaefer, 1959~. One dimension concerned warmth or acceptance and reflected the fact that some parents characteristically fee! and express considerable affection for their children-
202 DEVELOPMENT DURING MIDDLE CHILDHOOD taking pride in them and enjoying interacting with them while other parents express a cold, rejecting attitude, experiencing the child as burden- some and unrewarding. Most parents fall toward the warm end of this con- tinuum. The second major dimension, unrelated to the first, reflected the amount of control exercised, with some parents being highly restrictive and others highly permissive; most fall between the two extremes. A third di- mension in which parents differ emerged in some studies: the openness of their communication with their children (Baumrind, 1967~. Early research (Baldwin et al., 1945) stressed democratic child-rearing, which included the element of permissiveness. It also implied that parents felt the obligation to explain their requirements and disciplinary actions and that they allowed the child to participate in family decision making as much as possible. Efforts to relate these dimensions of childrearing to children's personality and social development had rather inconsistent results. Tight controls ex- ercised by parents sometimes were associated with the child's being inhibited and fearful, sometimes with anger and aggression, and sometimes with com- petent, mature functioning. It became apparent that the effect of firm con- trols depended on how parental demands were selected and imposed and on the affective context in which they occurred. Parental strictness that involves power assertion on the basis of their authority and power to punish has a different effect than strictness that is a consistent enforcement of rules grounded in discussion and explanation. Parental warmth generally a positive factor in development was also found to be variable in its effects, depending on whether it was accompanied by indulgence, overprotection, intrusive and anxious involvement in all aspects of the child's life, or reasonable rules firmly enforced. In addition, it mattered how contingent the parental behavior was whether parents' warm reactions occurred in response to something the child had done- reflecting sensitivity to the child's states and activities or occurred merely as an expression of the parent's mood. A number of studies also indicated that it was not sufficient merely to describe parents in terms of how frequently they used reasoning or explanation. The content of the reasoning was shown to be related to outcomes. In the next section, findings on the relationship between parent's child- rearing attitudes or behaviors and the characteristics of their school-age children are briefly summarized. Effects of Parental Variation Parent-child interaction studies involving young children tend to stress the effects of parental variations on the child's attachment to parents, co
CONTEXT OF THE FAMILY 203 operativeness with parents, and social interactions with peers. As children grow into the middle childhood years, these aspects of their development continue to be important, but new outcome variables especially relevant to this age are added: self-concepts, intemalization of moral values, social perception, prosocial and antisocial behavior. The discussion below summarizes studies that deal with preadolescent school-age children (see also Maccoby and Martin, 1983~. The reader is wamed that almost all of the research represents correlations between char- acteristics of parents and characteristics of children measured at the same time. Such studies tell us little about the direction of the influence processes that occur between parent and child but do serve to connect the parent- child relationship with school-age children's individual characteristics. For convenience, these characteristics are called "effects" or "outcomes" because they have been conceived that way in the literature. Some of the aspects of children's functioning discussed below are also discussed in other chapters in this volume; the focus here is on the way these child characteristics are related to the functioning of the families in which the children are reared. Aggression, Other Antisocial Behavior, and Undercontrol In a longitudinal study in Finland, Pulkkinen (1982) studied children in families when the children were ages 8, 14, and 20. She reported that predelinquent, and subsequently delinquent, behavior in the children was more probable in families in which the parents were relatively uninvolved with the children seldom talked with them, had little information about their activities and associates away from home, and were uninterested in the children's school progress. These parents tended not to listen to the cI~ildren's opinions when formulating rules or making demands. In short, the parents were indifferent and unresponsive or, to use Pulkkinen's term, "parent-centered. " The longitudinal study of children's development of ego controls by Block and Block (1980) has only begun to be reported. Their earlier findings on adolescents (Block, 1971) are consistent with the view that uninvolved, self-centered parenting is associated with undercontrolled, impulsive be- havior in children. Thus far, few additional data are available concerning the middle childhood period. Studies of the interaction patterns in the families of aggressive and pre- delinquent or delinquent children (e.g., Patterson, 1982) indicate that there are high rates of mutually coercive behaviors between parents and all the children in the family, but especially with the child who has been identified by the school, parents, or correctional authorities as out of control. Families
204 DEVELOPMENT DURING MIDDLE CHILDHOOD of these children seem to lack mechanisms for shutting off cycles of hostile interaction among family members, so that hostility escalates to high levels. Family members avoid one another when possible, have few joint activities, and are ineffective in their efforts to solve day-to-day problems. The quality of interaction in the family and in the aggressive child's extrafamilial be- havior has improved by teaching parents to state clear expectations of the child, to monitor the child's compliance, to use firm and consistent but nonviolent means of enforcing their requirements, to give praise and re- wards for good behavior, and to engage in problem-solving discussions of unresolved issues that lead to anger and depression. Studies of adolescent delinquents and young adult criminals consistently show that such persons have high rates of conduct problems during their middle childhood years. Furthermore, the incidence of conduct disorders is associated with pathology in the families of the deviant children: family discord, poor supervision of children, family instability that involves placing the children in nonfamily care for periods of time, and parental mental illness or criminality (Farrington and West, 1981; McCord, 1979; Robins, 1974; Rutter et al., 1975~. To a considerable extent, then, the antisocial forms of adult deviance have their roots in family dysfunction during middle childhood. Perhaps some of the dysfunction can also be traced back to early childhood. It has been shown that antisocial children tend to be deficient in social perspective taking; that is, they do not empathize with others' distress and tend not to consider how others will view their actions (Chandler, 1973; Chandler et al., 1974~. Such findings suggest that parenting techniques that foster children's acquisition of social-cognitive skills may be instrumental in helping children to regulate their aggressive impulses. There is very little research on the relationship of child-rearing practices to children's social cognitions and empathic responses, but there is some evidence that other- oriented induction-that is, consistently reminding children of how their behavior looks to others and how it will affect others is important (see Bearison and Cassell, 1975~. Self-Concepts Two aspects of children's self-concepts have been studied in relationship to parental practices. The first is the child's self-esteem, and at first glance the results do not appear to be consistent. When parents are strict in the sense of imposing many restrictions, their children's self-esteem tends to be low. When parents are strict, however, in the sense of exercising firm control-monitoring their children's behavior and following through on
CONTEXT OF THE FAMILY 205 requirements the children's self-esteem tends to be high. The critical factor appears to be whether parental strictness is accompanied by emotional sup- port, commitment to the children's welfare, and an open interchange of ideas between parent and child. Parental warmth is consistently associated with children's self-esteem (Loeb et al., 1980), while physical punishment and psychological punishment withdrawal of love-are associated with low self-esteem. A second aspect of parental influence on school-age children's self-con- cepts concerns the children's `'inner locus of control"-their sense of being in control of the events affecting their lives. When parents of inner-Iocus children are observed attempting to teach or help their children with a task, their approach is to make suggestions that improve the child's chances of arriving at a solution, while parents of children who have an external locus of control are likely to issue specific directives or take the task out of the child's hands. Internalization of Moral Values Children tend to show signs of conscience-confess misdeeds, fee} guilt- and make relatively mature moral judgments when their parents seldom use direct, unqualified power assertion. Parents of more morally mature children, when training and directing their children, stress the consequences of the children's actions for others. When these parental techniques are accom- panied by withdrawal of love and arousal of the child's guilt feelings, the child's intemalized moral values tend to be fairly rigid and rule oriented. When they are accompanied by firm follow-through on parental demands and low guilt induction, however, the child's moral orientation tends to be more flexible (M.~. Hoffman, 1970; Salzstein, 1976~. Prosocial Behavior Prosocial behavior includes helpful and emotionally supportive acts di- rected toward others. Parents' use of other-oriented induction has been found to be associated with prosocial behavior in school-age children (see Radke- Yarrow et al., 1983, for a review). In addition, experimental studies in which nonparental adults have tried to train children in prosocial behavior have found that emphasizing a victim's distress, and the victim's relief when helped, effectively fosters children's prosocial behavior. This training is most effective if the teacher has previously established a nurturant relationship with the child (Yarrow et al., 1973~. Effective helping of others depends somewhat on children's ability to take the perspective of others to un
206 DEVELOPMENT DURING MIDDLE CHILDHOOD derstand others' situations and feelings. As noted earlier, parents foster this ability by using other-oriented induction rather than authoritarian, directive techniques (Bearison and Cassell, 19751. Competence Baumrind (1973), in a longitudinal study of the development of com- petence in children, stressed three aspects of competence: cognitive abilities, social assertiveness, and social responsibility, i. e., effective cooperation with others, self-regulation of antisocial behavior. In a brief prepublication sum- mary of findings on the interactions of parents with their 8- to 9-year-old children, Baumrind reported that, in both boys and girls, social responsibility is associated with parents' responsiveness to their children's needs and bids for parental attention. For boys, social responsibility is enhanced if the parents make fairly strict requirements for proper behavior. Cognitive com- petence is associated in both sexes with high parental demands; in boys, however, parental responsiveness contributes as well. Social assertiveness in boys is not associated with either demanding parents or responsive parents; girls with demanding parents, however, tend to be socially assertive. Baum- rind suggested that each sex needs to be pushed by parents in order to develop a characteristic that is outside stereotypic sex roles: social asser- tiveness in girls, social responsibility in boys. Many of the correlations emerging from the studies reported above are small. Much of the variance in children's characteristics cannot be attributed to parents' behavior as measured thus far. Furthermore, some inconsistent themes have emerged from the work summarized above. One theme is that the development of competencies in children requires the firm exercise of parental control. For example, in Patterson's (1982) interventions with parents of antisocial children, parents were taught to closely monitor their children's behavior, clearly require certain conforming behaviors from them, and consistently punish infractions. Thus, the pattern of behavior that was encouraged for parents was fairly power assertive. The second theme is that power-assertive parenting is counterproduc- tive it promotes neither the internalization of adult values nor self-control. Indeed, some studies indicate that power-assertive pressure toward a partic- ular value or behavioral goal causes children to devalue that behavior (l~ep- per, 1983~. Can both themes be correct? Some researchers have attempted to reconcile them, but more work is needed. At present, it appears that parents foster optimum development in middle childhood by teaching, guiding, and emotionally supporting their children.
CONTEXT OF THE FAMILY 207 Children ages 6-12 do not usually learn competence and self~sufficiency by fending for themselves. At this age, parental guidance in the form of power assertion is less effective than firm control combined with warm respon- siveness and the exhortation that children's actions have consequences for themselves and others. Perhaps, however, we should consider the possibility that different patterns of change are optimal for families with different initial interaction pattems; that is, perhaps lax parents need to acquire power- assertive skills, and power-assertive parents need to acquire intemalization- fostering skills. We can only assume that the desirable balance between the two themes is linked to the age of the child. GROUP DIT[k;RENCES IN CHILD-REARING Historically, researchers have compared parents from different socioeco- nomic levels or ethnic groups with respect to their child-rearing values and their characteristic modes of interacting with their children. In recent years, interest has focused on family structure with regard to the following com- parisons: single-parent and two-parent families; families with two employed parents and those with one employed parent; families with high versus low father involvement in child-rearing; and reconstituted, i.e., stepparent, fam- ilies versus families with two biological parents. We will focus first on the demographic comparisons, then on family structural ones. Socioeconomic Status Child,Rear~ng in Different Social Classes When parents are grouped by income, education, or occupation or in- dexed by a combination of these average group differences in child-rearing consistently appear, although the overlap between groups is great (for reviews see Bronfenbrenner, 1958; Hess, 1970; Laosa, 1981; for studies providing data on specific issues, see Bee et al., 1969; Hill and Stafford, 1980; Newson and Newson, 1976; Shipman et al., 1976; Stafford, 1980; Yankelovich et al., 1977; Zill, in press). Major contrasts that have emerged with some consistency are that middle-class parents, compared with working- or lower- cIass parents: · Have higher rates of interaction with their children and are more re- sponsive to their children's bids for attention. Shipman and coworkers ~ 1976) showed that higher-income and better-educated mothers of 8.5- to 9.5-year' old children were more likely than their lower-status counterparts to give
208 DEVELOPMENT DURING MIDDLE CHILDHOOD informative-interactive responses to children's difficult questions; to know the child's teacher by name; and to take the child on outings outside the home. · Use more elaborated language in talking with their children. · Are more permissive toward sexual behavior and more accepting of children's displays of parent-directed anger; they are also less restrictive about the range of activities allowed the child. At the same time, they place somewhat greater demands on the child for maturity, achievement, and independence. · Place less value on obedience and respect for authority and more value on the child's happiness, curiosity, creativity, and eagerness to learn. · Are less power assertive and more democratic; that is, they are more likely to allow the child a voice in family decision making, more willing to listen to the child's point of view, less likely to give direct orders, and more likely to orient and motivate the child when tasks are undertaken. These differences are reflected also in disciplinary techniques: middle-cIass parents less often use physical punishment and more often use reasoning and ex- planation. Their reasoning more frequently takes the form of drawing chil- dren's attention to the effects of their behavior on others. · Show a lower incidence of clearly hostile or rejecting attitudes toward their children. The large majority of parents, regardless of education and income, are reasonably warm and accepting, and scores on summary measures of warmth or affection are frequently not significantly different for the social class groups. The differences appear at the lower end of the distribution. It seems clear that the sociocultural milieu in which parents function does affect their child-rearing. It is not clear, however, at what point along the continuum the greatest shifts in child-rearing occur. Various cutoff points have been used in different studies that compared groups. Some reports suggest that the major differences are between upper-middle-cIass parents and everyone else. It may be that with respect to several important aspects of child-rearing, such as the amount of affection expressed to the child, upper middle-, middle-, and working-class families are much alike, while it is the disorganized, multiproblem families (sometimes called the "under' class") who deviate in these aspects. Although the underclass represents a small proportion of all families, it is large enough to affect the means of any lower-cIass or working-class group in which it is included. We do not know, however, whether there are sharp transition points along the socioeconomic continuum in terms of how parents relate to their children. It remains to be discovered whether such sharp breaks exist, and if they do, where they occur with respect to various aspects of parenting.
CONTEXT OF THE FAMILY 209 Another question is: What aspect of socioeconomic status is most closely linked with what aspects of child-rearing? Laosa ~1982) found that education carries far more weight than income in predicting the teaching styles of Chicano mothers, while others emphasize the importance of affluence or economic deprivation in influencing parental behaviors. This question alerts us to a point that has been stressed by a number of writers: that social class is not a meaningful variable in itself; rather it is a convenient index for variations in the life experiences of parents and children. These writers urge that we delve beneath the mere facts of parental education, income, and occupation, and investigate the processes whereby these demographic facts are translated into family life. We need to ask two questions: What are some hypotheses concerning why the child-rearing methods and values of parents are linked to their social class? And given that the linkages exist, what are their implications for children? Some Models of Socioeconomic Differences in Child-Rear~ng Socioeconomic differences in child-rearing can be conceptualized in a number of ways. A first point of view, the cultural fag model, comes primarily from cross-cultural studies and argues that when simple, preliterate societies undergo modernization, parents will begin to teach children some of the skills and values they will need for functioning in a complex socioeconomic system (e.g., increased orientation to time). Within a society, families differ in the degree to which they are subject to modernization pressures. Hence, group differences in child-rearing emerge (Inkeles, 1960~. In advanced in- dustrialized societies there are presumably subcultural pockets that continue to resemble simpler levels of social organization. A second mode! may be called the rote-status transmission model. From this point of view parents who are in occupations that call for self-regulation and initiative and whose work involves symbols rather than materials are likely to value and foster independence and self-direction in their children, while parents who are in subordinate roles in authority structures will value and foster obedience and be less concerned with their children's inner mo- tivations and autonomy (Kohn, 1963, 1977~. A variant of this view is that a lower-class position in society implies powerlessness in many spheres of life, not only the workplace, and that parents who fee! powerless transmit to their children the message that they too are powerless (Sennett and Cobb, 1982~. This is done in a variety of ways, including low mutuality in com- munication (i.e., less negotiation). When poverty is chosen as an alternative life-style, however, so that it does not imply powerlessness, it is not accom
210 DEVELOPMENT DURING MIDDLE CHILDHOOD panted by the child-rearing styles that usually characterize lower-class parents (Weisner, 1982). Another model, called the expander functioning model, emphasizes the skills that are acquired through education. Education improves parents' vo- cabularies, improves their knowledge about issues, and promotes advanced styles of thinking and communicating with others. In particular, it brings about a shift from restricted to elaborated communicative codes (Bernstein, 1961; Hess and Shipman, 19671. Education also motivates parents to take an interest in their children's intellectual development, as manifested in more frequent reading to the child and the teaching of literacy skills (Laosa, 1982~. The stress~inte7ference model emphasizes the destabilizing nature of lower- class life. Major and minor crises (an unexpected bill, a bounced check, trouble with the car, accidents, something lost or stolen, loss of a job, divorce) function as stressors and, at least for some parents, directly affect the way they deal with children. Typically, parents under stress reduce their interaction with the children and become impatient, irritable, and peremp- tory. These parental reactions exacerbate the irritable behavior of the chil- dren, which adds to parental stress; thus, a cycle of deteriorating parent- child relationships is set in motion (Hetherington et al., 1982; Patterson, 1982~. Many kinds of crises are considerably more common in impoverished families. In those families stressors may have a multiplicative, rather than additive, effect. (Rutter, 1981:2091. Many, though not all, impoverished families appear to live close to the point at which additional external stressors would produce disorganization of family functioning loss of routines, abut siveness, or some degree of neglect or indifference toward children. Social- class differences in childrearing, then, may be viewed as a reaction to the differential levels of stress impinging on parents in different social classes (Maccoby and Martin, 1983~. Still another view is that what appear to be social-class differences in child-rearing are actually products of other demographic and ecological con- ditions with which socioeconomic status is correlated. Thus, Burrows (1981) reported that while substantial social-class differences exist in Mexican fam- ilies in the ways that parents deal with their children and children react to their parents, these differences are accounted for by the large family size and crowded living conditions of lower-class families, and once these variables are entered into a predictive equation, socioeconomic status produces no . a. slgnl~lcant variance. We must not overlook the possibility of reverse-order effects; that is, families may have drifted into the lower socioeconomic levels of society
CON17IXI OF THE FAMILY 211 because of some pathology in the family members. Individuals who suffer from mental illness, alcoholism, drug addiction, etc., have difficulty holding jobs and maintaining an adequate income; as young people, they have difficulty meeting the demands of a school regime, complete fewer years of schooling, and are disadvantaged by these deficits in their life chances later on. Hence, the disorganized and distressed quality of family life and the deficits in parenting that are found in a subgroup of lower-status families appear to be a manifestation, in some cases, of the same intrapersonal factors that led them into their lower status rather than an outcome of being poor or uneducated. No doubt there are circular processes at work: being poor or uneducated increases the risk of various pathologies; these pathologies in' crease the risk of being poor and uneducated. Socioeconomic Differences in Childrearing and Their Possible implications for Children Given that child-rearing varies in some respects according to the socio~ economic status of parents, what are the implications of such variations' Historically, researchers examined such differences in an effort to explain behavioral differences among children of different socioeconomic back' grounds. In particular, researchers asked whether certain aspects of child- rearing by lower-status parents are conditions that underlie the higher rates of delinquent and predelinquent behavior and the higher risk of poor ace' demic performance among lower-status children. A central idea here is that parental behavior serves as the functional mediator between a family's social status and those aspects of children's behavior that are linked to social class. That is, parents' education or position in the social structure has an impact on how they raise their children via some of the mechanisms listed above- and these cIass-linked variations in child-rearing produce differential be- havior patterns in children, some of which adapt children to the social situation in which they have been reared or maladapt them to other social milieus. Thus a conservative cycle may be set in motion, tending to produce in children the characteristics that will keep them in the same social class occupied by their parents. Laosa ~1982) argued, for example, that the teach' ing style employed by poorly educated mothers with their preschool children is dissimilar to the style the children encounter from teachers when they enter school; thus these children are poorly adapted to the school environ- ment. We suspect, too, that the themes of child-rearing wherein the major social-cIass groups differ (see listing above) are such that middle-class families more effectively achieve the parent-child cooperation that is needed for the
212 DEVELOPMENT DURING MIDDLE CHILDHOOD coregulation processes of middle childhood. If this is so, we have a tentative explanation for higher rates of out-of-control, antisocial behavior in {ower- status children. We must ask, however, if variations in parental behavior are indeed the main conduits whereby socioeconomic milieu is translated into meaningful experiences for children. Zill (in press), in a cautionary note, reports that maternal education is an independent and more powerful predictor of various child outcomes in his study than are the measured child-rearing variables through which maternal education presumably manifests itself. Of course, the problem may be that Zill's child-rearing measures were too narrow or too unreliable. But the question remains: What are the linkages between a social indicator such as mother's education, her child-rearing practices, and children's class-linked characteristics? We cannot attempt to generate a quantitative estimate of the variance in child-rearing that is accounted for by socioeconomic factors. All we can say is that it is large (see Newson and Newson, 1976) and surprisingly consistent across samples and measures. If it is true, in addition, that so- cioeconomic factors continue to account for a substantial portion of child outcomes after the class-linked variations in child-rearing practices have been partialled out, then serious consideration would have to be given to the possibility that parents are not the main channels whereby sociocultural influences are transmitted to children, but that the transmission occurs directly (e.g., through TV, neighborhoods, etc.) to both parents and chil- dren, producing cIass-linked characteristics in both parents and children that have little or no functional connection. Researchers have rarely at- tempted to unravel these causal complexities, no doubt because small sample sizes or samples lacking socioeconomic range preclude employing the mul- tivariate analyses that might be useful. One way to approach the problem is through some form of multiple regression in which socioeconomic status is taken out first, so that it can be seen whether parental behavior makes any additional contribution to a child outcome, and then to reverse the procedure, determining whether socioeconomic status makes an independent contribution once parental be- havior variations have been partialled out. These procedures, rarely used, examine the relative contributions of within-class versus between-class var- iations in parental behavior. Hess and Shipman (1967) have used multi- variate analysis in their work with a sample of black mothers and children from several social classes; they found first-order relationships between so- cioeconomic status and maternal teaching styles. Furthermore, child out- come variables were related to both socioeconomic status and matemal teaching styles. Multivariate analysis showed that significant variation in
CONTEXT OF HIE FAMILY 213 child outcomes was still accounted for by maternal teaching styles, even after socioeconomic status had been taken out, while the reverse was not true. Such results support the view that socioeconomic status had its primary effect via mother-child interaction. in general, however, there is scant information in the literature on middle childhood that reveals whether this pattern is a general one. By the prep adolescent years, many children begin to understand what their own family's position in the social structure is and develop attitudes about what this position implies in the way of opportunities and probabilities of success. Such cognitions clearly could contribute to chilciren's task motivations, among other things, over and above the effects of their cIass-linked expe- riences with their parents. Ethnic Group Differences We have limited information concerning how parents in ethnic cultures and subcultures differ from their mainstream counterparts in their interac- tions with their school-age children or concerning the effects, if any, these differences have on child development. The problems of identifying the links between ethnicity, parenting, and child outcomes are similar to those noted above with socioeconomic factors. Ethnic groups show great variability in the conditions of family life. Young, unmarried black mothers living on welfare payments in urban ghettos-so poignantly described by Carol Stack (1974) have life situations utterly different from middIe-cIass, two-parent black families with stable jobs and residences. Much parental behavior, such as maternal teaching styles, shows great variation within ethnic groups along socioeconomic lines (Hess and Shipman, 1967; Laosa, 1982~. Apart from socioeconomic factors, ethnic groups that may be perceived by outsiders as homogeneous subcultures in fact are not so. Hispanic families, for example, slider in their life-styles depending on whether their origins are Cuban, Puerto Rican, Central American, or Mexican, and although they presumably share the same language and many of the problems of bilingualism, even their linguistic usages differ considerably (Laosa, 1982~. Researchers some- times aggregate such subgroups in the interests of obtaining adequate sample sizes for ethnic comparisons, but in doing so they run the risk of obscuring important variations and relationships. Considering the great socioeconomic and other variation within ethnic groups, the dangers are obviously great that simple comparisons between black and white families or Hispanic and Anglo families will be confounded with these other variables. As is well known, black and Hispanic families typically have low average levels of income and parental education. An
214 DEVELOPMENT DURING MIDDLE CHILDHOOD excellent example of how these differences may generate what appear to be ethnic differences is provided by Laosa's reports (1981, 1982) on the teaching behavior of Mexican-American mothers. Observations of the mothers' teach- ing styles revealed that Chicano mothers, on the average, tended to use demonstration and directives in teaching their children to perform a manual task, while Anglo mothers tended to use questioning and praise. There were, however, no differences between Chicanos and AngIos in their teaching style when comparisons were made for subgroups matched on education. Laosa (1982) notated that education rather than level of occupation is the crucial factor determining Chicano-Angio differences in the aspects of child- rearing analyzed so far. Zill (in press) similarly found that a number of ethnic differences in child-rearing attitudes and values are a function of socioeco' nomic factors. When these are controlled, only a few distinctive ethnic c. aaracterlstlcs remain. Family structure is another source of confounding in comparisons between ethnic groups. Single-parent families constitute a considerably larger pro portion of black than white families, and it is rare to find studies in which comparisons have been made between minority and majority families who have been equated for both socioeconomic and family structural character- istics. Ethnicity and Family Influence in Middle Childhoo~l The limited research on families in minority subcultures focuses primarily on infants and preschool children. Only two large-scale interview studies with carefully selected representative samples provide some data on the middIe-childhood period: the national survey by Zill (in press) of 2,301 children ages 7-l ~ in 1976-1977 and the more limited General Mills Amer- ican Family study (Yankelovich et al., 1977) of 1,230 families with children under the age of 13, also conducted in 1976-1977. In addition, data are now available from a study of 764 sixth-grade children and their families in Oakland, California, the majority of whom were black (Medrich et al., 1982~. The latter study found lower levels of interaction between parents and children in black than white families; that is, black families had a lower incidence of all family members eating together, sharing hobbies, going places together, and parents facilitating the children's activities e.g., tak- ing children to lessons or sports practice sessions. Black children in this study spent more time watching TV. These differences persisted after equat- ing for education and income. We should note that this study was conducted in an urban area with a comparatively high proportion of educated, stable middle-income families. We do not know whether the same picture would emerge in an inner-city ghetto or among rural families.
CONTEXT OF THE FAMILY 215 The General Mills study (Yankelovich et al., 1977) reported that black parents place a higher value on achievement in both school and sports- than do white families and that their children, reporting their awareness of this pressure, concur. From the Zill (in press) study we leam that, with socioeconomic factors equated, Hispanic families place greater emphasis on such traditional values as respect for authority than do other ethnic groups, while black families are more likely to use physical punishment. The latter finding echoes themes that emerged from a variety of early studies that used both observational and interview data and showed black parents to be more restrictive and punitive than white parents, and in some instances more overprotective (see Peters, 1981, for a review). Interpreting Ethnic Group Differences The use of physical punishment provides illustration of a more general issue: the same child-rearing variable may not have the same meaning in different cultures or subcultures. For example, Baumrind reported (1973) that authoritarian child-rearing was associated with positive outcomes in the small group of black girls in his study, while the reverse was true for white girls. Some studies have indicated (Peters, 1976; Young, 1970) that, while the frequent use of physical punishment is associated with a number of adverse outcomes in mainstream children, it may not have these correlates in black children. Is physical punishment part of a different cluster of par- enting practices in black families? Does its greater frequency make it part of a cultural norm so that it is experienced as less punitive by the children involved? We do not know. We are merely alerted by this example to the dangers of generalizing the findings of mainstream-sample studies of child- rearing to minority cultures. Important issues are posed by the question of how to interpret ethnic differences in child-rearing pattems. Debates over the deficit versus the pluralistic interpretations continue (see Laosa, 1982; Peters, 1981~. The dangers of the ethnocentric imposition of middle-cIass white values on mi- nority cultures are frequently pointed out and need no further elaboration here. In the larger context of cross-cultural differences in child-rearing, LeVine (1979) argued that cultural groups develop a pattern of parental functioning that is adapted to the cultural, economic, and ecological niche in which the group functions. LeVine stressed that, by and large, such patterns are successful adaptations. It is only under conditions of sociocultural change that they may become maladaptive. Perhaps in the long run this point of view will enable us to arrive at some reasonable interpretations of subcultural differences in child-rearing. It is too early, however, to embark on this enterprise. As far as middle childhood is concerned, we lack the
216 DEVELOPMENT DURING MIDDLE CHILDHOOD necessary basic data on ethnic variations in child-rearing or their relationship to child outcomes. And analyses have not been performed that would allow us to understand the complex interweaving of such factors as income, ed- ucation, ethnicity, family size, urban residence, and single parenthood. Single-Parent Families We turn now to variations in family structure and will consider both the current changes in family structures and the implications for school-age children. At the time of this writing, approximately one-fifth of American children ages 6-12 are living with only one parent, and as many as 40 percent will probably live in a single-parent household at some time during their childhood (see Chapter I; see also Click, 1979~. The number of single- parent households has more than doubled in the last 20 years, ant] the majority of children in single-parent households (over 90 percent) live with their mothers. About three-quarters of the single-parent households result from the divorce or separation of married couples, although a growing mi- nority are composed of never-married mothers and their children. The ethnic differences in the incidence of single-parent households are extreme: in 1977, the percentage of children under 18 living in such households was Il.9 for whites, 19.5 for Hispanics, and 41.7 for blacks. Typically, single-parent families are poor. Writers have recently begun to point to the feminization of poverty (Pearce and McAdoo, in press); that is, increasingly, families below the poverty line are headed by a single female parent. The number of such families is increasing: 38 percent of white mother-only families and 86 percent of black mother-only families fall below the official poverty line, compared with 16 percent and 46 percent, respectively, of two-parent fam- ilies (Hill, 19831. The average income of single-parent households is much lower than two-parent households. Single-parent families are heterogeneous in a number of ways. Eiduson and colleagues (1982) described three kinds of never-married single mothers: a group of nest bedims, well-educated and self-supporting women who have elected to raise a child alone; a group of post hoc adapters who are not so wed! off as the nest builders but are possessed of sufficient personal and economic resources to manage their unplanned parenthood adequately; and a group of unwed mothers who did not welcome the pregnancy and are young, impoverished, poorly educated, and lacking in job skills. This last group turns either to welfare or parental assistance for economic support. Pre- dictably, the Eiduson group reported considerable variation among these subgroups in values and coping styles. Nevertheless, when the subgroups were combined as a composite of single parents, they differed from traditional
CONTEXT OF TIE FAMILY 217 families and other altemative life-style families in that they experienced more residential moves and other forms of stress than any other group, were less satisfied with their lives, and had more disorganized households. Single parents suffer from task overload. They must carry out alone the parenting responsibilities that are shared between two parents in intact households. In most cases, single mothers work full time; therefore, they have less time for parenting functions than the part-time employed mothers in dual-parent households. For the majority of single parents, the stress imposed by these realistic burdens is exacerbated by the emotional turmoil that attends and follows marital disruption. A number of studies of divorcing families (Hetherington et al., 1982; Wallerstein and Kelly, 1980; Zill, 1978) document what happens to the quality of parenting and the adjustment of children after families move from dual-parent to single-parent status. These studies include preschoolers and adolescents, but a substantial portion of the data deals with the middle-childhood period. From these and other sources, several consistent themes emerge: 1. Marital separation usually entails major emotional distress for children and disruption in the parent-child relationship. During the several years following separation, single parents when compared with their own earlier patterns of behavior and with parents in intact families typically show diminished parenting; that is, they tend to be self-preoccupied, less attentive and responsive to their children, less consistent, and more irritable. They spend less time with the children and tend to neglect such household routines as laundry, cleaning, regular meals, bedtimes, and baths. For many single parents, parenting functions return to a more organized and child-centered pattern as the stress of separation diminishes; this is more likely to happen when the parent forms new intimate ties and is no longer "single." As the custodial family stabilizes over time, contact with the noncustodial parent usually diminishes, so that the outside parent becomes peripheral in the lives of most children of divorce (Furstenberg et al., 19831. 2. The effect of marital disruption on children depends on the child's age at the time of the breakup. According to Wallerstein and Kelly (1980), the primary reaction of children ages 6-8 is marked grief and fear and an intense longing for the reconciliation of the parents; these reactions may continue into the later school-age years. By ages 9-12, children express a greater sense of shame over the divorce and more open anger toward one or the other of their parents usually the parent deemed responsible for the divorce; they are thus vulnerable to being drawn into alliances by one parent against the other. Furthermore, children of this age are more likely to reject a stepparent. At both ages, deterioration of the children's behavior at home
218 DEVELOPMENT DURING MIDDLE CHILDHOOD and at school may be seen: increases in aggression; difficulties in maintaining friendships with peers; resistance or excessive dependency toward teachers; deterioration of school work; and in some children, a drift toward predelin- quent activities and groups. 3. We know more about short~term effects of family disruption than we do about long-term effects. A remedy for this gap is long-term follow-ups of individuals whose families were disrupted during their childhood, although simple comparisons between such persons and persons from intact homes are likely to be confounded by socioeconomic status or ethnic differences. Some studies indicate that the long-term effects of broken homes may be modest for many of the children involved (Kulka and Weingarten, 1979), while other studies show continuing effects. A recent study of college stu- dents showed that students from intact homes have closer relationships with both parents but particularly fathers at this young adult period than students whose parents were divorced at an earlier time (Fine et al., 1983~. In addition, there is evidence for an intergenerational transmission effect- that is, for individuals from broken homes to have higher rates of disruption in their own marriages and to show deficits in parenting (see Rutter, 1979, for a review). The effects of early disruption may depend on the concurrence of other risk factors. Thus Rutter (1981) reported that broken homes are associated with children's later delinquency and criminality in inner-city London but not on the Isle of Wight and attributed the difference to the fact that the restoration of harmonious family relations following divorce is more likely to occur on the Isle of Wight than in inner~city London. The outcome measures in the existing studies of long-term effects did not include some of the effects that, from a theoretical standpoint, may be especially likely. Specifically, studies of early separation from attachment figures suggested that one of the sequelae of such separations might be increased vulnerability to disturbance over subsequent separations. General measures of happiness or the number of worries an individual has may not be fine-grained enough to pick up such specific effects. When studies simply compare the children of single-parent families with other children, the reasons for any differences are unclear. For example, Yankelovich et al. (1977) reported that single mothers are more likely than other mothers to say that they have sometimes lost control and punished their children more than they deserve. Children of single mothers are more likely to argue with their mothers and have more difficulty getting along with other children. We do not know whether such problems are most likely to appear in divorced families, in families where the mother has never married, or in families where the father is dead. We also leam from the
CONTEXT OF THE FAMILY 219 Oakland study (Medrich et al., 1982) that single parents were less likely to register their children for out-of-school activities or spend time taking them to such activities; however, we do not know which subgroup of single parents- widowed, divorced, separated, or never-married was responsible for the lower facilitation scores. The Zill (1978) report provides a more differentiated picture. In the case of a widowed single parent, the mental health of the mother and children is very similar to that in two-parent families with reasonably happy marriages. This finding strongly suggests that the distinctive characteristics of single- parent families are not solely the result of the absence of an adult male in the household. Nor can we make the case that two parents are in all respects better than one, although in terms of time pressures, they usually are. Are the characteristics of single-parent families primarily a result of the stress of marital disruption' To some extent they are, but it should be noted that the indicators of family distress are as great for the never-married mothers as for the divorced and separated ones. The loss of a parent, therefore, may not be the primary factor. Do the higher levels of distress in single-parent families stem primarily from their poverty? Again, in part, yes, but Zill (1978) showed that after scores were adjusted for income and education, distress remained relatively high in all the groups of singe parent families except the widowed ones. An unresolved issue concerns whether single-parent families suffer more difficulties in parent-child interaction than do discordant intact families experiencing much parental quarreling. Zill (1978) and Rutter (1982) both reported that signs of disturbance in children and distortions in parental behavior are as great in disharmonious two-parent families as they are in divorced or separated ones. In the study by Hetherington and colleagues (1982) and that by Wallerstein and Kelly (1980), however, the separation itself appears to have added considerable distress to all parties concerned, with the critical event being the departure of one parent from the household. Children often seem unaware of how serious their parents' dissension is and tend to be unprepared for the breakup. The split in living arrangements calls for the formation of a new kind of relationship with the noncustodial, visiting parent and for adaptation to a custodial parent who may be more vulnerable after the departure of the spouse. There is great variability among single parents in the effectiveness with which they cope with the special demands of their situation. Some function well, better than they did in unsatisfactory marriages. Many do not. Little is known about the factors that underlie these variations. Studies of divorce indicate that continued involvement by the former spouse in child care and economic support considerably eases the burdens of single-parenthood, as
220 DEVELOPMENT DURING MIDDLE CHILDHOOD does the formation of new intimate ties. Other kinds of support systems are valuable, particularly kin relationships in low-income minority communities. Minority families tend to have a larger number of kin living nearby, even though the number of nonparental adults living in the household may not be unusually high. While it is true, however, that minority single parents have considerable contact with kin, the nature of the contact is highly variable. Sometimes these contacts are supportive; in other cases they create additional strains and pressures (Stack, 1974~. Working Mothers The second major social change that has occurred in recent years in addition to the increase in single parenthood has been the substantial in- crease in the number of married women with school-age children who work outside the home. Reviews are now available that summarize what is known about the impact of matemal employment on family life and on the devel- opment of children (Hayes and Kamerman, 1983; Hoffman, 1979; Kamer- man and Hayes, 1982; Lamb, 1982; Moen, 1982~. For the purposes of this chapter, there are several limitations on much of the research. The majority of studies examine the impact of matemal employment on infants and prep school children. Furthermore, a number of studies merely record whether a mother is working or not, without specifying whether the employment is full time or part time. In fact, although at least one-half of married American women with school-age children are now working, less than one-third work full time, and there is reason to believe that this is an important distinction with respect to the socialization of school-age children. Does the employment of a mother place her children at any sort of risk? On the whole, the answer would appear to be no, but this answer must be qualified with "it depends." It depends, for one thing, on the socioeconomic situation of the family. If the mother is single, impoverished, and poorly educated, in many cases the children appear to benefit from the mother's employment, partly because of the increased income, but also because of the mother's improved morale and sense of competence. Possibly, too, the children benefit from their experiences in day care while their mothers are working. The impact of matemal employment also depends on the sex of the child. In poor families, there is evidence that the sons of employed mothers are somewhat less well adjusted than when the mother is not em- ployed; and in middle-cIass families the sons of employed mothers may perform more poorly in school than they otherwise would. The daughters of working mothers, however, fairly consistently display higher achievement motivation than the daughters of nonemployed mothers.
CONTEXT OF THE FAMILY 221 In the majority of homes with two working parents, the mother still assumes major responsibility for household management and care of children. She is the one who usually takes the child to the doctor or dentist, to after- schoo! lessons, or for outings on weekdays. And she also does most of the shopping, cooking, cleaning, and laundry. When she works full time in a 9 to 5 job, particularly when any other adults in the household are away from home at the same time, considerable task overload ensues. Working mothers report that time is their most precious and most scarce resource. They fee} they have too little time to spend with their children; their school- age children echo this feeling, complaining that their parents spend too little time with them (Yankelovich, 1977~. Recent studies of children at risk for delinquency highlight the impor- tance, for the avoidance of antisocial behavior, of parental monitoring of children's activities and whereabouts (Patterson, 1982; Pulkkinen, 1982~. An understudied issue concerns the methods whereby working parents mon- itor their children's after-school activities and the effectiveness with which they can do so. Typically, infants, toddlers, and preschoolers are placed in formal or informal day care while their parents work. Thus, there is an adult formally responsible for their supervision. When children enter school, par- ents feel that the need for such formal arrangements is greatly reduced, since school serves a major baby-sitting function. Furthermore, as children enter and progress through the school-age years, the supervision they require changes dramatically. Even during the hours when they are not in school, they are often away from the house, playing with other children or attending a variety of after-school activities. Thus, as we have noted earlier, the parent's task is no longer one of monitoring a child who is within sight or immediate call. In the task of coregulation, the parent and child must cooperate in exchanging information about their plans and whereabouts. The child must take some responsibility for checking in, and the parent must be skillful in determining whether the child has told the truth, kept the parent adequately informed, and lived up to agreements. Parents differ significantly in how well informed they are about where their school-age children are, who they are with, and what they are doing. The flexibility of the mother's working hours may make a considerable difference in the extent of parental monitoring. The Oakland study of sixth- grade children (Medrich et al., 1982) reported the following group differ- ences with respect to the percentage of families in which neither parent was at home after school: when the mother works full time, 66 percent; when she works part time, 20 percent; when she does not work, 12 percent. A possible implication of such differences comes from the study by Dornbusch et al. (1983), in which it was shown that in the age range 10-14, children
222 DEVELOPMENT DURING MIDDLE CHILDHOOD began heterosexual dating earlier when they lived in households in which no adult was at home after school. We know little about how parents monitor their children's activities during out~of~schoo! hours when parents are working and no adult is at home. Most schools do not provide after-school programs, and even when they do, children may choose not to attend. While a neighbor or relative may agree to "keep an eye" on a child after school and provide a place where the child can check in, we do not know whether these arrangements result in as effective monitoring of the child's activities as would be managed by the child's parents. We know that there are "latchkey" children (how many?) who are instructed to go home and remain at home until their parents return from work and whose mothers check on their whereabouts by phone; not all parents have easy access to telephones at the workplace. We can only speculate that in some undetermined proportion of cases, supervision is loose and the children are allowed to drift toward activities and companions that parents would disapprove. For example, the General Mills study (Yanke- lovich et al., 1977) showed that children of working mothers are more likely than children of nonworking mothers to have friends or acquaintances who have played hockey, gotten into trouble with police, run away from home, or experimented with drugs. Such predelinquent companionship has been shown to be a factor in the onset of delinquent or predelinquent behavior. Unfortunately, in the General Mills study the comparison between the children of working and nonworking mothers was not controlled for socio- economic or family structural variables; therefore, it is not possible to de' termine what the independent contribution of the mother's work status is when it is uncomplicated by these other factors. This brief review of some of the group differences in child-rearing patterns indicates that family function is related to the cultural and ecological niche in which the family is located (see Weisner, in this volume). Undoubtedly there are other aspects of family life- such as rural or urban living, the nearness of relatives, the presence of physical hazards in the neighborhood, or the degree of crowding that affect the child directly but also indirectly through their impact on parents and their children-rearing practices (Bron- fenbrenner, 1979~. To a considerable extent, parents determine their chil- dren's physical environments- the nature of the house and neighborhood in which they will live, the objects with which the house is provided but the parents are themselves constrained in making these choices by economic and cultural factors over which they often have little control. We have seen that sociocultural factors do not act singly. Family income, ethnicity, pa- rental education and employment, and family structure are intricately linked.
CONTEXT OF THE FAMILY 223 To date, children's environments have been conceptualized mainly as forces impinging on them from the outside, which act on children but are not acted on by children. As children enter the middle childhood period, however, they begin to exercise some selection with respect to the settings they will enter and their social companions. Furthermore, it begins to matter to them how they conceive of the environmental and interpersonal situations in which they find themselves, and the competencies that they believe they can bring to bear in coping with them. FAMILY ROLES AND SYSTEMS Much of the research that has been summarized so far has dealt with parent-child interactions as though the parent-child dyed were the central functional unit through which families influence the development of chil- dren. We know, however, that the reality is much more complex. The relationship of a parent to a child is influenced by how the two parents relate to each other. One parent may function as an executive for the other; some parents provide support for one another's child-rearing, with positive effects; discord between parents has been shown to weaken the child-rearing effectiveness of either or both parents (see Maccoby and Martin, 1983, for a review). There is some evidence that the presence of a normal parent with whom the child has a good relationship will moderate or even cancel out the disruptive influence of a pathological or highly distressed parent (Heth- erington et al., 1982; Rutter, 1971~. The number of children in the family as well as their sexes and spacing also affects the functioning of the family as a whole. There have been efforts to analyze the functioning of families as systems. For example, Bell and Harper (1977) discussed the upper and lower bounds that families set for the behavior of family members and the ways in which the family as a whole will move to redress the balance if the behavior of a member exceeds these bounds. Samaroff (1982) considered the family as a self-stabilizing system that can adapt to perturbations in the environment. In addition, clinicians have been concemed with the way in which thera- peutic interventions with one family member may affect the functioning of other family members, for better or worse. Thus far, however, these con- ceptualizations have seldom entered the mainstream of research on the family, and little consideration has been given to whether there are changes in the family as a system when the developmental level of the children in the family changes. In the remainder of this chapter ~ focus on two aspects of the functioning of families that have been relatively neglected so far: the
224 DEVELOPMENT DURING MIDDLE CHILDHOOD distinctive role of fathers and the role of sibling relationships in the func- tioning of families in the middle childhood years. Fathers' Roles The changes in women's roles particularly their increasing participation in work outside the home have been accompanied by ideological changes concerning fathers' roles in child-rearing. The movement toward a legal presumption of joint custody of children following divorce is one expression of the growing emphasis on the importance of fathers' playing a continuing role in the lives of their children. In intact families, at least among the current and recent college generation, fathers are expected to assume a larger share of day-to-day child-rearing duties. Increasingly, fathers are present at the birth of their children reflecting both parental wishes and the adap- tation of medical practice to accommodate the changes in views about fathering. A body of research on fathers has grown up in recent years (see Lamb, 1981, and Parke and Tinsley, 1983, for reviews). One question that has been of considerable interest to researchers is: To what extent are the recent changes in attitudes about the role of fathers reflected in the day-to-day lives of families7 First, a distinction should be made between competence and performance. Research with infants and young children has shown that fathers can perform competently all the child care activities that have tra- ditionally been part of the matemal role, beyond childbearing and breast- feeding, of course. There is no reason to believe that fathers' competency does not extend to the middIe-childhood period, although this question has not been studied. The frequency, however, with which fathers actually engage in child care activities seems largely unaffected by the changes in ideology and women's work patterns outlined above. Results from studies of fathers' participation in families with working compared with nonworking mothers are not entirely consistent, but the overall picture is one of very small increases only a few minutes per day if any, in the amount of fathers' time spent caring for children when mothers work. When modest increases do occur, they are likely to be confined to the period when there is at least one child under age 3. Some literature suggests that fathers begin to exercise a more equal role in child care only when they take over some of the executive-planning and management functions, and these appear to be difficult to divide between the parents. Furthermore, the traditional styles of interaction characteristic of the two parents remain remarkably unchanged even in the relatively rare cases in which fathers do take on a fairly equal share of the child care responsibilities (Lamb et al., 1982~.
CONTEXT OF THE FAMILY 225 The care-taking activities that loom large with infants and preschoolers diminish in importance as children enter the school-age period. We know little about the distinctive contribution of fathers at this time. Research has focused on a relatively narrow range of issues e.g., the changes in fathers' functioning following divorce, the impact of father availability on sex-role development in boys and girls, and occasionally the father's role in moti- vating children for school achievement. Clearly, there are many other things we would like to understand, such as how mothers and fathers function or fail to function to support one another's child-rearing activities. it seems likely that both parental teamwork and division of labor are necessary for maintaining the coregulation processes with school-age children. This ques- tion leads to a larger one: Do fathers function mainly as a parallel parent influencing children in the same degree and in the same way as mothers? Or do the two parents have distinctively different influences on the children, by virtue of their gender and their differentiated extrafamilial activities7 There has been some work on nontraditional families in which the fathers take on an equal or even primary role in child care (see Parke and Tinsley, 1983, for a summary). Though these studies deal with families with very young children, some results are relevant to older children. For one thing, shared caretaking arrangements may not be stable. In a study of families in which fathers took at least an equal caretaking role with mothers during the child's early years (Russell, 1982), the follow-up study showed that within only 2 years after the initiation of the study, three-quarters of the families had resumed to a traditional pattern of greater maternal than patemal in- volvement in child care. The fathers, however, who had formerly taken a large share of the child care responsibilities but no longer took this role, had a closer relationship with their children than fathers who had not been similarly involved. Longer-term follow~ups are needed to determine whether the early division of labor in child care has an impact on father-child re- lationships in middle childhood. Sibling Relationships The body of research on siblings and their role in the lives of children during middle childhood is sparse indeed. A recent review by Dunn (1983) concentrated on sibling relationships in early childhood, in part because more information was available for this age, but a number of points emerged from the review that have relevance for the middle childhood period. 1. The affective relationship between siblings is usually both intense and ambivalent. The relationship of older to younger siblings has some of the
226 DEVELOPMENT DURING MIDDLE CHILDHOOD elements of the parent~child relationship, in that the older child performs some caretaking, teaching, helping, dominating, and occasionally punish- ing. There is some evidence that older siblings age 6 or 7 are more effective in teaching a structured task to their younger siblings than are unrelated persons (Cicirelli, 19721. Further evidence of the role complementarily between siblings is evidenced by younger siblings who do more imitating and show more attachment behavior toward the older sibling than vice versa. At the same time there is considerable reciprocity in the relationship: intense pleasure in joint play, mutual empathy, and fairly frequent inter- changes of reciprocated anger and teasing. In the families of hyperaggressive children, the amount of quarreling with siblings is especially marked (Pat- terson, 1982~. 2. By the age of school entry, children who have siblings are spending considerably more time with them than they spend with their parents (Bank and Kahn, 1975~. 3. Poor sibling relationships during the preschool years predict both con- tinued poor sibling relationships by the age of school entry (Stilwell, 1983) and more antisocial behavior toward others outside the family in 8- or 9- year-old children (Richman et al., 19821. 4. The relationship between siblings is related to the mother's interaction with the children. In a study of fourth- and fifth-grade girls interacting with their mothers and with sisters two or three years younger than themselves, Bryant and Crockenberg (1980) found that mothers who are above average in responsiveness to their daughters' needs have children who show more prosocial and less hostile behavior toward their siblings than the children of less responsive mothers; furthermore, intersibling hostility is greater if the mother shows favoritism to one of the siblings. These findings are consistent with the research on younger children, which has underscored the important role played by the mother in shaping sibling relationships (Dunn and Ken- d~ick, 1982~. The findings also point strongly to the importance of consign Bring the family as a system rather than as a collection of dyadic relationships. 5. Sibling relationships do not always lead to sibling similarities. Indeed, one of the major things that older siblings leam after the birth of a younger sibling is the ways in which they are different from the younger child. Mothers stress the differences: "You used to do that [drink from a bottle; soil your diapers!, but now you're a big girl"; "the baby is too little to sit up at the table with us; isn't it nice that you can have dinner with Mommy and Daddy?" Despite occasional regression after the birth of a younger sibling, older siblings frequently show sudden gains in social maturity. Dunn (1983:801) stated: "Studies of middle childhood report that an age-gap of 2-4 years heightens the differences between sibling pairs." In general, however, re- search is only beginning to explore how the presence of siblings fosters
CONTEXT OF THE FAM1L Y 227 processes of differentiation and individuation, as distinct from mutual iden ... . tl~lcatlon. We have noted above that siblings are frequently dissimilar with respect to various personality and intellectual attributes. One source of this differ' entiation may lie in the family dynamics that lead different children to take up different roles and that cause parental reactions toward one child to be influenced by and differentiated from their relationship with another child. There are many things we do not know: Do parental reactions exaggerate sibling differences? How do the dynamics of same~sex and cross~sex familial relationships play themselves out at this period of the children's lives? The existence of substantial sibling differences provides an interesting counterweight to the earlier emphasis on socioeconomic factors and family structures. ~ have argued that a family's sociocultural milieu has important implications for the way the family functions. Siblings share their family's socioeconomic status, and they live together in either a single' or two Parent family with the same working or nonworking mother. Presumably, if stressful conditions impinging on the parents make them less responsive to their children, the impact of this change also falls on all children in the same family. Can the differences in the way parents treat different children in the same family be so great as to outweigh and affect each child more than do the common elements of the general family environment? A clue comes from the work of Elder ( 1974, 1979), who found that the economic stresses experienced by families during the Great Depression of the 1930s had neg- ative consequences for young children at the time of their fathers' unem' ployment but generally positive ones for adolescent children. Unemployed fathers tended to behave in arbitrary and punitive ways toward their children; older children could escape from these intrafamily tensions to some degree, while younger children could not. Furthermore, the older children could get part-time jobs and make positive contributions to the family economy. The Elder research highlights the fact that family characteristics may have profoundly different effects on children of different developmental levels. Some of the research on divorce, cited earlier, makes the same point. These considerations return us full circle to the main theme of the first part of this chapter: it is time for researchers to give more serious thought to the developmental status of children and the way in which this Bevels opment affects both parents' child-rearing methods and the more general sociocultural milieu in which the family Unctions. METHODOLOGY We have noted that the direction of effects in parent-child interactions has not been established. This problem is probably of greater importance in
228 DEVELOPMENT DURING MIDDLE CHILDHOOD middle childhood than at earlier periods, because by middle childhood chil- dren have developed considerable skit'! in putting pressure on their socializers to achieve their own goals and in moderating or deflecting the socialization pressures that reflect the socializers' goals. Methods for identifying the di- rection of effects are currently under active development and debate. Ex- perimental interventions, in which either the parent's behavior or the child's behavior is changed and changes in the partner noted, provide the most solid foundation for causal inference, although the strength of the conclu- sions will depend, among other things, on sample size, random assignment of cases to experimental and control groups, the use of placebos, and the blindness of observers. Pharmacological interventions (e.g., Barkley, 1981; Barkley and Cunningham, 1979; Barkley et al., 1983) provide excellent examples of such an approach. Interventions with clinical populations in- volving the training of a group of parents to improve their child-rearing skills are potentially powerful but often involve great difficulties in the selection of suitable control groups and the maintenance of observer blind ness. For ethical and practical reasons, however, experimental interventions are not possible for the study of many aspects of family functioning. In such cases, sequential data are of special value in unravelling causal sequences; such data can take the form of either ( 1 ) time-series or event-series coding of segments of ongoing parent-child interaction or (2) longitudinal data in which comparable variables are measured for both parent and child at suc- cessive points in time (see Maccoby and Martin, 1983; and Martin et al., 1981, for detailed discussion of these issues). Perhaps the most difficult problem in designing a longitudinal study so that it will permit such analyses is deciding what behaviors are to be con- sidered the same or even comparable from one age to another, because many behavioral systems change their form radically with children's development. Longitudinal work on the sequelae of early secure or insecure attachment, for example, has had to select outcome measures at preschool age quite different from the ones that indexed attachment at 12 or 18 months. And while change may not be so rapid between ages 6 and 12 as it was earlier, it is still great enough to create serious problems in longitudinal study. Age changes in the definition and measurement of a given attribute render ques- tionable the methods of longitudinal analysis that call for the partialling out of autocorrelations. Such techniques have been used most successfully over relatively short age spans for which the same measures can be used through the life of the study. We suspect that these issues of developmental change constitute one reason why using structural models for time-sequential data analysis has trickled only slowly into child development research.
CONTEXT OF THE FAMILY 229 Nevertheless, the time is ripe for a wider application of the sophisticated multivariate analysis techniques that have been developed in recent years for nonexperimental data. These techniques require large samples and mul- tiple measures of single constructs. Hence they are not likely to be feasible for studies calling for the kind of intensive assessments that can only be done with small samples. Nevertheless, when the number of variables is relatively small and it is possible to plan for a large sample, structural mod- eling procedures have begun to prove their power. An example comes from the work of Entwisle and Hay~uk (1982~. These investigators took note of the correlation between children's school achievement and their level of expectations for their achievement in school-related tasks. The study sought to determine which way the causal arrow pointed; by estimating altemative models, it showed that first-grade children's expectations determine their achievement rather than the reverse-a result that runs counter to findings with older children. In the section on child-rearing characteristics of parents from different ethnic groups, we noted that ethnicity as a variable is often confounded with other characteristics of these groups, notably eclucation and economic level. We implied that ethnic groups should be matched for these charac- teristics in order to obtain valid comparisons. The same kind of reasoning applies to comparisons of other groups. We need to introduce some cautions, however, concerning either planned or ex post facto matching. Selecting a subgroup of, say, single white mothers who match single black mothers with respect to other demographic variables results in the selection of a group of single white mothers who are unrepresentative of their own population. For longitudinal studies, this fact raises the issue of regression effects. Serious questions are being raised concerning the legitimacy of partialling out con- founding variables in attempts to isolate the effect of single variables in "causal" analyses of samples in which it has not been possible to assign cases randomly to groups. These problems do not render group comparisons useless for some purposes, of course, but they do call for a reconsideration of how to analyze nonexperimental studies (Cronbach, 1982~. The utility of causal models is currently being hotly debated. Clearly, there are trade-offs. Further application of these techniques in a variety of studies is needed before we will understand the limits of their applicability and can identify the kinds of studies in which the techniques will lead to different and more valid conclusions than would otherwise be possible. Interviews with teachers, parents, and children; Q-sorts by parents and teachers; and paper and pencil assessment procedures (e.g., personality as- sessment batteries, moral dilemma stories) have been used extensively in studies of school-age children. The self-report measures take advantage of
230 DEVELOPMENT DURING MIDDLE CHILDHOOD the fact that children of this age at least those beyond third grade are old enough to read questions, write out answers, and fully enter into con- versations with interviewers. Measures derived from child reports and parent reports, however, are fraught with problems of halo effects, inaccurate rec- ollections, and biases that reflect the subjects' wish to present themselves or their children in a favorable light. In recent years, these methods have been increasingly supplemented by observations of parent-child interaction. Observational data are valuable, expanding our knowledge beyond what was available from the interview and paper and pencil measures. Observational methods, however, have hazards. One-shot observations are of doubtful value; tO get stable scores on the characteristics of parents or children, it is necessary tO sample behavior across time and situations, and this is expensive. In addition, in some respects it is more difficult to conduct naturalistic observations of school-age children than with infants and preschoolers. M. Radke-Yarrow (personal communication) reported that when mothers have been trained as observers and dictate into a tape recorder brief summaries of disciplinary encounters with their children, children age 6 or 7 sometimes insist on dictating their own version of the episode, indicating their high level of awareness of being observed. Similar problems are encountered by those attempting to make inconspicuous observations on school playgrounds. It is difficult to prevent children from knowing they are observed. Further- more, much of the free-play interaction of school-age children occurs in settings ant! at times that preclude observation. In-home observations of older children are similarly susceptible to dis- tortion. It is unnatural to require family members to remain in the same room for periods of observation or to set other constraints that may be desirable to researchers (e.g., family members should not read or watch television). Most observational studies have turned to structured problem' solving or teaching-learning kinds of interaction. It may be, however, that direct observations of any sort necessarily lose their ecological validity with the increasing age of the child. It appears that consideration of other as- sessment methods, and ways to improve them, is in order. Some recent data, as yet unpublished, by Jacques Wright indicate that when observational scores are aggregated across a considerable number of occasions or situations or both, the aggregate scores correspond well (cor- relations in the 70s or higher) with the ratings on comparable dimensions made by adults who are familiar with how the children behave in a variety of settings. The ratings by familiar peers are also highly related to aggregated observational scores that reflect how the children actually behave; this re- lation is particularly strong with respect to antisocial behavior (see also Epstein, 1980, on the importance of aggregation). Thus, for research that
CONTEXT OF THE FAMILY 231 seeks to identify children's stable individual characteristics, it may be both necessary and defensible to rely more heavily on ratings, Q-sorts, and nom- ination procedures rather than behavioral observations that are restricted in frequency and cross~situational scope (see also Calms and Green, 19791. Peer nominations are important in assessing the social behavior of school- age children, but their use is constrained by ethical considerations. Re- searchers must not exacerbate the tendency of school-age children to neg- atively label their peers. Considering that each method has known strengths and weaknesses, researchers more often are fuming to multimethod assess- ment batteries for both parent and child. CONCLUSION The first major theme of this chapter is that research on socialization has not been sufficiently developmental in concept. ~ have argued that the middle childhood period has its distinctive patterns of parent-child rela- tionships and its distinctive socialization agenda, both of which need to be understood in terms of the developmental level children have reached by the time they enter this period and the normative developmental changes they undergo as they traverse it. ~ have traced some of these changes and suggested how they might be implicated in the patterns of parent-child interaction that characterize middle childhood. ~ have suggested that some of the traditional variables chosen by students of socialization, such as the frequency of reward or punishment, may not be as appropriate for the middle childhood period as they are for younger ages, and that we must be alert for the emergence of significant new parent-child interaction variables as chil- dren progress through the developmental timetable. Specifically, ~ have suggested that child-rearing shifts as children enter the school-age period, changing from largely face-to-face control, management, and teaching to more distal processes. These processes call for complex cooperation co- regulation-between parent and child. ~ have urged that research should focus on how these coregulation functions are carried out and how they change as children become more competent participants. The second major theme concerns individual differences differences among families in the way they rear their children during the middle childhood years and the possible effects of these variations. ~ have considered the evidence concerning differences in the way families in different social groups function and have discussed a number of viewpoints concerning how these group variations have evolved. While there are some replicable relationships between socioeconomic status and child-rearing, much less is known con- ceming how child-rearing is influenced by other aspects of the sociocultural
232 DEVELOPMENT DURING MIDDLE CHILDHOOD niche in which families function. That we could list many alternative ex- planations for the known variations in childrearing underscores our igno- rance about the causal chain leading from the social and environmental conditions of family life to the Sanctioning of parents to the various devel- opmental paths taken by children. ~ have argued that these causal networks become more complex as children enter the age period when they begin to have some control over the selection and modification of their own envi- ronments. Longitudinal research is needed to clarify the effect of early family interaction patterns on later ones in the development of individual differ ences. REFERENCES Armentrout, V.A., and Burger, G.K. 1972 Children's reports of parental childrearing behaviors at five grade levels. Developmental Psychology 7:44-48. Baldwin, A.L. 1946 Differences in parent behavior toward three' and nine~year~old children. Joumal of Per- sonality 15: 143- 165. 1955 Behavior and Development in Childhood. New York: Dryden Press. Baldwin, A.L., Kalhorn, ]., and Breese, F.H. 1945 Pattems of parent behavior. Psychological Monographs 58:1-75. Bank, S., and Kahn, M.D. 1975 Sisterhood-brotherhood is powerful: Sibling subsystems and family therapy. Family Proc- ess 14:311-337. Barkley, R.A. 1981 The use of psychopharmacology to study reciprocal influences in parent-child interaction. Journal of AbnorTr~1 Child Psychology 9:303-310. Barkley, R.A., and Cunningham, C.C. 1979 The effects of methylphenidate on the mother-child interaction of hyperactive children. Archives of General Psychiatry 36:201-208. Barkley, R.A., Cunningham, C.E., and Karlsson, I. 1983 The speech of hyperactive children and their mothers: Comparison with normal children and stimulant drug effects. Joumal of Learning Disabilities 16: 105- 110. Baumrind, D. 1967 Childcare practices anteceding 3 patterns of preschool behavior. Generic Psychology Mon- ographs 75:43-88. 1971 Current patterns of parental authority. Developmental Psychology Monographs 4(1):part 2. 1973 The development of instrumental competence through socialization. In A.D. Pick, ea., Minnesota Symposium on Child Psychology. Vol. 7. Minneapolis: University of Minnesota Press. Bearison, D. I., and Cassell, T. Z. 1975 Cognitive decentration and social codes: Communication effectiveness in young children from differing family contexts. Developmenm1 Psychology 11:29-36. Becker, W.C. 1964 Consequences of different kinds of parental discipline. In M.L. Hoffman and L.W. Hoffman, eds., Review of Child Development Research.Vol. 1. New York: Russell Sage Foundation.
CONTEXT OF THE FAMILY 233 Bee, H.L., Van Egeren, L.F., Streissguth, A.P., Nyman, B.A., and Leckie, M.S. 1969 Social class differences in maternal teaching strategies and speech patterns. Developmental Psychology 1:726-734. Bell, R.Q. 1977 Research strategies. In R.Q. Bell and L.V. Harper, eds., Child Effects on Adults. Hillsdale, N.J.: Lawrence Erlbaum. Bell, R.Q., and Harper, L.V. 1977 Child Effects on Adults. Hillsdale, Ad.: Lawrence Erlbaum. Bemstein, B. 1961 Social class and linguistic development: A theory of social reaming. In A.H. Halsey, ]. Flavell, and C.A. Anderson, eds., Education, Economy, and Society. New York: Free Press. Block, J. 1971 Lives Through Time. Berkeley: Bancroft Books. Block, J.H., and Block, J. 1980 The role of ego-control and ego-resiliency in the organization of behavior. In W.A. Collins, ed. , Minnesota Symposium on Child Psychology. Vol. 13. Hillsdale, N.J.: Lawrence Erlbaum. Bowlby, J. 1969 Attachment. New York: Basic Books. Bronfenbrenner, U. 1958 Socialization and social class through time and space. In E.E. Maccoby, T.M. Newcomb, and E.L. Hartley, eds., Readings in Social Psychology. New York: Holt. 1979 The Ecology of Human Development. Cambridge, Mass.: Harvard University Press. Brown, A.L., and DeLoache, ].S. 1978 Skills, plans and self~regulation. In R. Siegler, ea., ChildTen's Thinking: What Develops? Hillsdale, N.J.: Lawrence Erlbaum. Bryant, B., and Crockenberg, S. 1980 Correlates and dimensions of prosocial behavior: A study of female siblings with their mothers. Child Development 51:529-544. Burrows, P.E\. 1981 Parent-Child Behavior as a Function of Socio-economic Level and Family Size in Mexican Families. Paper presented at the biennial meeting of the Society for Research in Child Development, Boston, April. Caims, R.B., and Green, ].A. 1979 How to assess personality and social pattems: Observations or ratings? In R.B. Caims, ea., The Analysis of Social Interactions. Appendix A. Hillsdale, N.~.: Lawrence Erlbaum. Chandler, M. 1973 Egocentrism and anti-social behavior: The assessment and training of social-perspective taking skills. Developmental Psychology 9:326-336. Chandler, M., Greenspan, S., and Barenboim, C. 1974 Assessment and training of role-taking and referential communication skills in institu' tionalized emotionally disturbed children. Devek~pmental Psychology 10:546-553. Chapman, M. 1979 Listening tO reason: Children's attentiveness and parental discipline. Me7T§l-PalmeT Queerly 25:25 1-263. Cicirelli, V.G. 1972 Concept learning of young children as a function of sibling relationships to the teacher. Child Development 43:282-287. Clifford, E. 1959 Discipline in the home: A controlled observational study of parental practices. 1Our7uz1 of Genetic Psychology 95:45-82.
234 Cronbach, L.J. DEVELOPMENT DURING MIDDLE CHILDHOOD 1982 Designing Evaluations of Educational and Social Programs. San Francisco: ]ossey-Bass. Damon, W. 1977 The Social World of the Child. San Francisco: ]ossey-Bass. Dodge, K. 1980 Social cognition and children's aggressive behavior. ChiLl Development 51:162-171. Dornbusch, S. M., Carlsmith, ]. M., Bushwall, S.]., Leiderman, H., Hastorf, A. H., Gross, R.T., and Rutter, P. 1983 Single Parents, Extended Households and the Control of Adolescents. Paper read at the Pacific Sociological Meetings, San Jose, Calif., April. Dunn, J. 1983 Sibling relationships in early childhood. Child Development 54:787-811. Dunn, ]., and Kendrick, C. 1982 Siblings: Siblings, Love, Envy and Understanding. Cambridge, Mass.: Harvard University Press. Eiduson, B., Kornfein, M., Zimmerman, I., and Weisner, T. 1982 Comparative socialization practices in traditional and altemative families. In M.E. Lamb, ea., Non-Traditional Families: Parenting and Child Development. Hillsdale, N.].: Lawrence Erlbaum. Elder, G.H. 1974 Children of the Great Depression. Chicago: University of Chicago Press. 1979 Historical changes in life patterns and personality. In P.B. Baltes and O.G. Brim, eds., Life Span Devek~pment and Behavior. Vol. 2. New York: Academic Press. Emmerich, W. 1962 Variations in the parent roles as a function of parents' sex and the child's sex and age. Merrill Palmer Quarterly 8:1-1 1. Entwisle, D.R., and Hayduk, L. 1982 Early School~ng. Baltimore: Johns Hopkins University Press. Epstein, S. 1980 The stability of behavior II. Implications for psychological research. American Psychologist 35:790-806. Farrington, D.P., and West, D.~. 1981 The Cambridge study in delinquent devel`~pment. In S.A. Mednick and A.E. Baert, eds., Prospective Longituclinal Research. Oxford, England: Oxford University Press. Fine, M.A., Moreland, ].R., and Schwebel, A.I. 1983 Long-term effects of divorce on parent-child relationships. Devek~pmental Psychology 19:703- 713. Furstenberg, F.F., Nord, C.W., Peterson, J.L., and Zill, N. 1983 The life course of children of divorce: Marital disruption and parental contact. American Sociological Review 48:656-668. Garmezy, N. 1983 Stressors of childhood. In N. Garmezy and M. Rutter, eds., S`Tess, Coping and Development in Children. New York: McGraw-Hill. Glick, P.C. 1979 The children of divorced parents in demographic perspective. Journal of Social Issues 35:170-182. Goodenough, F.L. 1931 Anger in Young Children. Minneapolis: University of Minnesota Press. Grusec, J.E., and Kuczynski 1980 Direction of effect in socialization: A comparison of the parents' versus the child's behavior as determinants of disciplinary techniques. Developmental Psychology 16:1-9.
CONTEXT OF THE FAMILY Grusec, ].E., and Redler, E. 235 1980 Attribution reinforcement and altruism: A developmental analysis. Developmental Psy- chology 16:525-534. Halverson, C.F., and Waldrop, M.F. 1970 Matemal behavior toward own and other preschool children: The problem of "ownness." Child Development 41:838-845. Hayes, C.D., and Kamerman, S.B., eds. 1983 Children of Working Parents: Experiences and Outcomes. Panel on Work, Family, and Community, Committee on Child Development Research and Public Policy. Washing' ton, D.C.: National Academy Press. Hess, R.D. 1970 Social class and ethnic influences on socialization. In P.H. Mussen, ea., Carmichael's Manual of Child Psychology. Vol. II. New York: John Wiley & Sons. 1981 Approaches to the measurement and interpretation of parent child interaction. In R.W. Henderson, ea., PaTent-Child Interaction: Theory, Research and Prospects. New York: Act ademic Press. Hess, R.D., and Shipman, V.C. 1967 Cognitive elements in maternal behavior. In J.P. Hill, ea., Minnesota Symposium on Child Psychology. Vol. 1. Minneapolis: University of Minnesota Press. Hetherington, E.M., Cox, M., and Cox, R. 1982 Effects of divorce on parents and children. In M. Lamb, ea., Non-Traditional Families: Parenting and Child Development. Hillsdale, N.].: Lawrence Erlbaum. Hill, C.R., and Stafford, F.P. 1980 Parental care of children: Time diary estimate of quantity, predictability and variety. Journal of Human Resources 15:219-239. Hill, M.S. 1983 Trends in the economic situation of U.S. families and children. In R. Nelson and F. Skidmore, eds., American Families and the Economy. Committee on Child Development Research and Public Policy. Washington, D.C.: National Academy Press. Hoffman, L.W. 1979 Maternal employment. American Psychologist 34:859-865. Hoffman, M.L. 1970 Moral development. In P.H. Mussen, ea., Carmichael's Manual of Child Psychology. Vol. 2. New York: John Wiley & Sons. Inkeles, A. 1960 Industrial man: The relation of status to experience, perception and value. American Jounmd of Sociology 66: 1-31. Kamerman, S.B., and Hayes, C.D., eds. 1982 Families That Work: Children in a Changing World. Panel on Work, Family, and Come munity, Committee on Child Development Research and Public Policy. Washington, D.C.: National Academy Press. Kohn, M.L. 1963 Social class and parent-child relationships: An interpretation. American Journal of So- cio~gy 68:471-480. Class and Confer. laity; A Study in Values. Second edition. Chicago: University of Chicago Press. Krauss, R.M., and Glucksberg, S. 1969 The development of communication: Competence as a function of age. Child Development 40:255-266. Kulka, R.A., and Weingarten, H. 1979 The long-term effects of parental divorce in childhood on adult adjustment. Journal of Sock} Issues 35:50-78. 1977
236 DEVELOPMENT DURING MIDDLE CHILDHOOD Lane, D.M., and Pearson, D.A. 1982 The development of selective attention. Mernll-Palmer Quarterly 28:317-337. Laosa, L.M. 1981 Maternal behavior: Sociocultural diversity in modes of family interaction. In R. W. Henderson, ea., Parent-Child lnteracticm: Theory, Research and Prospects. New York: Ac- ademic Press. 1982 School, occupation, culture and family: The impact of parental schooling on the parent' child relationship. Journal of Educational Psychology 74:791-827. Lamb, M., ed. 1981 The Role of the Father in Child Development. Second edition. New York: John Wiley & Sons. 1982 Non-Tradiiional Families: Parenting and Child Development. Hillsdale, Ad.: Lawrence Erl' baum. Lamb, M.E., Frodi, A.M., Hwang, C.P., and Frodi, M. 1982 Varying degrees of paternal involvement in infant care: Attitudinal and behavioral correlates. In M. E. Lamb, ea., Non-Traditional Families: Parenting and Child Development. Hillsdale, N.J.: Lawrence Erlbaum. Lasko, ].K. 1954 Parent behavior toward first- and second-born children. Genetic Psychology Monographs 49:97-137. Lepper, M.R. 1983 Social control processes and the internalization of social values: An attributional per' spective. In E.T. Higgins, D.N. Ruble, and W.W. Hartup, eds., Social Cognition and Social Development. London and New York: Cambridge University Press. LeVine, R.A. 1979 Cross Cultural Research on Child Development. Invited address presented at the meeting of Society for Research in Child Development, San Francisco, March. Loeb, R.C., Horst, L., and Horton, P.J. 1980 Family interaction patterns associated with self-esteem in pre-adolescent girls and boys. Merrill Palmer ~uaTterly 26:203-217. Maccoby, E.E. 1983 Social-emotional development and response to stressors. In N. Garmezy and M. Rutter, eds., Stress, Coping and Development in Children. New York: McGraw-Hill. 1984 Socialization and developmental change. Child Development. April. Maccoby, E.E., and Martin, J.A. 1983 Socialization in the context of the family: Parene-child interaction. In E.M. Hether' ington, ea., Manual of Child Psychology. Paul Mussen, general ed. Vol. 4, fourth edition. New York: John Wiley & Sons. Martin, J.A., Maccoby, E.E., Baran, K.W., and Jacklin, C.N. 1981 Sequential analysis of mother-child interaction at 18 months: A comparison of microan- alytic methods. Developmenm1 Psychology 1 7:146-157. Marvin, R.S. 1977 An eehological-cognitive model for the attenuation of mother-child attachment behavior. In T. Alloway, P. Pliner and L. Krames, eds., Attachment Behavior. New York: Plenum Press. McCord, J. 1979 Some child-rearing antecedents of criminal behavior in adult men. Joumal of Personality and Social Psychology 37:1477-1486. McCord, W., and McCord, J. 1960 Origins of Alcoholism. Stanford, Calif.: Stanford University Press.
CONTEXT OF ITIE FAMILY 237 Medrich, E.A., Roizen, ]., and Rubin, V; 1982 The Serious Business of Growing Up. Berkeley: University of California Press. Moen, P. 1982 The two-provider family: Problems and potentials. In M.E. Lamb, ea., Non-Traditional Families: Parenting and Child Development. Hillsdale, Ad.: Lawrence Erlbaum. Newson, ]., and Newson, E. 1968 FORT Years Old in an Urban Community. Chicago: Aldine. 1976 Seven Years Old in the Home Environment. New York: John Wiley & Sons. 1977 Perspectives on School at Seven years Old. Winchester, Md.: Allen & Unwin. Parke, R.D., and Tinsley, B.R. 1983 Fatherhood: Historical and contemporary perspectives. In K.A. McCluskey and H.W. Reese, eds., Life Span Developmental Psychology: Historical and Cohort Effects. New York: Academic Press. Patterson, G.R. 1982 coercive Family processes. Eugene, Oreg.: Castalia Press. Pearce, D., and McAdcx,, H.P. In Women and Children: Alone and in Poverty. press Peters, M.F. 1976 Nine Black Families: A Study calf Household Management and Child Rearing in Black Families with Working Mothers. Ph.D. dissertation, Harvard University. 1981 Parenting in black families with young children. In H. P. McAdoo, ea., Black Families. Beverly Hills: Sage Publications. Pulkkinen, L. 1982 Self-control and continuity from childhood t<, adolescence. In P. B. Baltes and O.G. Brim, eds., Life-Span Development and Behavior. V`~14. New York: Academic Press. Radke-Yarrow, M., Zahn-Waxler, C., and Chapman, M. 1983 Children's prosocial disp`,siti`,ns and hehavi`'r. In E.M. Hetherington, ea., Manual of Child Psychology. Paul Mussen, general ed. Vol. 4, fourth ed. New York: John Wiley & Sons. Richman, N., Graham, P., and Stevenson, ]. 1982 Preschool to School: A BehaviouTal Study. London: Academic Press. R<,herts, G.C., Block, J.H., and Block, ]. 1981 Continuity and Change in Parents' Child-Rearing Practices. Paper presented at meetings of the Society fair Research in Child Development, Boston. Rollins, L.N. 1974 Deviant Children GTOW UP. Huntington, N.Y.: Robert Krieger. Robins, L.N., and Ratcliff, K.S. 1980 Childhood conduct discarders and later arrest. In L.N. Robins, P.J. Clayton, and l.K. Wing, eds., The Social Consequences calf Psychiatric Illness. New York: Brunner/Mazel. Rowe, D.C., and Plc~min, R. 1981 The importance <,f non-shared environmental influence in hehavi<,ral devek~pment. Developmental Psych<,k~gy 17:517-531. Russell, G. 1982 Shared-caregiving families: An Australian study. In M.E. Lamb, ea., Non-Traditional Families: Parenting and Child Devek~pment. Hillsdale, N.].: Lawrence Erlhaum. Rutter, M. 1971 Parent-child separation: Psych`,k~gical effects on the children. Journal of Child Psychology and Psychiatry 12:233-260. 1979 Maternal deprivation, 1972-1978: New findings, new concepts, new approaches. Child Development 50:283-305.
238 DEVELOPMENT DURING MIDDLE CHILDHOOD 1981 Maternal Deprivation Reassessed. Second edition. Harmondsworth, England, and New York: Penguin Books. 1982 Epidemiological~longitudinal approaches to the study of development. In W.A. Collins, ea., The Concept of Development. Vol. 15: Minnesota Symposium on Child Psychology. Hillsdale, N.].: Lawrence Erlbaum. 1983 Stress, coping and development: Some issues and some questions. In N. Garmezy and M. Rutter, eds., Stress, Coping and Development in Children. N.Y.: McGraw-Hill. Rutter, M., Yule, B., Quinton, D., Rowlands, O., Yule, W., and Berger, M. 1975 Attainment and adjustment in two geographical areas III. Some factors accounting for area differences. British Journal of Psychiatry 126:520-533. Salzstein, H.D. 1976 Social influence and moral development: A perspective on the role of parents and peers. In T. Lickona, ea., Moral Development and Behavior: Theory, Research and Social Issues. New York: Halt, Rinehart & Winston. Samaroff, A.]. 1982 Development and the dialectic: The need for a system approach. In W.A. Collins, ea., Minnesota Symposium on Child Development. Vol. 15, Hillsdale, N.J.: Lawrence Erlbaum. Scarr, S., Webber, P.L., Weinberg, R.A., and Witeig, M.A. 1981 Personality resemblance among adolescents and their parents in biologically related and adoptive families. Journal of Personality and Social Psychology 40:885-898. Schaefer, E.S. 1959 A circumplex model for maternal behavior. Journal of Abno7rrml and Social Psychology 59:226-235. Sears, R.R., Maccoby, E.E., and Levin, H. 1957 Patterns of Child Rearing. Evanston, Ill.: Row, Peterson. Selman, R.L., and Byrne, D.F. 1974 A structural developmental analysis of levels of role taking in middle childhood. Child Development 45:803-806. Sennett, R., and Cobb, ]. 1982 The Hidden Injuries of Class. New York: Random House. Shipman, V.C., McKee, ].D., and Bridgeman, P. 1976 Stability and Change in Family Status, Situational and Process Variables and Their Relation so Children's Cognitive Performance. Princeton, N.].: Educational Testing Service. Sroufe, L.A., and Waters, E. 1977 Attachment as an organizational construct. Child Development 48:1184-1199. Stack, C.B. 1974 Sex roles and survival strategies in an urban black community. In M.Z. Risaldo and L. Lamphere, eds., Women, Culture and Society. Stanford, Calif.: Stanford University Press. Sternberg, R.J., and Powell, ].S. 1983 The development of intelligence. In ]. Flavell and E. Markman, eds., Cognitive Devel- opment, Vol. 3. Manual of Child Psychology, fourth edition. Paul Mussen, general ed. New York: John Wiley & Sons. Stilwell, R. 1983 Social Relationships in Primary School Children as Seen by Children, Mothers, and Teachers. Unpublished Ph. D. dissertation, University of Cambridge. Wallerstein, ].S., and Kelly, ].B. 1980 Surviving the Breakup: How C)uLlren and Parents Cope With Divorce. New York: Basic Books. Weisner, T.S. 1982 As we choose: Family life styles, social class, and compliances. In ].G. Kennedy and R. Edgerton, ea., Culture and Ecology: Eclectic Perspectives. A special publication of the American Anthropological Association, No. 15.
CONTEXT OF THE FAMILY 239 Whalen, C.K., Henker, B., and Dotemoto, S. 1980 Methylphenidate and hyperactivity: Effects on teacher behaviors. Science 208:1280- 1282. 1981 Teacher response to the methylphenidate (Ritalin) versus placebo status of hyperactive boys in the classroom. Journal of Child Development 52:1005-1014. Yankelovich, D., Clark, R., and Martire, G. 1977 General Mills American Family Report. Minneapolis, Minn.: General Mills, Inc. Yarrow, M.R., Scott, P.M., and Waxier, C.Z. 1973 Leaming concern for others. Developmental Psychology, 8:240-260. Young, V. 1970 Family and childhood in a southern Negro community. American Anthropologist 72:269 288. Zill, N. In press 1978 Divorce, Marital Happiness, and the Mental Health of Children: Findings from the FCD National Survey of Children. Paper prepared for NIMH Workshop on Divorce and Children, Bethesda, Md., February. Happy, Healthy and lusecuTe. New York: Cambridge University Press.
