Paul R Amato. Handbook of Contemporary Families. Editor: Marilyn Coleman & Lawrence H Ganong. Sage Publication. 2004.
One of the goals of science social science as well as natural science is to formulate principles and explanations that apply to a wide range of phenomena, irrespective of time and place. Many theorists, for example, assume that the basic principles of children’s learning (reward, punishment, identification, modeling, internalization of social rules) are applicable across cultures and historical periods. According to this view, the basic processes that underlie child socialization are the same, irrespective of whether children grow up in an urban area in the United States in the 21st century, a Puritan colony in New England in the 17th century, or an extended family in India in the 19th century.
Although the attempt to discover general principles of social behavior is a useful goal, social scientists can never be free entirely of historical and cultural context, for two reasons. First, many of the phenomena of interest to social scientists change over time. This statement applies to divorce, which changed from a relatively uncommon form of behavior at the beginning of the 20th century to a relatively common form of behavior by the end of the 20th century. Second, social scientists are influenced by the worldviews of the particular eras in which they are trained and conduct their research. As I show below, the demand for divorce among married couples, the laws regulating divorce, and public attitudes toward divorce shaped the views and research agendas of family scholars during the 20th century.
The purpose of this chapter is to summarize social scientific research about divorce within a social and historical context. This purpose is best achieved by dividing U.S. history into three periods: (a) 1900 to 1960, a time when most family scholars believed that divorce harmed children and that the increasing rate of divorce was a serious social problem; (b) the 1960s through the 1980s, a time when many family scholars viewed children as resilient and redefined divorce as a transition consistent with the goals of resolving destructive marital relationships and maximizing personal happiness; and (c) the 1990s to the present, a time when family scholars moved toward a more complex synthesis of earlier perspectives. These three periods can be distinguished not only in terms of scientific views about divorce but also in terms of the frequency of divorce, the legal regulation of divorce, and public attitudes toward divorce. Although researchers have studied the causes as well as the consequences of marital disruption, I focus most of my discussion in this chapter on research dealing with the effects of divorce on children.
Divorce as a Social Problem: 1900-1960
The Divorce Rate
Data on divorce in the United States are available from 1860 to the present. A useful way to track the frequency of divorce over time is to calculate the refined divorce rate, which is the number of divorces in a given year per every 1,000 married women. The divorce rate increased from about 2 in 1865 to about 4 in 1900. Divorce continued to increase during the early part of the 20th century, reaching about 8 in 1940. The divorce rate spiked sharply upward during World War II (the first half of the 1940s). After the war, the divorce rate declined, and by the end of the 1950s, it was back to where it had been 20 years earlier. Despite the postwar decline in divorce, marital dissolution became increasingly common in American society during the second half of 19th century and the first half of the 20th century.
During this time, divorce rates also responded (involving short-term increases and decreases) to specific events and changing social circumstances. The divorce rate stabilized during the Great Depression of the 1930s, a time when many couples could not afford to divorce. The surge in divorces during World War II indicates that wartime marriages were not very stable, partly because military service separated many husbands and wives for long periods. The 1950s, in contrast, represented a relatively “profamily” period in U.S. history. In addition to the declining divorce rate, the marriage rate was high, and fertility increased dramatically (the baby boom). Several forces came together to create the stable, two-parent, child-oriented family of the 1950s: a strong economy, real increases in men’s wages, veterans taking advantage of the G.I. Bill to obtain university educations, the growth of home construction in the suburbs, and a desire on the part of many people to move beyond the tumultuous war years and to concentrate on home and family life. (See Cherlin, 1992, for a detailed discussion.)
The Legal Regulation of Divorce
During the 19th century and throughout most of the 20th century, divorces were granted only when one spouse demonstrated to the court’s satisfaction that the other spouse was “guilty” of violating the marriage contract. Although divorce laws varied from state to state, grounds for divorce often included infidelity, physical or mental cruelty, and abandonment. Under this system, the “innocent” spouse often received a better deal than the guilty spouse with respect to the division of marital property, alimony, and child custody. In other words, the law “punished” the spouse who was guilty of undermining the marriage (Katz, 1994).
Note that this system did not allow for the possibility that two spouses simply might be unhappy with their marriage and wish to go their separate ways. Instead, by making it difficult to divorce, and by requiring one spouse to accept responsibility for violating the marriage contract, the law affirmed its commitment to the norm of marital permanence. Despite the difficulty and cost of proving fault in court, however, demand for divorce increased throughout this period. To accommodate this growing demand, the courts gradually broadened the grounds for marital dissolution, and an increasing proportion of divorces were granted on the relatively vague grounds of “mental cruelty” (Emery, 1988).
Public Attitudes toward Divorce
Historically, Americans have disapproved of divorce. Few attitudes surveys were conducted before the 1960s. Nevertheless, most religions have discouraged divorce, and Americans are a religious people. The Catholic Church explicitly forbids marital dissolution, although an annulment is possible if couples can demonstrate that reasonable expectations for marriage were not met (for example, the marriage was not consummated). Mainline Protestant denominations, along with Judaism, allow divorce but believe that it should avoided if at all possible. Because many people viewed divorce as a form of deviance (or a sin), divorced individuals were, to a certain extent, stigmatized although the stigma was stronger for women than for men (Kitson, 1992). Before the 1960s, divorce was a campaign liability, even for a man. Adlai Stevenson (who was divorced in 1949) served as the Democratic Party’s candidate for president in 1952. During the campaign, Dwight Eisenhower (the Republican Party candidate) raised Stevenson’s divorce as a campaign issue an issue that resonated strongly among women voters (Rothstein, 2002). Partly for this reason, Stevenson lost the election.
Social Scientific Views of Divorce
Like the general public, early family scholars took a dim view of divorce. Developmental theories of the time, such as Freudian theory, assumed that children needed to grow up with two parents to develop normally. For this reason, most social scientists in the first half of the 20th century saw the rising level of marital disruption as a serious social problem. Sociologists, in particular, were concerned that one of the fundamental institutions of society the family was being undermined. In the 1920s and 1930s, social scientists published books with titles such as The Marriage Crisis (Groves, 1928), Family Disorganization (Mowrer, 1927), and Marriage at the Crossroads (Stekel, 1931). Curiously, few studies focused on children, presumably because the idea that divorce was bad for children seemed self-evident. Instead, researchers focused primarily on factors that promoted or eroded marital happiness (e.g., Burgess & Cottrell, 1939; Terman, 1938).
Ernest Burgess was one of the most influential of these early family scholars (Burgess & Cottrell, 1939; Burgess & Locke, 1945). Burgess argued that the increasing rate of divorce reflected a transformation in the nature of marriage: That is, marriage was changing from a formal social institution to a private arrangement based on companionship. By institution, Burgess meant a fundamental unit of social organization a formal status regulated by social norms, public opinion, law, and religion. According to Burgess, the industrialization and urbanization of the United States were weakening the institutional basis of marriage. As parents, religion, community expectations, and patriarchal traditions exerted less control over individuals, marriages were based increasingly on the mutual affection and individual preferences of spouses. Burgess referred to the new model (and ideal) of marriage as companionate marriage.
Companionate marriages are held together, not by bonds of social obligation, but by ties of love, friendship, and common interest. And unlike institutional marriage, companionate marriage allows for an ample degree of self-expression and personal development. Burgess further argued that in an urbanized society in which interaction takes place largely within the context of impersonal, secondary relationships, companionate marriage becomes the most important source of social and emotional support in people’s lives. Because companionate marriage was central to people’s well-being, and because institutional barriers to divorce were becoming weaker, an increasing number of spouses were electing to leave unsatisfactory marriages to find happiness with alternative partners.
The notion that marriage should be based on mutual affection, sexual attraction, and equality began to gain public acceptance during the first few decades of the 20th century. Psychologists, educators, and social service providers applied these ideas in their professional practice, and it was in this context that marital counseling emerged as a discipline, with its goal being to help couples achieve emotional closeness and sexual satisfaction through improved communication and conflict management. By the 1950s, the great majority of Americans, irrespective of social class, accepted the companionate model of marriage as the cultural ideal (Mintz & Kellogg, 1989). Although Burgess believed that the ascendance of companionate marriage was responsible for the increase in divorce, he also believed that the divorce rate would decline once society adjusted to this new arrangement. In particular, he (and others) believed that a combination of public education and marital counselingbased on emerging social scientific information about marriagewould help couples to achieve stable and happy marriages in the new era. Burgess and his colleagues did not foresee the massive increase in marital disruption that was shortly to occur.
The Divorce Revolution: 1960 through the 1980s
The Divorce Rate
After declining for more than a decade, the divorce rate increased sharply during the 1960s and 1970s. The divorce rate reached a peak in 1980, then declined modestly. Currently, the rate of divorce is about 20, which means that about 2% of all marriages end in divorce every year [(20/1,000) x 100]. Although the 2% figure may seem low, it is based on a single year. By applying duration-specific probabilities of divorce across all the years of marriage, it is possible to project the percentage of marriages that will end in divorce. Using this method, demographers estimate that about one half of first marriages, and about 60% of second marriages, will end in divorce (Cherlin, 1992). These figures represent a historically high level of marital disruption in the United States. In comparison, about one eighth of all marriages ended in divorce in 1900, and about one fourth of all marriages ended in divorce in the 1950s (Preston & McDonald, 1979).
The Legal Regulation of Divorce
During the first half of the 20th century, all U.S. states granted divorces under a fault regime. In 1953, Oklahoma became the first state to allow no-fault divorce (Vlosky & Monroe, 2002). Although fault-based divorces still were possible, spouses could dissolve marriages even if neither spouse had committed a serious marital offense. In these cases, “incompatibility” became the grounds for divorce. Several other states added no-fault options during the next decade, including Alaska in 1963 and New York in 1967 (Vlosky & Monroe, 2002).
The most dramatic change in divorce law occurred in California in 1969. In that year, the California legislature threw out fault-based divorce entirely and replaced it with no-fault divorce. Under the new legislation, only one ground for divorce existed: the marriage was “irretrievably broken” due to “irreconcilable differences” (Glendon, 1989). Moreover, courts in California granted divorces even if one spouse wanted the divorce and the other did nota system known as unilateral no-fault divorce. The assumption underlying unilateral no-fault divorce is that it takes two committed spouses to form a marriage; if one spouse no longer wishes to remain in the relationship, then the marriage is not viable.
Other states quickly followed California’s lead, and by the mid-1980s, no-fault divorce existed in all 50 states. Most states adopted versions of unilateral no-fault divorce. Other states introduced no-fault divorce but only by mutual consent. In Pennsylvania, for example, if one spouse wants a divorce and the other does not, then a no-fault divorce can be obtained only after the spouses are separated for 2 years. Otherwise, the spouse who wants the divorce must prove fault (Schwartz, 1999). Although some states require mutual consent, and although many states have retained fault-based divorce alongside no-fault divorce by mutual consent, most divorces in the United States today take place under unilateral no-fault divorce regimes (Katz, 1994).
States introduced no-fault divorce for several reasons. First, it was widely known that many couples who wished to dissolve their marriages colluded to fabricate grounds for divorce. The recognition that many couples were making a mockery of the law was one factor leading to the reformation of divorce law (Katz, 1994). In addition, legal scholars increasingly accepted the proposition that spouses had a legal right to end their marriages if they were incompatible, had fallen out of love, or were no longer happy living together (Glendon, 1989). Finally, fault-based divorce is an inherently adversarial procedure. Legislators believed that no-fault divorce would lessen the level of animosity between former spouses and hence make it easier for them to cooperate in raising their children following marital dissolution (Glendon, 1989; Katz, 1994).
Some scholars believe that the liberalization of divorce laws stimulated further demand for divorce (e.g., Parkman, 2000). Other scholars disagree with this claim (Glenn, 1999). Even if legal changes encouraged more couples to divorce, however, this effect probably was modest. Changes in the law, as well as changes in divorce rates, were primarily reflections of the broad shift from institutional marriage to companionate marriage, as described earlier (Burgess & Cottrell, 1939; Burgess & Locke, 1945).
Public Attitudes toward Divorce
Public attitudes toward divorce became more liberal during the 1960s and 1970s. For example, between 1965 and 1976, the percentage of respondents who felt that divorce laws were “too strict” increased (Cherlin, 1992). In another study, the percentage of women who disagreed with the statement “When there are children in the family, parents should stay together even if they don’t get along” increased from 51 in 1962 to 80 in 1977 (Thornton & Young-DeMarco, 2001). As noted earlier, Adlai Stevenson’s divorce was a major campaign issue in the 1952 presidential election. In contrast, Ronald Reagan ran successfully for president in 1980, and his opponent, Jimmy Carter, did not raise Reagan’s divorce as a campaign issue (Rothstein, 2002). Indeed, Reagan ran on a profamily platform, despite his history of troubled relationships with his adult children. By the end of the 1970s, the great majority of Americans viewed divorce as an unfortunate but common event, and the stigma of divorce, although still present, was considerably weaker than in earlier eras. The growing acceptance of divorce was consistent with a larger cultural shift that occurred during the 1960s and 1970s. According to Furstenberg and Cherlin (1991, Chap. 6), people’s expectations for personal happiness from marriage increased throughout the 20th century. With personal fulfillment becoming the main criterion by which people judged their marriages, spouses tended to seek divorces when they became unhappy with their relationships, even if the marriage did not include serious problems such as abuse. Moreover, rather than condemning the decision to divorce, friends and family members tended to support spouses who left unsatisfying marriages. This shift in people’s views, and the corresponding decrease in social pressure to stay married, is consistent with Burgess’s notion of the decline of institutional marriage and the rise of companionate marriage, as described earlier. (Also see Bellah, Madsen, Sullivan, Swidler, & Tipton, 1985.)
Social Scientific Research on Children and Divorce
During the 1960s and 1970s, research on the effects of divorce on children proliferated. Congruent with popular opinion, most family scholars in the 1960s assumed that children who grew up without two biological parents in the household were prone to a variety of emotional, behavioral, and academic problems. Studies conducted in the 1960s and early 1970s appeared to support this assumption. Research indicated that children from single-parent households were substantially overrepresented in samples of juvenile delinquents (Glueck & Glueck, 1962) and children with behavioral or emotional disorders (McDermott, 1970).
In 1971, psychologists Judith Wallerstein and Joan Kelly began an influential longitudinal study of 60 families and 131 children (ages 3 to 18). Wallerstein and Kelly published a variety of academic papers during the 1970s that described these families, and they later summarized their results in a widely read book (Wallerstein & Kelly, 1980). According to the authors, 5 years after divorce, one third of children were adjusting well and had good relationships with both parents. Another group of children (more than one third of the sample) were clinically depressed, were doing poorly in school, had difficulty maintaining friendships, experienced chronic problems such as sleep disturbances, and continued to hope for the reconciliation of their parents. A second book published at the end of the 1980s was based on a 10-year follow-up of the sample (Wallerstein & Blakeslee, 1989). In this book, Wallerstein claimed that children with divorced parents often developed into troubled youth who were prone to psychological distress and anxious about forming relationships. As she stated, “Almost half of the children entered adulthood as worried, underachieving, self-deprecating, and sometimes angry young men and women” (p. 299). Although not all children were affected negatively, this study confirmed many people’s fear that divorce was a serious risk factor for long-term child maladjustment.
Although early research supported the belief that divorce was intrinsically bad for children, other studies in the 1970s challenged this dominant view. For example, Mavis Hetherington and her colleagues (Hetherington, 1979; Hetherington, Cox, & Cox, 1982) studied 144 preschool children, half from recently divorced families and half from nondivorced families. During the first year of the study, children with divorced parents exhibited more behavioral and emotional problems than did children with continuously married parents. Two years after divorce, however, most children with divorced parents no longer exhibited an elevated number of problems, although a few problems lingered for boys. Contrary to Wallerstein’s findings, Hetherington’s research suggested that most problems associated with marital disruption improve once family members have time to adjust to their new circumstances. (It is difficult to compare the Hetherington and Wallerstein studies directly. Compared with Hetherington, Wallerstein used a broader age range of children, did not have a comparison group of married-couple families, and used a qualitative rather than quantitative approach.)
During the 1970s and 1980s, an increasing number of family scholars began to question the assumption that divorce inevitably harms children. In an influential review of the literature, Herzog and Sudia (1973) argued that single-parent households (usually headed by single mothers) tend to be poor and that children from poor families are especially likely to drop out of school or become delinquent. Therefore, many of the problems exhibited by children in single-parent families may be due to poverty rather than parental absence. Subsequent studies provided support for this conclusion. For example, in a national study, Guidobaldi, Cleminshaw, Perry, and McLaughlin (1983) found that without controls for income, children in divorced families scored significantly lower than children in nondivorced families on 27 out of 34 outcomes. Controlling for income, however, reduced the number of significant differences to 13, which suggested that income accounted for most of the differences between children with divorced and continuously married parents.
In another widely cited review, Robert Emery (1982) pointed out that marital conflict among continuously married parents has negative effects on children comparable to those attributed to parental divorce. Given that marital conflict precedes divorce, it is plausible that much of the apparent “effect” of divorce on children is due to interparental discord rather than parental absence. This view was consistent with research showing that children with continuously married, discordant parents have as manyor more problems than children with divorced parents (Long, 1986). Also consistent with this explanation were studies showing that children’s well-being is negatively related to the level of conflict between parents following divorce (Guidobaldi, Perry, & Nastasi, 1987; Kurdek & Berg, 1983).
Using meta-analytic techniques, Amato and Keith (1991) summarized 92 studies of children and divorce that had been published between 1953 and the end of the 1980s. Their meta-analysis confirmed that children with divorced parents, compared to children with continuously married parents, scored lower (on average) on measures of academic achievement, conduct, psychological well-being, self-esteem, and social competence. However, the effect sizes in this literature were small rather than large, with the median effect size being 0.14 of a standard deviation. An effect size of this magnitude indicates a substantial degree of overlap between children with divorced and continuously married parents, with many children in the former group doing well and many children in the latter group doing poorly. The meta-analysis also found that effect sizes tended to be smaller in more recent studies, which suggested that the effects of divorce were becoming weaker over time.
By the late 1970s and the 1980s, many family scholars had adopted the position that research on divorce had relied too heavily on a family deficit perspective (Demo & Acock, 1988; Marotz-Baden, Adams, Buech, Munro, & Munro, 1979). According to these critics, the deficit perspective inaccurately assumed that any departure from the nuclear family is deviant and therefore deleterious to children’s well-being. Although existing research suggested that children with divorced parents are at risk for a variety of problems, most of these studies were weak methodologically, the differences between children with divorced and continuously married parents were small in magnitude, and mostperhaps allof these differences could be accounted for by factors such as low income and interparental discord before and after divorce.
In contrast to the deficit perspective, these scholars advocated a framework that I refer to as a family pluralism perspective. This new perspective assumed that family structure has few intrinsic consequences (negative or positive) for children’s well-being and that children can develop successfully in a variety of family forms. Even if divorce created problems for children in the past, these negative effects have declined because society no longer harshly stigmatizes children with single parents. And although divorce may be a stressful experience, most children are resilient (rather than vulnerable) and are capable of adjusting well after a period of time. According to advocates of this new perspective, family processes (such as the level of marital conflict and the closeness of parent-child relationships) are more important than family structure in understanding children’s development and adjustment. Consequently, rather than “privilege” the traditional two-parent family and stigmatize divorced parents and their children, advocates of family pluralism emphasized the strengths of all types of families.
The perspective of family pluralism, by breaking with the notion that single-parent families have uniformly harmful effects on children, provided a good fit with the historical eraan era in which half of all first marriages ended in divorce, the general public had become tolerant of divorce (including divorces involving children), and the legal system had made it easier for couples to dissolve their marriages. Family pluralism acknowledged that divorce was a necessary safety valve for individuals in marriages marked by violence, substance abuse, and destructive behavior. Moreover, the pluralism perspective suggested that it was neither necessary nor desirable to sacrifice personal happiness for the sake of an unsatisfying marriage. According to this perspective, the provision of developmentally appropriate levels of warmth and supervision by competent, well-adjusted, and happy parents was the key to children’s well-being. Consequently, children were likely to be better offrather than worse offwhen unhappily married parents divorced to find greater happiness elsewhere.
The Emergence of a Middle Ground: 1990 to the Present
The Divorce Rate
After reaching a peak in the early 1980s, the divorce rate began to decline. At first, it was not clear whether this change represented a temporary fluctuation or the beginning of a long-term trend. However, the decline in divorce persisted throughout the 1980s and well into the 1990s, albeit modestly.
The Legal Regulation of Divorce
During the 1990s, the Marriage Movementa loosely aligned group of academics, social scientists, religious leaders, legislators, family therapists, and members of the publicemerged in the United States. This movement was diverse politically, with liberal as well as conservative elements. What connected these individuals was a concern that the institution of marriage in the United States had become weaker, with detrimental consequences for children, adults, and society in general. (See Blanken-horn, Bayme, & Elshtain, 1990; Council on Families in America, 1995; Ooms, 2002; Parke & Ooms, 2002, for varied perspectives on the Marriage Movement.)
Some conservative members of this groups adopted the goal of divorce law reform and attempted to lower the frequency of marital dissolution by placing restrictions on unilateral no-fault divorce. During the 1990s, legislators in nearly a dozen states introduced bills that would require the consent of both spouses for a no-fault divorce if they had dependent children. Fault-based divorce would be available without mutual consent in cases of abuse, desertion, or adultery. Other bills attempted to lengthen the waiting period before divorce or to require marital counseling (with the goal of attempting a reconciliation) before divorce. Despite attracting a good deal of media attention, none of these bills was passed in any state (Crouch, 2002).
Divorce law reformers were more successful with the introduction of covenant marriage in three states. Louisiana was the first state to introduce this legislation in 1997, followed by Arizona and Arkansas. Under this system, couples choose between two types of marriage: a standard marriage or a covenant marriage. To obtain a covenant marriage, couples must attend premarital education classes and promise to seek marital counseling to preserve the marriage if problems arise later. Unilateral no-fault divorce is not an option for ending a covenant marriage. To terminate a covenant marriage, one spouse must prove fault, although a couple also can obtain a divorce after a 2-year separation. (For a detailed description of covenant marriage in Louisiana, see Thompson & Wyatt, 1999). Although proponents of covenant marriage see it as a way to strengthen marriage and lower the divorce rate, only a small percentage of couples in Louisiana have chosen covenant marriages (Sanchez, Nock, Wright, Pardee, & Ionescu, 2001).
The impact of the divorce law reform movement has been modest, but other legislative efforts to strengthen marriage (rather than restrict access to divorce) have met with more success. Although it was not widely recognized at the time, the 1996 federal welfare reform legislation referred to promoting marriage and encouraging the formation and maintenance of two-parent families as explicit policy goals (Ooms, 2002). Since that time, several state governments have enacted legislation and programs to strengthen marriage and reduce divorce. For example, Oklahoma provides publicly funded premarital education classes to a wide range of couples, including poor, unmarried parents. Florida decreased the fee for a marriage license, along with the waiting period between obtaining a license and getting married, for couples who have taken a premarital education course. Florida also requires all high school students to take a course on marriage skills. In Arizona, Florida, and Utah, couples are given marriage materials (booklets or videos) that include information on how to build strong marriages, the effects of divorce on children, and available community resources. (See Parke & Ooms, 2002, for details on these policies.) Studies to assess the effectiveness of these interventions, however, have not been conducted.
Public Attitudes toward Divorce
Contrary to trends in the 1960s and 1970s, American’s attitudes toward divorce were less positive at the end of the 20th century. Two national surveys, one carried out in 1980 (Booth, Johnson, White, & Edwards, 1981) and the other carried out in 2000 (Amato, Johnson, Booth, & Rogers, 2003), provide the latest information on this issue. These two independent samples each involved about 2,000 married individuals. Agreement with the statement “Couples are able to get divorced too easily today” increased from 33% in 1980 to 47% in 2000. Correspondingly, agreement with the statement “The personal happiness of an individual is more important than putting up with a bad marriage” declined from 74% in 1980 to 64% in 2000. Overall, a scale based on these and other items (scored to reflect support for the norm of lifelong marriage) increased by more than one third of a standard deviation during this 20-year period. This strengthening of attitudes in support of lifelong marriage was apparent for wives as well as husbands.
Despite this shift in attitudes, it would be a mistake to assume that the American public is against divorce. Instead, people appear to be deeply ambivalent. For example, in the 2000 survey (Amato et al., 2003), only a minority of people (17%) agreed with the statement “It’s okay for people to get married thinking that if it does not work out, they can always get a divorce.” But at the same time, as noted earlier, the majority of people believe that personal happiness is more important than remaining in an unhappy marriage. Similarly, a poll by Time Magazine in 2000 found that 66% of people believed that children are better off after “a divorce in which the parents are more happy” than in “an unhappy marriage in which parents stay together mainly for the kids.” At the same time, however, 64% of people believed that children always or frequently are harmed when parents get divorced (Kirn, 2000).
Social Scientific Research on Children and Divorce
During the last decade of the 20th century, research on divorce became considerably more sophisticated. Compared with studies in prior decades, studies conducted in the 1990s were more likely to rely on large, randomly selected samples; to employ control variables to decrease the likelihood of observing spurious associations; to use psychometrically sound measures; and to include multiple indicators of child outcomes (see Amato, 2001, for a review). These methodological improvements were made possible partly by the public release of large, federally funded, longitudinal studies of families, such as the National Survey of Families and Households, the National Longitudinal Survey of Youth (Mother and Child Supplement), and the National Adolescent Health Study. The availability of these data sets made it possible for researchers to employ multiple respondents and multiple waves of data and to design studies in which children are assessed before as well as after parental divorce. (Although some large data sets, such as the National Survey of Children, were collected as early as the 1970s, this was the exception rather than the rule.) The widespread use of new statistical methods, such as structural equation modeling, pooled time-series analysis, growth curve analysis, and hierarchical linear modeling, complemented the availability of more complex and sophisticated data.
The scholarly literature on divorce since 1990 parallels, in certain respects, the increasing concern with marital stability along with an increase in ambivalence among policy makers and the general public. Some scholars argued, on the basis of research conducted mainly in the late 1980s and 1990s, that married individuals are healthier and happier than single individuals (see Waite, 1995, for a review). Similarly, scholars argued that children are better off in two-parent families than in divorced families (Popenoe, 1996; Wallerstein, Lewis, & Blakeslee, 2001). Other scholars contested this claim, arguing that research continues to reflect an unwarranted, ideologically driven bias in favor of marriage and traditional nuclear families (Coontz, 1992; Demo, 1992; Stacey, 1996). This debate became one of the central (and most interesting) features of family studies in the 1990s.
In the midst of this debate, a growing body of evidence suggested that the deficit perspective and the family pluralism perspective both represented simplified, one-sided accentuations of reality (Cherlin, 1999). One of the central assumptions of the pluralism perspective is that the negative effects of marital disruption on children are only temporary. Contrary to this assumption, however, studies since 1990 have demonstrated that parental divorce is associated with a variety of problems for adult offspring, including chronic psychological distress (Cherlin, Chase-Lansdale, & McRae, 1998), low socioeconomic attainment (McLanahan & Sandefur, 1994), seeing one’s own marriage end in divorce (Bumpass, Martin, & Sweet, 1991), and having weak ties with parents (Amato & Booth, 1997). A second assumption is that the apparent “effects” of divorce are due to unmeasured variables, such as poverty or the disturbed family relationships that precede divorce. However, research in the 1990s demonstrated that the links between parental divorce and child problems persist even after parents’ social class, parents’ marital happiness, parents’ personality characteristics, parents’ attitudes and religiosity, and genetic factors are controlled for. Although it is possible that some as yet unmeasured factor is responsible for parental divorce as well as problems among children, most family scholars now agree that divorce has at least some negative implications for children (see Amato, 2000, for a review). A third assumption is that the effects of divorce on children have become weaker (or even nonexistent) as divorce has become more frequent. This assumption was called into question by a recent meta-analysis, which indicated that the estimated effects of divorce on children were stronger (rather than weaker) in the 1990s than in the 1980s (Amato, 2001). Although the explanation for this increase is unclear, this finding indicates that the gap in well-being between children with divorced and continuously married parents is not likely to go away soon.
Now consider the family deficit perspective. The central assumption of this perspective is that deviations from the nuclear family result in serious problems for the great majority of children. Contrary to this assumption, however, the problems associated with divorce do not appear to be shared by allor even mostchildren. For example, Hetherington (Hetherington & Kelly, 2002), in summarizing nearly 30 years of results from her longitudinal studies, estimated that 20% to 25% of children with divorced parents reach adulthood with significant problems in psychological or social functioning, compared with about 10% of children with continuously married parents. Similarly, Amato (in press) estimated that parental divorce leads about 10% of offspring to have lower levels of psychological well-being in adulthood than they would have had if their parents had remained married. These results suggest that although marital disruption results in long-term problems for a significant minority of children, it does not condemn the majority of children to a lifetime of misery.
Given the limitations of these two perspectives, most studies published in the 1990s have focused not on whether divorce is harmful for children but on the mechanisms through which divorce affects children, as well as the circumstances under which divorce can have harmful, neutral, or positive consequences for children. A large number of studies have shown that divorce affects children through a few key mechanisms: (a) The stress of divorce tends to disrupt the quality of parenting from custodial parents; (b) living in a single-parent household often undermines the quality of relations with noncustodial parents; (c) divorce typically is followed by a decline in household income; (d) divorce tends to exacerbate conflict between parents, causing many children to feel “caught in the middle” between warring parents; and (e) divorce frequently is followed by other stressful events for children, such as moving, parental remarriage (which children often oppose), and additional parental divorces. Other studies show that protective factors (such as social support from peers, close relationships with grandparents and other involved adults, active coping styles on the part of children, and the availability of therapeutic interventions) can buffer the negative effects of divorce (see Amato, 2000, for a review).
The delineation of these mechanisms helps us to understand why the consequences of divorce vary from one child to the next. For example, when divorce is followed by inept parenting on the part of the custodial parent, a loss of contact with the noncustodial parent, a substantial decline in household income, additional conflict between parents, and a period of residential instability, then divorce is likely to have harmful effects on children. In contrast, when divorce is followed by authoritative parenting on the part of the custodial parent, a close relationship with the noncustodial parent, little or no decline in household income, a cooperative relationship between parents, and few additional stressful transitions, then divorce is likely to have minimal effects on children. When divorce is followed by a mix of these circumstances, then divorce is likely to have mixed effects on children as well.
Another line of research is pertinent to this issue. A series of studies have shown that when parents exhibit high levels of chronic, overt conflict (frequent loud arguments, expressions of hostility, and violence), children are better off, in the long run, if parents divorce rather than remain married. If parents exhibit relatively little overt conflict, however, then children are better off if parents remain married, even if parents are dissatisfied with the relationship (Booth & Amato, 2001; Hanson, 1999; Jekielek, 1998). These studies are of interest because they reveal conditions under which divorce is a better alternative for children than remaining in a two-parent family. Using longitudinal data, Amato (2002) estimated that about 40% of marriages that end in divorce fit this pattern.
Overall, research in 1990 led to a growing consensus among family scholarsa consensus that represents a middle ground between those who view divorce as being a catastrophic event (the deficit perspective) and those who view divorce as a relatively benign, transitory event (the pluralism perspective). This emerging perspective which I refer to as a contingency perspective sees divorce as a stressor and a risk factor for subsequent academic, behavioral, emotional, and social problems. However, depending on family circumstances before and after divorce, and depending on the presence of a variety of protective factors in children’s environments, divorce may be harmful or benign to children (Amato, 2000). Moreover, research in the 1990s consistently demonstrated that children do best when they grow up with two happily and continuously married parents (Amato & Booth, 1997; Cherlin, 1999; Hetherington & Kelly, 2002). Consequently, it is necessary to take into account family structure (divorce) as well as family process (the quality of family relationships) to understand children’s long-term development and well-being. In this sense, the contingency perspective of the 1990s represented a synthesis of the earlier deficit and pluralism perspectives.
Proponents of this new perspective value the benefits of marriage and are wary of the risks associated with divorce, but they also recognize that some marriages are harmful and some divorces are beneficial. This perspective is congruent with social trends in the 1990sa decade in which divorce rates declined modestly, policy interventions attempted to improve marital quality and stability, and public attitudes became increasingly supportive of the norm of lifelong marriage (but also continued to value the freedom to leave troubled marriages). The dissemination of research findings on the value of strong marriages for children and adults may have influenced public attitudes toward divorce and helped to propel legislative changes to support marriage. At the same time, the growing respect for marriage and the declining status of divorce in the popular culture and in political circles made it easier for family scholars to combine ideas from the deficit and pluralism models into a new synthesis.
During the 20th century, the frequency of divorce, the legal regulation of divorce, the public’s view of divorce, and social scientific conclusions about divorce changed in ways that were consistent and mutually reinforcing. Before the 1960s, divorce was relatively uncommon, laws in every state made it difficult to divorce, and the general public disapproved of divorce. Correspondingly, most social scientists believed that divorce was harmful to children, and they viewed the increase in divorce as a serious problem. During the 1960s, 1970s, and 1980s, the divorce rate increased dramatically, the legal system made divorce easier to obtain, and the general public became more accepting of divorce. These changes created a context within which family scholars could move beyond the limitations of the deficit model and see marital dissolution in a more favorable light. In the 1990s, the divorce rate declined, state governments enacted programs to strengthen marriage, and support for the norm of lifelong marriage increased among the public. Correspondingly, research in the 1990s repeatedly demonstrated that divorce can be harmful under some circumstances and that children develop best when they grow up with two continuously and happily married parents. It is difficult to predict the next stage of social change, but it seems likely that the present period will continue for some time, as social scientists, legislators, the media, and the general public continue to sort out the roles of marriage and divorce in the American family system.
Where is research on children and divorce headed? Two improvements seem likely. First, future research will become more sophisticated methodologically. The growing availability of national, longitudinal studies of children and families, innovations in statistical modeling, and the development of new measurement techniques will provide a clearer understanding of how marital dissolution affects children over the life course. Second, future research will become more sophisticated theoretically. Most scholars now recognize that asking whether divorce harms of helps children is too simple a question. Instead, future research will make progress in delineating the mechanisms through which divorce affects children and in specifying the conditions under which marital disruption has varied consequences for children. Moreover, an increasing percentage of children are born outside marriage mainly to parents who are either cohabiting or in romantic relationships. Because most of these parents are poor, and because these parents are at high risk for relationship dissolution, researchers refer to these unions as “fragile families” (McLanahan, Garfinkel, Reichman, & Teitler, 2002). Yet we know little about how these informal family arrangements affect children. In keeping with new demographic realities, future researchers will need to broaden their conceptual and operational definitions of “two-parent families” and “divorce” if they are to understand how these trends are shaping children’s lives.
In conclusion, social research does not take place in a cultural and historical vacuum. Instead, family scholarship both shapes and is shaped by events and processes in the larger society. This conclusion does not mean that social research is incapable of generating objective, factual knowledge. Indeed, family research has provided a great deal of useful information about the circumstances under which divorce disadvantages children. Instead, this conclusion indicates that family scholars draw on current worldviews to challenge, refine, and synthesize the knowledge and perspectives of the past. Social scientific information about families should be seen, not as something independent of culture and history, but as part of our culture and history.