Marital Dissolution

Abstract

Marital dissolution or divorce is one of the most dramatic events that can befall a family. Divorce is also relatively ubiquitous in the Western hemisphere. In this chapter, first, we present the statistical picture, putting today’s situation in historical context. Next, we explore the antecedents or predictors of divorce, distinguishing between “macro” level (factors that move the culture as a whole toward greater or lower rates of divorce) and “micro” level (factors that move specific couples to divorce) influences. We also briefly review theories regarding the ways in which these factors influence divorce. Then we turn to the consequences of divorce for mothers and fathers, distinguishing between the legal, behavioral, emotional, social, health, and financial consequences. Then we review the effects of divorce on the children involved, noting how the parents’ responses often modulate the impact on children. Because divorce is so fully intertwined with the legal system, we next discuss the legal issues and processes involved when parents divorce. Our final sections cover, respectively, issues, processes, and policies under current debate; methodological issues arising in the study of divorcing parents and a concluding section.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Marital dissolution or divorce is one of the most dramatic events that can befall a family: Hardly anyone who has undergone a divorce regards it as “just another” transition or event. For many, in fact, their divorces or their parents’ divorces are life-defining events, around which all other experiences are organized: “before the divorce” vs. “after the divorce.”

Divorce is also relatively ubiquitous in the Western hemisphere. As shown in the next section, divorce today is so commonplace that even those who are not directly affected by divorce invariably know families and individuals who are so affected. That this reality is a relatively recent phenomenon, largely confined to the last 5 decades, is not always appreciated.

This chapter will largely focus on divorcing families in the United States that include children. There is a separate and largely nonoverlapping literature on childless dissolutions not summarized here (for space reasons); that literature shows that the effects tend to be milder and of shorter duration (Masheter, 1991; Metts & Cupach, 1995) than those described here. In the sections that follow, first, we present the statistical picture, putting today’s situation in historical context. Next, we explore the antecedents or predictors of divorce, distinguishing between “macro” level (factors that move the culture as a whole toward greater or lower rates of divorce) and “micro” level (factors that move specific couples to divorce) influences. Macro-level factors are of greatest interest to sociologists, demographers, economists, historians, policy specialists, and legal scholars whereas micro-level factors are of primary interest to psychologists, family scholars, and therapists. Because this Handbook addresses an interdisciplinary audience, it is appropriate (though unusual) that we consider both sets of antecedents. Consistent with another goal of this Handbook, we briefly review theories regarding the ways in which these factors influence divorce. Then we turn to the consequences of divorce for mothers and fathers, distinguishing between the legal, behavioral, emotional, social, health, and financial consequences. Then we review the effects of divorce on the children involved, noting how the parents’ responses often modulate the impact on children. Because divorce is so fully intertwined with the legal system, we next discuss the legal issues and processes involved when parents divorce. Our final sections cover, respectively, issues, processes and policies under current debate; methodological issues arising in the study of divorcing parents and a concluding section.

Divorce in Historical Perspective

At the founding of America, divorce was extremely rare: only nine divorces occurred in the entire 72-year history of the Plymouth colony (Riley, 1991). It was not until after the Revolutionary War that courts, as opposed to legislatures, took jurisdiction over divorce. Once that occurred, more reliable record-keeping began. There are several indices frequently used to track divorce prevalence: (1) the absolute number of divorces granted; (2) the “crude rate” (i.e., the number of divorces per 1,000 people); (3) the divorce-to-marriage rate (i.e., number of divorces divided by the number of marriage licenses granted); (4) and the “refined rate” (i.e., the number of divorces per 1,000 married women over 15). The latter is the index most demographers prefer, because it is the closest to the index (5, i.e., the “risk rate”) that would be most informative (i.e., what percentage of married couples eventually get divorced). However, short of tracking each individual couple, the latter isn’t readily calculable and can only be estimated by making assumptions. The risk rate for women born between 1948 and 1950 is estimated at 42 % (Bumpass, Martin, & Sweet, 1991; Schoen & Standish, 2001; Teachman, Tedrow, & Hall, 2006).

Figure 21.1 shows the refined divorce rate (index 4) from 1870 to 2000. Fewer than 2 marriages per 1,000 married women ended in divorce in 1870; the figure rose to 18 per 1,000 married women in 2000, with several obvious secular trends over the period. There was an upward “blip” following WWI, as well as a huge but very temporary spike during WWII; both are usually explained as responses to the hardships placed on marriage by the vicissitudes of war. But the pattern from the late 1960s to the late 1970s is the one of greatest interest to contemporary family scholars; during that decade, the divorce rate doubled (Michael, 1988). Since then, the rate has drifted rather steadily but gradually downward. Possible explanations of recent patterns need to account for both of these tendencies.

Fig. 21.1

Divorces per 1,000 married women aged 15 and older by year, 1870–2000 (adapted from Jacobson, 1959; Preston & McDonald, 1979)

Antecedents, Causes, or Predictors of Divorce

Macro Level Antecedents

Scholars have explored four groups of factors to account for the divorce rate trends starting around 1968: demographic, legal, economic, and attitudinal/cultural.

Demographic factors. Several demographic trends are related to divorce rates. For example, the older the age at marriage, the less likely the couple is to divorce (Thornton & Young-DeMarco, 2001). Because couples have recently waited longer to marry, this may account for the leveling of the divorce rate in recent years; indeed, Heaton (2002), found that this factor can account fully for the divorce trends of the 1975–1995 period. However, it fails to account for the dramatic earlier increases. Similarly, educational attainment is associated (negatively) with divorce (Bumpass et al., 1991; Heaton, 2002). The tendency for average levels of education to increase steadily over time corresponds to the reduction in divorce rates since 1980, but does not explain the increase during the 1970s, when education levels increased as well.

Legal factors. One of the most obvious factors that coincided with the staggering increase in the divorce rate was the liberalization of divorce laws. Prior to 1969, couples seeking divorce had to prove that one spouse had violated the marriage contract. In that year, the no-fault/unilateral divorce movement began in California and was emulated across the nation in the next decade. Thereafter, one and only one spouse had merely to proclaim the marriage “irretrievably broken” for the divorce to be granted (Amato & Irving, 2006). Note that this timing coincided almost perfectly with the increase in the divorce rate, leading some (e.g., Allen, 1992; Nakonezny, Shull, & Rodgers, 1995; Parkman, 2000) to conclude that liberalization of the divorce code was the primary reason for the increase in divorce. Although this claim makes some intuitive sense (of course, divorce rates increase when divorces are easier to obtain), it ultimately must be rejected for two reasons. First, it does not account for the more recent downturn in divorce (the laws remain the same now as they were in the 1970s; Wolfers, 2006); second, more careful analysis, including state-by-state calculations (Ellman & Lohr, 1998; Peters, 1992) suggest that changes in divorce laws tended to follow, rather than lead, increases in the divorce rate. Thus, during the critical decade something else appears to have caused both the increase in divorce and the passage of legislation facilitating divorce.

Economic factors. One such likely candidate variable is female labor force participation. The percentage of women with school age children working outside the home increased dramatically, from 28 % in 1950 to 68 % in 1986 (Hochschild & Machung, 1989). In 1970, for the first time in history, a majority of women were employed (Bergmann, 1986). Having independent incomes allowed women who were unhappy in their marriages to contemplate divorce. Schoen, Astone, Rothert, Standish, and Kim (2002) found that wives’ incomes were indeed linked to divorce—but only for marriages that were unhappy. Nonetheless, female labor force participation fails to account fully for changing divorce rates, because it increased gently, year by year, before and after the critical decade, whereas divorce rates increased exponentially only after 1968 (Michael, 1988). Further, the percentage of women working outside the home has continued to increase from the 1980s until today, while the divorce rate has declined over that period.

Cultural/attitudinal factors. The final category of variables often implicated by scholars involves cultural and attitudinal factors. There is little doubt that the public generally became more accepting of divorce during the late 1960s and 1970s (Thornton, 1989). Commenting on such features of the “me first” generation (Bellah, Madsen, Sullivan, Swidler, & Tipton, 1985), Amato and Irving (2006) observed that “American culture has become increasingly individualistic, and people have become inordinately preoccupied with the pursuit of personal happiness. Because people no longer wish to be hampered with obligations to others, commitment to traditional institutions that require these obligations, such as marriage, has eroded” (p. 51). These normative changes, perhaps better than the other categories of variables reviewed above, appear to account for both the dramatic rise and the mild decline in divorce over the last 50 years.

However, two questions need further explication: (1) what factors precipitated these changes in values; and (2) to what extent were the changes pervasive and universal vs. specific to certain demographic groups or regions? Addressing such issues definitively is difficult, of course, because quantifying precise features of cultural changes over time is a daunting task.

The first author (working with Jenessa Shapiro) hit upon a promising device to elucidate these issues. Reasoning that popular magazines both reflected and galvanized distinctive cultural views, Braver and Shapiro speculated that subscription rates to certain of these magazines across times and locales could provide an empirical window onto these trends. They thus obtained state-by-state, year-by-year subscription data for the following four magazines: Lady’s Home Journal (read almost entirely by women with fairly traditional values and interests); Playboy (glorifying male hedonism); Cosmopolitan (representing lifestyle advice for “fun, fearless females” seeking empowerment, self-improvement, and sexual fulfillment); and Ms. (representing the feminist ideology closely associated with the Women’s Movement). Arraying these subscription rates in a multi-level model against the year-by-year, state-by-state (crude) divorce rates, Braver and Shapiro found (in results not previously published) that changes in divorce rates at the state level were well matched by the state’s trends in subscriptions to Ms. Magazine, and were opposite (though not significantly) to its trends in subscriptions to Lady’s Home Journal. Importantly, they found virtually no association between the state’s divorce rate and its subscriptions to Playboy or Cosmopolitan. Taken together, these data provocatively suggest that some, but not all, value changes are associated with changes in divorce rates. Changing levels of interest in the aspects of self-fulfillment and self-empowerment captured in feminist ideology and its antithesis seemed important, whereas changes in the emphasis on either male-specialized or female-specific self-fulfillment and hedonism were unrelated to divorce trends. Strengthening this interpretation, since around 1970, about 2 in 3 divorces have been sought by wives (Ahrons & Rodgers, 1987; Braver, Whitley, & Ng, 1993; Pettit & Bloom, 1984), whereas previously, “most divorces were the man’s idea” (DeWitt, 1992, p. 54). In addition, findings show that the more that an individual woman agrees with the precepts of the Woman’s Movement, the more likely she is to divorce (Finlay, Starnes, & Alvarez, 1985).

Micro-level Antecedents

Whatever the divorce rate at a particular time, some couples divorce and others do not. There has been extensive research on the micro-level factors that predict divorce (Rodrigues, Hall, & Fincham, 2006; White, 1990). In addition to those demographic variables mentioned earlier (age at marriage and education levels), race is associated with the risk of divorce: African-American and interracial couples are more likely to divorce than Anglo-American couples (Bramlett & Mosher, 2002; Heaton, 2002). Having lived together before marriage is another risk factor (Bumpass et al., 1991; Heaton, 2002) as is having divorced parents (Amato, 1996; Wolfinger, 1999, 2000). This “intergenerational transmission of divorce” seems best explained by the relatively weak commitment to marriage on the part of offspring with divorced parents (Amato & DeBoer, 2001). Belonging to certain religions is associated with reduced risk as well (Bramlett & Mosher, 2002), especially when individuals are highly religious, in whatever faith (Bramlett & Mosher, 2002; Mahoney et al., 2001). Generally, income reduces the risk of divorce (Kurdek, 1993), but as wives earn more, and account for a greater proportion of family income, the risk of divorce increases (Rogers, 2004). A qualification to this pattern are results in a study by Sayer and Bianchi (2000) that wives’ financial independence predicted divorce only if they were unsatisfied in their marriages. Again, we note that the lack of income that once restrained unhappily married women from divorce has become weakened by their collective economic advances. Personality factors, especially neuroticism, are also associated with an elevated risk for divorce (Karney & Bradbury, 1995; Kelly & Conley, 1987). A study by Kiernan (1986) reported, for example, that neuroticism measured in adolescence predicted women’s divorce by age 32.

Notwithstanding these demographic factors, the most important micro-level predictors clearly involve features of the marital relationships themselves. Before describing the empirical findings, however, we note several relevant theories.

Social exchange theory. The social exchange theory of divorce proposes that couples are constantly (perhaps subconsciously) evaluating their marriages (and other relationships). They evaluate relationships in terms of reward-cost ratios to either make a “profit” or at least reduce their perceived losses. If their perceived costs become too high and continual losses are experienced, then divorce becomes more likely. Spouses are also more likely to choose divorce when the barriers to divorcing are lower (i.e., resulting in reduced costs) and the alternatives to staying married seems more attractive (i.e., the alternative relationships or circumstances—even being single—are either more rewarding or less costly than the current marital relationship) (Levinger, 1979). Nobel Prize winner Gary Becker (1993) has formulated an entirely economic or rational choice theory of marriage positing that men and women “attempt to maximize their utility by comparing benefits and costs.... they divorce if that is expected to increase their welfare” (Becker, pp. 395–396).

Vulnerability—stress—adaptation. From this perspective, three factors are highlighted which combine to impact likelihood of divorce: (1) Enduring vulnerabilities—the attributes such as personality and personal background that each spouse brings to the marriage; (2) stressful events with which the partners must cope; and (3) adaptation processes—experiences during the marriage describing the efforts to deal with stress (Karney & Bradbury, 1995). For example, if the couple has disparate views about spending (enduring vulnerabilities), their differences might not cause marital difficulties, until and unless they experience financial setbacks (stressful events). In that event, if they have good conflict resolution skills (adaptation processes), the strength of the marriage might not be affected; But vulnerabilities plus stressful events plus poor adaptation processes are hypothesized to decrease marital quality, which in turn increases marital instability.

Cascade theory. Gottman and colleagues have engaged in extensive research observing, videotaping, and coding couples discussing their disagreements; from this, they have developed a comprehensive theory of the “trajectory to divorce” (Gottman, 1993). According to this trajectory, couples who eventually divorce are hypothesized to have been unhappily married for some time, and only then to seriously consider dissolution, and only then to actually separate and then divorce. The happiness of the marriage is thought to be predicted by each of their behaviors during their interactions or negotiations. If both spouses exhibit a higher quantity of positive (e.g., agreement, validation) than of negative (e.g., disagreement, dismissal, belittling) behaviors, the couple is deemed to be “regulated.” In “nonregulated” couples, one or both partners emit more negative than positive acts. Nonregulated couples are hypothesized to experience early divorce (first 7 years). In contrast, it is the lack of positive affect (enjoyment of being together, satisfaction) that should predict later (years 7 through 14) divorce (Gottman & Levenson, 2000).

Findings. There is substantial empirical support for all the above theories. It has also been found, perhaps surprisingly, that marital dissatisfaction is a necessary, but not a sufficient, explanation for divorce. Some couples remain deeply dissatisfied, or disengaged, but do not divorce. For example, Davila and Bradbury (2001) found that spouses who were concerned about abandonment and their worthiness of love stayed married even when distressed. Further, although there is a correlation between the two spouse’s marital happiness, only couples in which husbands are unhappier than wives have increased risks of divorce (Gager & Sanchez, 2003). The attributions spouses (especially distressed wives, Bradbury & Fincham, 1992; Gottman, Coan, Carrere, & Swanson, 1998) make about their own and their partners’ behavior are also informative (Bradbury & Fincham, 1990).

Whereas some researchers have simply asked people why they are unhappy in their marriages or why they sought divorce, other researchers consider the partners’ “accounts” less valid because of retrospective biases, self-serving attributions, face-saving attempts, and cognitive dissonance. Nonetheless, some interesting findings regarding accounts have surfaced. For one thing, the reasons given are generally more benign than might have been thought. “Gradual growing apart, losing a sense of closeness”; “serious differences in lifestyle”; “not feeling loved or appreciated” were ranked 1–3 by wives in Braver and O’Connell’s (1998) study (cf. Booth & White, 1980; Gigy & Kelly, 1992; Kitson & Sussman, 1982). None of these divorces would have been allowed if fault had to be established; in fact, the only classic “fault ground” to make the top ten was “husband’s extramarital affair.” For another, the reasons men and women give are very different, leading to the idea that “his” and “her” divorce are quite distinct. In fact, husbands are often quite unclear about why their divorce happened, but wives rarely are (Amato & Previti, 2003; Kitson, 1992).

Consequences of Divorce for Parents

Legal Consequences of Divorce for Parents

Divorce typically has radically different legal consequences for mothers and fathers. In 68–88 % of cases, mothers get physical custody of the children, fathers do so in only 8–14 % of cases, and joint physical custody is specified in 2–6 % (Argys et al., 2007; Braver & O’Connell, 1998; DeLusé, 1999; Emery, 1994; Fox & Kelly, 1995; Logan, Walker, Horvath, & Leukefeld, 2003; Maccoby & Mnookin, 1992; Nord & Zill, 1996; Saluter & Lugaila, 1998; Seltzer, 1990). The amounts of time the children spend with non-primary parents (often called visitation, access, contact, or parenting time) has increased recently. Prior to the 1980s, most decrees specified that children should spend every other weekend with those parents, amounting to about 14 % of the children’s time (Kelly, 2007). Two studies of randomly selected case files in Arizona conducted 10 years apart (Braver & O’Connell, 1998; DeLusé, 1999) found non-trivial increases in the parenting time ordered. Venohr and Griffith (2003) found that, by 2001–2002, almost half of the decrees specified 24–32 % of the days per year and another fifth specified 33–50 % of the parenting time for the non-primary parents. Fewer than one in five specified as little as the traditional 14 %. By 2008, 45 % of the decrees specified 15–35 %, 7 % specified 35–49 % of the child’s time with the father, and 22 % essentially divided parenting time equally (Venohr & Kaunelis, 2008). In Washington State, 46 % of fathers obtained at least 35 % parenting time in 2007–2008 (George, 2008) and in Wisconsin, 24 % had equal parenting time decreed in 2003 (Brown & Cancian, 2007).

Legal custody (who has the legal authority to make decisions regarding medical or educational issues) is more variable from state to state; rates of joint legal custody range from 21 % (Seltzer, 1990, national data) to 76 % (Maccoby & Mnookin, 1992, California data) to 93 % (Douglas, 2003, New Hampshire) and have also changed much more than levels of physical custody over time. Interestingly, both sets of changes appear spontaneous, i.e., not based on corresponding revisions of formal policy. During one longitudinal study conducted by the first author (Braver, Shapiro, & Goodman, 2005), joint legal custody doubled (from about 1/3 to 2/3) over the 3 year course of the study, although there were no discernable changes in any formal or official standards. Rather the informal “culture” among the relevant professionals (judges, attorneys, custody evaluators, mediators, etc.) changed, possibly in response to evidence (e.g., Gunnoe & Braver, 2001; Maccoby & Mnookin, 1992) that it generally had beneficial effects.

This change in the informal culture of professionals paralleled changes in public opinion. In 2006, 86 % of the voters responded affirmatively to the following advisory (i.e., non-binding) ballot question in Massachusetts: “There should be a presumption in child custody cases in favor of joint physical and legal custody, so that the court will order that the children have equal access to both parents as much as possible, except where there is clear and convincing evidence that one parent is unfit, or that joint custody is not possible due to the fault of one of the parents.” Braver, Fabricius, and Ellman (2008) gave the identically worded statement to a representative sample of adult citizens in Tucson, AZ, inviting respondents to indicate how much they agreed with the statement on a 7-point Likert scale. Ninety percent responded on the “agree” side; 57 % responded 7 (=“strongly agree”) and another 30 % responded 6 (=agree). There were no significant differences by gender, age, education, income, whether the respondents themselves were currently married, had ever divorced, had children, had paid or received child support, or by their political ideology.

To explore the depth of commitment to the preference, Votruba (2008) asked a different representative sample of Tucson, AZ, citizens about a hypothetical custody case. Participants were told that the mother and father divided pre-divorce child care “about like average families in which both parents work full-time (both M-F, 9-to-5).” The parents were further described as reasonably good parents who deeply loved their children, with a family life that was quite average, and children who were normally adjusted. Respondents were asked how they would award parenting time if they were judges. About 75 % chose the option, “Live equal amounts of time with each parent.” Almost all the remainder chose “Live with mother, see father a lot.” This was in marked contrast with how they thought parenting time would actually be allocated in “today’s courts and legal environment”—fewer than 25 % thought the equal time arrangement would prevail. All of the others thought that mothers would be awarded most of the parenting time.

This difference between what most people think proper and what they expect courts to assign appears to fuel disapproval of courts. When Braver et al. (2008) asked respondents to “describe the ‘slant’ of the … legal system, as a whole, toward divorcing parents,” 81 % reported that it favors mothers, and only 16 % saw the system as unbiased. This result corresponds well with what divorcing fathers themselves answered in a separate investigation (Braver & O’Connell, 1998): 3/4 thought that it favored mothers and not a single father thought that the system favored fathers. Mothers tended to agree that the system was slanted in their favor: while 2/3 thought it was balanced, three times as many mothers thought it favored mothers as thought it favored fathers. Moreover, only 36 % of experienced divorce attorneys believe that the system is not slanted, 59 % believe that it favors mothers, and only 5 % believe that it favors fathers (Braver, Cookston, & Cohen, 2002).

Psychological and Emotional Consequences of Divorce for Parents

Because divorce has been rated the number one life stressor (Dohrenwend & Dohrenwend, 1974; Holmes & Rahe, 1967), it is not surprising that divorced parents are more likely to suffer psychological and emotional problems than married parents, although most parents are not permanently damaged by divorce. Divorced parents have higher risks of depression, anxiety, and unhappiness, physical illnesses, suicide, motor vehicle accidents, alcoholism, homicide, and overall mortality (e.g., Aseltine & Kessler, 1993; Bloom, Asher, & White, 1978; Davies, Avison, & McAlpine, 1997; Gove & Shin, 1989; Gove, Style, & Hughes, 1990; Hemstrom, 1996; Joung et al., 1997; Kitson, 1992; Kposowa, Breault, & Singh, 1995; Lorenz et al., 1997; Simon & Marcussen, 1999). Involvement in new relationships (e.g., Wang & Amato, 2000) and remarriage (e.g., Demo & Acock, 1996) reduce the risk of such consequences as do religious or personal beliefs and values accepting of divorce (Booth & Amato, 1991; Simon & Marcussen, 1999).

Theories. Several theorists have attempted to explain how divorce affects psychological well-being. Amato (2000) proposed a Divorce-Stress-Adjustment Model, in which the path between divorce and adjustment was mediated by stressors such as sole parenting responsibility, loss of emotional support, continuing conflict with ex-spouses, economic decline, and other stressful divorce-related events. The path to adjustment is also moderated by protective factors such as individual, interpersonal, and structural resources, the definition and meaning of divorce to the individuals, and demographic characteristics, such as gender, age, race, ethnicity, and culture. An alternative model posits the reverse—that divorce is driven by preexisting, stable personality characteristics. Individuals who are poorly adjusted (i.e., those with more severe negative emotional, behavioral, and health outcomes, and difficulty functioning in new roles), often those who divorce and never remarry, may select into divorce and out of remarriage (e.g., they may be more restless or mentally unstable prior to their first marriage) (Davies et al., 1997; Kelly & Conley, 1987; Kitson, 1992; Kurdeck, 1990; Mastekaasa, 1994).

Which gender does better emotionally after divorce? Despite media portrayals to the contrary (e.g., First Wives Club), women tend to show greater emotional adjustment and recovery than do men following divorce (Ahrons & Rodgers, 1987; Braver & O’Connell, 1998; Chiraboga & Cutler, 1977; Wallerstein & Kelly, 1980) for several reasons. First, women tend to be better than men at seeking, building, and using social support networks that buffer the stresses that accompany divorce (Chiraboga, Coho, Stein, & Roberts, 1979; Hughes, 1988; Keith, 1986; Kitson, 1992; McKenry & Price, 1995; Umberson, Chen, House, Hopkins, & Slaten, 1996), and children themselves can be sources of support for custodial parents (Blankenhorn, 1995; McKenry & Price, 1995). Also, mothers must “hold it together” for the sake of the children, whereas noncustodial fathers often do not have this sobering responsibility.

Second, as mentioned earlier, women usually initiate divorce (Ahrons & Rodgers, 1987; Braver, Whitley, & Ng, 1993; Pettit & Bloom, 1984). Spouses who initiate divorce tend to experience more stress before the actual decision to divorce, but relief afterwards. In contrast, spouses who do not initiate divorce experience the most stress once the divorce process commences. Third, men are more likely than women to use ineffective or harmful methods of coping with the stress of divorce, more often turning to substances and alcohol to help them cope (Baum, 2003; Umberson & Williams, 1993).

Fourth, role change may be one of the most important factors contributing to the distress and unsuccessful adjustment of fathers (Umberson & Williams, 1993) and the successful adjustment of mothers (Wallerstein & Kelly, 1980). Women are more likely to view divorce as a “second chance”—mothers report improved work opportunities, social lives, happiness, and self-confidence (Demo & Acock, 1996). Along with divorce, women often gain higher status within-family roles (e.g., head of household, breadwinner, etc.) while men often acquire low-status responsibilities (e.g., gain domestic roles) and are confused or frustrated by their new roles as noncustodial parents (Braver & O’Connell, 1998; Umberson & Williams, 1993).

Finally, divorce settlement satisfaction also affects custodial and noncustodial parents’ emotional well-being differentially (Sheets & Braver, 1996). Fathers frequently feel as though they have experienced gender discrimination at the hands of the legal system, often for the first time in their lives (Braver & O’Connell, 1998). In contrast, women tend to report higher levels of satisfaction with most divorce settlements, including custody, finances, visitation, and property (Sheets & Braver, 1996).

Economic Consequences of Divorce for Parents

Most parents experience substantial financial setbacks when they divorce. The costs of the divorce itself may be very expensive, depending on the state in which it occurs, the complexity of the case, the degree of contentiousness and disagreement, and the use of litigating attorneys as opposed to alternate modes of dispute resolution. When there are few disagreements and the parties do not hire lawyers (in many states, pro se—without lawyers—divorces are the norm) the costs can be as low as a few hundred dollars. On the other hand, some divorces cost well over $100,000 in legal and associated costs. By one popular estimate, the average divorce costs about $20,000 per couple (McDonald, 2009).

Second, and more enduringly, there will be added ongoing costs associated with running a second household. Most of the literature suggests that this hardship falls disproportionately on mothers (Bartfeld, 2000; Bianchi, 1992; Bianchi, Subaiya, & Kahn, 1999; Burkhauser, Duncan, Hauser, & Bernsten, 1990, 1991; Corcoran, 1979; David & Flory, 1989; Duncan & Hoffman, 1985; Espenshade, 1979; Garfinkel, McLanahan, & Hanson, 1998; Hoffman & Duncan, 1985; Holden & Smock, 1991; Peterson, 1996; Sayer, 2006; Smock, Manning, & Gupta, 1999; Sorenson, 1992; Teachman & Paasch, 1994; Weiss, 1984). Weitzman’s (1985) findings that women (and children) suffered a 73 % decline in their standards of living after divorce, while fathers enjoyed a 42 % increase, were the most widely cited, but were later recanted because they were erroneous (Peterson, 1996; Weitzman, 1996).

In contrast, Braver et al. (2005; see also Braver, 1999; Braver & O’Connell, 1998) have contended that, if proper accounting is made, the postdivorce circumstances of fathers and mothers are largely equal in the short term, while in the long run, the majority of divorced mothers fare better than their ex-husbands. To understand the debate requires understanding the operational definition of “standard of living.” Most researchers focus on the income-to-needs ratio, in which the household’s annual income is divided by the Federal Poverty Threshold (FPT) for comparable households. Because child support is very frequently paid by one divorced parent and received by the other (and less frequently, alimony is also paid), both are virtually always subtracted from the payer’s annual income and added to the recipient’s before division by the FPT. But Braver and his colleagues (2005) argued that at least two crucial, yet obvious factors have been typically omitted when such calculations are made.

First, all such calculations are based on gross income, yet, only after tax income can be used to support families. It turns out that custodial parents are taxed far more advantageously than noncustodial parents. Through such tax devices as the Head of Household filing status, the Earned Income Credit, and the Child Tax Credit, the IRS in effect subsidizes the standard of living in the custodial but not noncustodial households. Second, most of the above researchers have assumed that, other than child support, 100 % of the children’s expenses are borne by custodial parents, while noncustodial parents were assumed to pay nothing: no child meals, no child transportation costs, $0 to entertain the children, nothing to provide room for the children in their homes, and no share of medical insurance or medical expenses, etc. In other words, most analyses do not take into account any kind of visitation expenses, nor any direct payments by noncustodial parents for the children, although these are often appreciable (Fabricius & Braver, 2003). Braver and Stockburger (2004) and Rogers and Bieniewicz (2004) specify a set of reasonable and robust assumptions, concerning the cost of children relative to adults, and the proportion of child’s expenses that travel with the child, and economies of scale, that can be used to correct estimates for those expenses borne by noncustodial parents instead of the custodial parents. Using such assumptions, Braver and O’Connell (1998) and Braver (1999) found that the average standards of living shortly after divorce for mothers and fathers were equivalent.

And what of the longer term? Few researchers have studied anything beyond about 18 months after the divorce, but two very common events become significant as time progresses. First, the salaries of many custodial mothers increase: Duncan and Hoffman (1985) found that, by 5 years after divorce, women who remained single increased their standard of living by 34 %. Men’s salaries do not increase similarly because most already earn close to their maximum capacity at the time of divorce. Second, most divorced parents remarry as time progresses. According to Bumpass, Sweet, and Castro-Martin (1990), about two-thirds of divorced mothers and about three-quarters of divorced fathers remarry. When they do, the economics change again. When mothers remarry, they gain more income than expenses, whereas fathers do the reverse (Fabricius, Braver, & Deneau, 2003). Thus, remarriage tends to make mothers’ standards of living higher than fathers’. If the parents’ standards of living were about equal shortly after the divorce, these two factors combine to make the long-term financial effects of divorce, on average, more favorable to mothers than to fathers.

Consequences of Divorce for Parenting

The first year or two following divorce are typically chaotic and highly stressful for both parents and this may adversely affect parent–child relationships (Hetherington, 2003). Divorce often leads to decline in the use of positive parenting strategies (responsiveness) and increases in reliance on negative parenting strategies (e.g. harshness; Harold & Conger, 1997) by both custodial mothers and noncustodial fathers (Kline-Pruett, Williams, Insabella, & Little, 2003; Sturge-Apple, Gondoli, Bonds, & Salem, 2003), although the long-term effects on their behavior may be different.

Custodial parents. Many divorced custodial mothers engage in coercive exchanges with their sons that are characterized by punitive discipline, irritability, an escalation of conflict, and aggressiveness (Hetherington, 1993) while struggling to monitor and supervise the activities of their youngsters (Hetherington & Stanley-Hagen, 2002). Interestingly, fathers who gain custody may have more difficulty than custodial mothers supervising and monitoring their adolescents’ behavior (Maccoby, Buchanan, Mnookin, & Dornbusch, 1993). Children of divorce often have more autonomy and decision-making power than children in non-divorced families (Hetherington & Clingempeel, 1992). Custodial parents’ use of praise, warmth, and other positive parenting strategies are often disrupted by divorce, leading a substantial number of children to emotionally disengage from their families (Hetherington, 1993).

Noncustodial parents. Parents who enter noncustodial status face very different parenting challenges (Maccoby & Mnookin, 1992) because the amount of contact and involvement with their children will substantially decline for many. The contact the average noncustodial fathers tend to have with their children is increasing. Older research (e.g., Amato, 1986; Fulton, 1979; Furstenberg & Nord, 1985; Hetherington, Cox, & Cox, 1982; Hetherington & Hagan, 1986) had shown very low levels of contact, with many fathers disengaging completely. However, later research (Braver, Wolchik, Sandler, & Sheets, 1993b; Braver, Wolchik, Sandler, Fogas, & Zvetina, 1991; Bray & Berger, 1990; Maccoby, Depner, & Mnookin, 1988; Seltzer, 1991) reported higher levels of contact and evidence provided by Cooksey and Craig (1998) indicated that this pattern change was a cohort difference (i.e., current generations of divorced fathers visit more). Several researchers (Fabricius & Hall, 2000; Parkinson, Cashmore, & Single, 2005; Smith & Gallop, 2001) also have reported that both children and fathers wish for more contact, but this is often precluded by the divorce decree, which accords more closely to mothers’ preferences for relatively little visitation for fathers. Interestingly, some children are able to maintain close relationships with nonresident fathers even who they have very little contact (Maccoby et al., 1993).

Many noncustodial fathers have difficulty adapting to their new roles because there are no scripts defining noncustodial relationships (Wallerstein & Corbin, 1986; Wallerstein & Kelly, 1980). Braver and O’Connell (1998) argued that a number of fathers feel “parentally disenfranchised,” believing they have only a limited amount of control over child-rearing issues and have roles that are not valued by their children’s mothers or by the legal system. Noncustodial parents also must cope with changes in the quality of their relationships with their children (Amato & Gilbreth, 1999). Many become very permissive and assume the role of companion rather than disciplinarian or teacher (Hetherington, 1993).

Several theorists have attempted to synthesize information about the factors that predict the amount of contact into theories of father–child relationships.

Interactionist-feminist theory. Arendell (1992a, 1992b, 1994, 1995) adapted an interactionist-feminist perspective when interpreting results from her qualitative study of 75 recently divorced nonresident fathers. In this perspective, father absence is a masculinist strategy to control situations through conflict and tension. While most fathers complained of “injustice, discrimination, resistance, and frustration and discontent” and expressed rage at the legal system and at their ex-wives, this was due to their use of a “masculinist discourse of divorce,” which employed the “rhetoric of rights.”

Family systems theory. The family systems perspective (Arditti & Kelly, 1994) explains father–child relationships by noting that they occur in the context of the mother–father relationship. They note that even when the formal husband/wife relationship ceases to exist, the continuing relationship between parents, vis-a-vis their children, usually necessitates some level of mutual involvement, requiring divorced parents to engage in frequent interactions. The theory posits that fathers who had closer and less conflictual relationships with their ex-wives had better relationships with their children. Those who feel that custody and visitation arrangements are unjust and unsatisfactory are expected to have had poorer relationships with both their children and their ex-wives (Madden-Derdich & Leonard, 2000, 2002).

Role-identity theory. Role-identity theory (Ihinger-Tallman, Pasley, & Buehler, 1993; Minton & Pasley, 1996; Stone & McKenry, 1998) is a “mid-range” theory in which the father’s parenting role identity is expected to predict paternal involvement and child well-being. Father’s role identity included the dimensions of satisfaction with the father role, perceived competence in that role, his investment in the role and its salience. This role identity can be moderated by such factors as the co-parental relationship, mother’s views of the father’s parenting, father’s emotional well-being, the encouragement father receives from others to engage in parenting, and the father’s dissatisfaction with the legal system and the custody and visitation arrangements.

Role-enactment theory. Leite and McKenry (2002) reformulated the above theory into role-enactment theory. In this theory, a new predictor is father’s role satisfaction, which is assumed to be reduced by the fact that aspects of the father role remain salient to many nonresidential fathers despite the ambiguity and barriers they encounter. A second added predictor is “institutional role clarity,” the degree to which arrangements that define expected levels of contact between nonresidential fathers and children were clearly specified in the legal documents. The greater the role satisfaction and the clearer the institutional role, the greater the contact.

Resource theory. Foa and Foa’s resource theory (1980) served as the basis for Rettig, Leichtentritt, and Stanton’s (1999) model emphasizing that, through the normal give-and-take of everyday relationships, people exchange resources. How these resource exchanges flow between partners determines interpersonal behaviors and satisfaction. When viewing noncustodial fathers through this lens, the father’s own perceived economic and social psychological well-being, his communication with the mother during co-parenting, and their degree of conflict were predicted to affect paternal involvement with children.

Social exchange theory. Social exchange theory (Braver, Wolchik, Sandler, & Sheets, 1993a) closely resembles resource theory. It postulates that fathers decide how much to invest in father–child relationships by implicitly comparing the rewards associated with those relationships with their costs; the more positive the reward-to-cost trade-off, the more fathers will invest in relationships. For example, fathers who enjoyed visits more had greater benefits, while those who fought less with wives had reduced costs; either would increase the trade-off, and therefore the level of contact. The model was strongly supported in longitudinal analyses by Braver, Wolchik, Sandler, Sheets, Fogas, et al. (1993b).

Consequences of Divorce for Inter-parental Relationships

When childless couples divorce, they typically have either no or relatively neutral relationships with their ex-spouses years later. In contrast, because couples with children “divorce each other but not their children,” they continue to have relatively involved relationships throughout most of their lives. Thus, most divorcing couples with children suffer high degrees of conflict with ex-partners which persist for 3 years or more after their divorces are finalized (Ahrons & Wallisch, 1986; Masheter, 1991). As time continues, about a quarter achieve working “co-parental” relationships (Adamsons & Pasley, 2006; Ahrons, 1981; Whiteside, 1998), half have almost no contact with the other and engage in “parallel parenting” (Ahrons, 1994; Maccoby & Mnookin, 1992) while the final quarter continue to display great hostility (Ahrons, 1994).

As suggested above, the parents’ conflict with one another is likely to affect the relationships that parents have with their children. For example, many divorcing parents who are conflicted put children in the middle of the differences by denigrating one another or by sending messages to the other parents through the children (Arbuthnot & Gordon, 1997; Buchanan & Heiges, 2001). Such experiences may lead children to feel that they need to take sides in the conflict.

Consequences of Divorce for Children

Developmental Course of Child–Parent Relationships

Parent–child relationships are generally believed to play a crucial role in shaping children’s development and adjustment. Consistent with this belief, considerable efforts have been made to examine the developmental course of these relationships on the welfare and adjustment of children. The research has been extensive, with focus on social skills, psychological health and symptomatology, school behavior and performance, as well as educational attainment, relationship success, delinquency, and life time earnings. Some aspects of these relationships that are influential are those that have more positive effects on children’s development than others as well as the effects of parent–child separations and relationship disruptions on children’s subsequent adjustment. In particular, Kelly and Lamb (2000, 2003; Lamb & Kelly, 2009), among others, have documented how an understanding of normative developmental phases informs both our understanding of how parental separation and divorce may affect children’s development and adjustment as well as how the design of postdivorce living arrangements most likely will benefit children.

As described by Bowlby (1969), and largely confirmed by subsequent research (for detailed review, see Thompson, 2006), infant–parent attachments pass through several developmental phases, during the first of which infants learn to discriminate among adult caregivers and gradually develop emotional attachments to them. Between 7 and 24 months, attachments become increasingly apparent, as infants preferentially seek to be near and interact with specific regular caregivers by whom they are more easily soothed than by strangers. Contrary to Bowlby’s initial speculation and widespread “common sense,” there is considerable evidence that most infants in two-parent families form attachments to both parents at about the same age, around 6–7 months (see Lamb, 2002a, for a review), even though fathers typically spend less time with their infants than mothers (Pleck & Masciadrelli, 2004). This indicates that, although a threshold level of interaction is crucial for attachments to form, time spent interacting is not the only critical dimension.

The amounts of time that infants spend with their two parents do not appear to determine whether or not the attachment relationships with either are insecure or secure. However, the relative prominence of the two parents in caring for and interacting with their children does appear to affect the relative importance of the two relationships with respect to their impact on later development (Lamb & Lewis, 2011). Nonetheless, both relationships remain psychologically important even when there are disparities between the two parents’ levels of participation in child care.

Parents are not equivalently sensitive, of course, and individual differences in responsiveness affect the quality or security of the individual attachment relationships that form. Specifically, insecure attachments are more likely to develop when parents are less sensitive, and it is quite common for infants to be securely attached to one person and insecurely attached to another (Ahnert, Pinquart, & Lamb, 2006), presumably because one parent has been sensitive while the other has been insensitive. The quality of both mother- and father–child relationships remains the most reliable predictor of individual differences in psychological, social, and cognitive adjustment in infancy, as well as in later childhood (Lamb & Lewis, 2011; Thompson, 2006).

Importantly, infants and toddlers need regular interaction with their “attachment figures” in order to foster, maintain, and strengthen their relationships (Lamb, 2002a; Thompson, 2006). This means that young children need to interact with both parents in a variety of contexts (feeding, playing, diapering, soothing, reading, putting to bed, etc.) to ensure that the relationships are consolidated and strengthened. In the absence of such opportunities for regular interaction across a broad range of contexts, infant–parent relationships may weaken rather than grow stronger. When toddlers are separated for as little as a few days from all of their attachment figures (for example, both parents) simultaneously, intense distress and disturbances may persist for as long as 6 months after reunion (Bowlby, 1973; Heinicke, 1956; Heinicke & Westheimer, 1966; Robertson & Robertson, 1971). Reactions are muted, but not eliminated, when children are cared for by other attachment figures or sensitive substitute caregivers during the separation (Robertson & Robertson). Extended separations from parents with whom children have formed meaningful attachments are thus undesirable because they unduly stress developing attachment relationships (Bowlby, 1973). The loss or attenuation of important attachment relationships may cause depression and anxiety, particularly in the first 2 years of life, when children lack the cognitive and communication skills that would enable them to cope with loss. The absence of regular contact slowly erodes relationships, such that, over time, parents who do not interact regularly with their infants effectively become strangers.

Relationships with parents continue to play a crucial role in shaping children’s social, emotional, personal, and cognitive development into middle childhood and adolescence (Lamb & Lewis, 2005, 2011). Indeed, the quality of the mother- and father–child relationships remain the most reliable correlates of individual differences in psychological, social, and cognitive adjustment in infancy, as well as in later childhood (Lamb & Lewis, 2011; Thompson, 2006). Children are better off with insecure attachments than without attachment relationships because these enduring ties play essential formative roles in later social and emotional functioning.

Disruptions in Parent–Child Relationships

There is also a substantial literature documenting the adverse effects of disrupted parent–child relationships on children’s development and adjustment, with a linear relationship between age of separation and later attachment quality in adolescence. The weakest attachments to parents are reported by those whose parents separated in the first 5 years of their lives (Woodward, Ferguson, & Belsky, 2000). Similarly, in a retrospective study of adolescents whose parents had divorced, Schwartz and Finley (2005) found that the age at time of divorce was associated with ratings of both paternal involvement and nurturance, indicating that the earlier the separation, the greater the impact on the quality of children’s relationships with their fathers.

Particularly over the last 2 decades, many large-scale datasets have been plumbed in efforts to understand the effects of parental separation or divorce on the children involved. The results of this research are remarkably consistent: Researchers have clearly demonstrated that, on average, children benefit from being raised in two-parent families rather than separated, divorced, or never married single-parent households (Amato, 2000; Aquilino, 1996; Carlson, 2006; Clarke-Stewart & Brentano, 2006; Clarke-Stewart, Vandell, McCartney, Owen, & Booth, 2000; Hetherington, 1999; Hetherington & Kelly, 2002; McLanahan, 1999; McLanahan & Sandefur, 1994; McLanahan & Teitler, 1999; Simons & Associates, 1996; Simons, Lin, Gordon, Conger, & Lorenz, 1999), although there is considerable variability within groups, and the differences between groups are relatively small. Indeed, although children growing up in fatherless families are, on average, disadvantaged relative to peers growing up in two-parent families with respect to psychosocial adjustment, behavior and achievement at school, educational attainment, employment trajectories, income generation, antisocial, and criminal behavior, as well as intimate relationships, the majority of children with divorced parents enjoy average or better-than-average social and emotional adjustment as young adults (Booth & Amato, 2001; Clarke-Stewart & Brentano, 2006; Hetherington & Kelly, 2002; Kelly & Emery, 2003). Approximately 20–25 % (some studies suggest 30–35 %) of children in post-separation and divorced families give evidence of adjustment problems, compared to 12 % (some studies suggest as much as 15 %) in married families. Thus, the majority of children from separated families evince no psychopathology or behavioral symptoms, although they are likely to experience psychic pain for at least some time (Emery, 1998; Hetherington & Kelly, 2002; Laumann-Billings & Emery, 2000).

Factors That Predict Child Adjustment to Divorce

Such dramatic individual differences in outcomes force us to identify more precisely both the ways in which divorce/single parenthood may affect children’s lives and the factors that might account for individual differences in children’s adjustment following their parents’ separation. Three inter-related factors (economic stress, conflict between parents, and changes in the children’s relationships with their parents) appear to be important but we focus here especially on both conflict and children’s relationships with their parents.

As mentioned above, conflict between the parents commonly precedes, emerges or increases during the separation and divorce processes, and often continues for some time beyond them. Inter-parental conflict is an important predictor of children’s psychosocial maladjustment just as marital harmony, its conceptual inverse, appears to be a reliable correlate of positive adjustment (Cummings, Merrilees, & George, 2010; Johnston, 1994; Kelly, 2000). The negative impacts of high levels of marital conflict on the quality of parenting of both mothers and fathers have been well documented. In general, parental conflict is associated with more rejecting, less warm, and less nurturing parenting by mothers, and with fathers’ withdrawal from parenting and increased engagement in intrusive interactions with their children (Cummings & Davies, 1994; Grych, 2005). Anger-based marital conflict is associated with filial aggression and externalizing behavior problems, perhaps because such parents and children have similar difficulty regulating negative affect (Katz & Gottman, 1993). These and other data support the observation that some of the “effects of divorce” are better viewed as the effects of pre-separation marital conflict and violence (Kelly, 2000).

Unfortunately, the adversarial legal system tends to promote conflict between already vulnerable parents because of its win-lose orientation and the way it fosters hostile behaviors and demands. Although the adversarial process purports to focus on children’s “best interests,” parents’ psychologically-driven legal strategies more often represent their own needs and perceived entitlements, and the effect is to diminish the possibility of future civility, productive communication, and cooperation (Kelly, 2003).

The quality, quantity, and type of parent–child relationships also powerfully affect the post-separation/divorce adjustment of school-aged children and adolescents. As mentioned earlier, deterioration in the quality of parenting after separation has long been recognized (Belsky, Youngblade, Rovine, & Volling, 1991; Clarke-Stewart & Brentano, 2006; Hetherington, 1999; Sturge-Apple, Davies, & Cummings, 2006; Wallerstein & Kelly, 1980). Many parents are preoccupied, stressed, emotionally labile, angry, and depressed following separation, and their “diminished parenting” includes less positive and affectionate involvement as well as more coercive and harsh forms of discipline. Additional intrapsychic factors affecting the quality of parenting include the parents’ psychological adjustment, violence, and levels of conflict. External factors such as absorption in dating, new partners, cohabitation, remarriage, poverty, and financial instability are also associated with reductions in the quality of parenting (Amato, 2000; Hetherington, 1999; Kelly, 2000; Pruett, Williams, Insabella, & Little, 2003; Simons et al., 1999; Wallerstein & Kelly, 1980).

Because single mothers need to work more extensively outside the home than do married or partnered mothers, parents spend less time with children in single-parent families and the levels of supervision and guidance are lower and less reliable than in two-parent families (Hetherington & Kelly, 2002; McLanahan, 1999). Reductions in the level and quality of parental stimulation and attention may affect achievement, compliance, and social skills while diminished supervision makes antisocial behavior and misbehavior more likely (Hetherington & Kelly, 2002).

Many researchers have identified specific aspects of parenting that can moderate the impact of separation and divorce on children’s social, emotional, and academic adjustment, thereby protecting children against some of the harmful impacts of high conflict. Effective parenting by separated mothers is characterized by warmth, authoritative discipline (setting limits, noncoercive discipline and control, enforcement of rules, appropriate expectations), academic skill encouragement, and monitoring of the children’s activities (Amato & Fowler, 2002; Buchanan, Maccoby, & Dornbusch, 1996; Hetherington, 1999; Martinez & Forgatch, 2002; Simons et al., 1999). As described in more detail below, more positive adjustment following divorce is also associated with such effective paternal behaviors as active involvement (help with homework and projects, emotional support and warmth, mutual discussion of the children’s problems, and involvement in school (Amato & Fowler, 2002; Amato & Gilbreth, 1999; Hetherington, 1999)).

Divorce commonly disrupts one of the child’s most important and enduring relationships, that with his or her father. As Amato (e.g., Amato & Gilbreth, 1999) has shown with particular clarity, however, the bivariate associations between father absence and children’s adjustment are much weaker than one might expect. Indeed, Amato and Gilbreth’s (1999) meta-analysis revealed no significant association between the frequency of father–child contact and child outcomes, largely because of the great diversity in the types of “father-present” relationships. We might predict that contacts with abusive, incompetent, or disinterested fathers are likely to have much different effects than relationships with devoted, committed, and sensitive fathers. As expected, Amato and Gilbreth (1999) found that children’s well-being was significantly enhanced when their relationships with nonresident fathers were positive, when the nonresident fathers engaged in “active parenting,” and when the contact was frequent. Dunn, Cheng, O’Connor, and Bridges (2004), Simons and Associates (1996), Hetherington, Bridges, and Insabella (1998), and Clarke-Stewart and Hayward (1996) likewise reported that children benefited when their nonresident fathers were actively involved in routine everyday activities, and this conclusion was clearly supported in recent analyses by Carlson (2006) of data from the National Longitudinal Study of Youth. Carlson showed that father involvement was associated with better adolescent adjustment and that paternal involvement partially mediated the effects of family structure (notably divorce or single parenthood) on adolescents’ behavioral outcomes. Similarly, higher levels of paternal involvement in their children’s schools was associated with better grades, better adjustment, fewer suspensions, and lower dropout rates than were lower levels of involvement (Nord, Brimhall, & West, 1997). Overall, active engagement in a variety of specific activities and ongoing school-related discussions between fathers and their adolescents significantly lowered the probability of school failure.

Another meta-analysis indicated that, on multiple ratings of emotional and behavioral adjustment and academic achievement by mothers, fathers, teachers, and clinicians, children in joint physical custody were better adjusted than children in sole custody arrangements (Bauserman, 2002). In fact, children in shared custody were as well adjusted as children whose parents remained married. Although joint physical custody parents reported less past and current conflict than did sole physical custody parents, conflict did not explain the superiority of the children in joint custody arrangements. Again, the clear implication is that active paternal involvement, not simply the number or length of meetings between fathers and children, predicts child adjustment. This suggests that postdivorce arrangements should specifically seek to maximize positive and meaningful paternal involvement rather than simply allow minimal levels of visitation. As in non-divorced families, in other words, the quality of continued relationships with the parents—both parents—is crucial (Kelly & Lamb, 2000, 2003). Stated differently and succinctly, the better (richer, deeper, and more secure) the parent–child relationships, the better the children’s adjustment, whether or not the parents live together (Lamb, 2002a, 2002b).

A recent longitudinal study of representative samples of adolescents living in low income neighborhoods in Boston, San Antonio, and Chicago nicely illustrated the associations over time between nonresident paternal involvement and adolescent delinquency, after statistically controlling for the effects of influences such as demographic factors and the quality of mother–child relationships (Coley & Medeiros, 2007). As expected, nonresident paternal involvement was associated with less delinquency overall; importantly, higher paternal involvement was associated with declines in delinquency over time, particularly among adolescents who were more involved with delinquent activities. In addition, as delinquency increased, paternal involvement increased too, suggesting that fathers were responding to changes in their children’s problem behavior. Similarly, in another longitudinal study of adolescents, Menning (2006) showed that adolescents whose nonresident fathers were more involved were less likely to start smoking.

Given the demonstrated importance of fathers’ active participation and effective parenting, the influence of maternal attitudes on the extent of paternal involvement in the marriage and following separation and divorce is important (Cowdery & Knudson-Martin, 2005; Pleck, 1997). Mothers can be influential “gatekeepers” of paternal involvement through attitudes and behaviors that limit or facilitate fathers’ opportunities to develop close relationships with their children. Mothers’ traditional attitudes toward women’s roles, identities linked primarily to caregiving, and perceptions that mothers are more competent at child care than fathers are associated with more active inhibitory gate-keeping, particularly following separation. These attitudes and perceptions by mothers predict lower levels of father involvement (Allen & Hawkins, 1999; Fagan & Barnett, 2003).

Implications for Divorce Policy

Although a number of factors help account for individual differences in the effects of divorce on children, therefore, the ability to maintain meaningful relationships with both parents does appear to be of central importance. Writing on behalf of 18 experts on the effects of divorce, Lamb, Sternberg, and Thompson (1997, p. 400) observed more than a decade ago that: “To maintain high-quality relationships with their children, parents need to have sufficiently extensive and regular interactions with them, but the amount of time involved is usually less important than the quality of the interaction that it fosters. Time distribution arrangements that ensure the involvement of both parents in important aspects of their children’s everyday lives and routines…are likely to keep nonresidential parents playing psychologically important and central roles in the lives of their children.”

In order for parents to have a positive impact on their children’s development, therefore, it is important that parents be integral parts of their children’s lives. This remains especially important as children get older and greater portions of their time are occupied outside the family by virtue of friendships, extracurricular activities, sports, and the like. At all ages, it is important for parents to know teachers and friends, what’s happening at school or preschool, how relationships with peers are going, what other activities are important or meaningful to the children, etc., and to be aware of daily ups-and-downs in their children’s lives. It is hard to do this without regular and extensive first hand involvement in a variety of contexts.

As Kelly and Lamb (2000; Lamb, 2002b; Lamb & Kelly, 2001, 2009) reiterated, the ideal situation is one in which children with separated parents have opportunities to interact with both parents frequently in a variety of functional contexts (feeding, play, discipline, basic care, limit-setting, putting to bed, etc.). The evening and overnight periods (like extended days with naptimes) with nonresidential parents are especially important psychologically for infants, toddlers, and young children. They provide opportunities for crucial social interactions and nurturing activities (including bathing, soothing hurts and anxieties, bedtime rituals, comforting in the middle of the night, and the reassurance and security of snuggling in the morning) that 1–3 h long visits cannot provide. According to attachment theory (Lamb, 2002a), these everyday activities promote and maintain trust and confidence in the parents, while deepening and strengthening child–parent attachments, and thus need to be encouraged when decisions about access and contact are made.

One implication of these findings is that even young children should spend overnight periods with both parents when both have been involved in their care prior to separation, even though neo-analysts have long counseled against this (Kelly & Lamb, 2000; Lamb & Kelly, 2001). As Warshak (2000) has pointed out, the prohibition of overnight “visitation” has been justified by prejudices and beliefs rather than by any empirical evidence. When both parents have established significant attachments and both have been actively involved in the child’s care, overnight “visits” help consolidate attachments and child adjustment, not undermine them. Consistent with this reasoning, the results of research by Pruett and her colleagues showed that regular overnight visits were associated with better adjustment on the part of toddlers and young children (Pruett et al., 2003). Parents who have been actively involved before divorce but are then denied overnight access to their children are thereby excluded from an important array of activities, and the strength or depth of their relationships suffer as a result.

To minimize the deleterious impact of extended separations from either parent, attachment theory tells us there should be more frequent transitions than would perhaps be desirable with older children (Kelly & Lamb, 2000). To be responsive to young children’s psychological needs, in other words, the parenting schedules adopted for children under age 2 or 3 should involve more transitions, rather than fewer, to ensure the continuity of both relationships and to promote the children’s security and comfort. Although no empirical research exists testing specific parenting plans following separation, it is likely, for example, that infants and toddlers would remain most comfortable and secure with schedules allowing the children to see their nonresident fathers at least three times a week, including at least one overnight extended stay (assuming that the fathers are competent, experienced parents), so that there is no separation of greater than 2–3 days. From the third year of life, the ability to tolerate longer separations begins to increase, so that most toddlers can manage two consecutive overnights with each parent without stress.

Recently, Fabricius and Luecken (2007) found that the quality of relationships between university students and their divorced fathers were better the more time they had lived with them, no matter how much parental conflict there has been, although the amount of exposure to parental conflict generally had negative effects. Inter-parental conflict should thus be avoided wherever possible, but its presence should not be used to justify restrictions on children’s access to either of their parents.

Legal Processes Associated with Parental Divorce

When parents divorce, they must make formal—and legally-binding—arrangements about various matters, including child support, payment of the children’s medical expenses and insurance, and, most importantly, custody and parenting time. Only 2–10 % of divorcing couples in the USA have their custody provisions decided by a judge (Braver & O’Connell, 1998; Logan et al., 2003; Maccoby & Mnookin, 1992); the remainder reach agreements themselves that court officials then “rubber stamp.” Many steps are often involved in reaching such agreements.

Parents typically have one or more court appearances or “settlement conferences” before final resolution. Often judges exhort parents to settle their differences and, when they are making too little progress, may order them to special classes, and appoint mediators or custody evaluators. Often, but not always, lawyers are involved. In about 30 % of divorces, neither parent is represented; in another 30 % only one (most commonly the mother) is; and in the remainder both are (Braver & O’Connell, 1998; Logan et al., 2003; Maccoby & Mnookin, 1992).

Many parents attend “Parent Education” classes (Blaisure & Geasler, 1996) which have recently become very popular (Arbuthnot, 2002; Blaisure & Geasler, 2000), although there is little evidence that they have the intended effects: facilitating negotiations, preventing couples from litigating, and improving the quality of their agreements (Braver, Salem, Pearson, & DeLusé, 1996; Douglas, 2006; Goodman, Bonds, Sandler, & Braver, 2004; Pollet & Lombreglia, 2008; Sigal, Sandler, Wolchik, & Braver, 2011).

About one quarter of parents enter into mediation, which involves neutral professionals helping couples resolve disputes about custody, parenting time, and sometimes financial issues (Braver & O’Connell, 1998; Emery, 1994; Kelly, 2004). In some states, mediation is not confidential, and if the parents do not forge agreements, the mediators make influential recommendation to the courts (Douglas, 2006). Australia is pioneering “child inclusive” divorce mediation in which the children meet separately with specially trained intermediaries (McIntosh, Wells, Smyth, & Long, 2008).

About 5–10 % of the time, experts (usually psychologists) are jointly hired by the parents to perform custody evaluations and make non-binding recommendations either to the parents or directly to the court. However, there have been several recent critiques of the legal and ethical propriety as well as the value of custody evaluations (Bow & Quinnell, 2002, 2004; Martin, 2005; Tippins & Wittman, 2005; Emery, Otto & O’Donohue, 2006).

After one or more of these interventions, the final arrangements may diverge substantially from the parents’ initial intentions. In Braver and O’Connell’s (1998) study, for example, 70 % of the mothers initially wanted sole legal custody, and the remainder wanted joint custody, whereas 75 % of the fathers wanted joint custody, and the remainder were equally split between wanting sole maternal and sole paternal custody. Similarly, Maccoby and Mnookin (1992) reported that 82 % of the mothers wanted sole maternal physical custody, while equivalent thirds of the fathers wanted joint, sole paternal, or sole maternal custody arrangements. Two-thirds of the University students surveyed by Fabricius and Hall (2000) reported that their mothers had wanted to be primary residential parents, and almost two-thirds reported that their fathers had wanted equal or nearly equal living arrangements or to be their primary residential parents. In all three studies, however, the parents ultimately were twice as likely to reach agreements reflecting the mothers’ rather than the fathers’ preferences.

Why do mothers’ preferences tend to prevail? Some have speculated that fathers’ stated preferences are only bargaining positions later negotiated away in exchange for concessions on child support (Neely, 1984; Singer & Reynolds, 1988; Weitzman, 1985), but all three of the empirical investigations that have specifically investigated this claim have failed to find support (Braver & O’Connell, 1998; Maccoby & Mnookin, 1992; Venohr & Griffith, 2003).

Instead, it is likely that fathers’ low level of persistence follows the guidance they receive from judges, attorneys, custody evaluators, parent educators, and mediators (Braver & O’Connell, 1998; Maccoby & Mnookin, 1992; Mnookin, 1984; Mnookin & Kornhauser, 1979). But this is not inevitable: DeLusé (1999) found that fathers who attended Parent Education classes negotiated significantly more parenting time, presumably because instructors had explained the courts’ desire to keep both parents involved.

Lawyers may play a particularly important role. At an Arizona State Bar convention, Braver et al. (2002) distributed a custody scenario to Family Law attorneys who were randomly assigned to “represent” either the mother or the father. Although the facts in the scenario were neutral, the attorneys were more likely to tell mothers that they would probably prevail in seeking physical custody and the majority of parenting time. Note that even if courts were indeed completely gender-neutral, lawyers’ views can discourage worthy fathers from pursuing their desired arrangements, creating a self-fulfilling prophecy.

Current Debates About Divorce

Because its effects on both parents and children are so dramatic, and because divorce is so common, many aspects of current policy and practice are the focus of intense debate.

No-fault/unilateral divorce. Beginning in the late 1960s, as explained above, no-fault divorce became the standard throughout the United States. Very importantly, divorces also became unilateral, because reluctant parties could do nothing to prevent, delay, or avoid divorce. As noted earlier, the surge in the divorce rates coincided with the passage of these laws, leading some advocates to press for a return to fault-based divorces (Brinig & Buckley, 1998). Making divorce harder to obtain would perforce reduce the rate of divorce, but efforts to repeal no-fault divorce laws are unlikely to prevail in the current cultural climate (DiFonzo, 1997). There is, however, public support for better public education about marriage, to prevent couples from later seeking divorce. At least one Catholic diocese increased the length of marriage preparation classes for churchgoers from 6 to 9 months in 2010, and many courts now require parties considering separation to attend cautionary classes.

Custody standards. There has been a considerable debate about the legal standards that guide decisions about custody. Since 1970, the prevailing standard has been the “Best Interests of the Child” standard (BIC). The BIC is generally considered an improvement over past standards because it accords primacy to children’s needs, and is egalitarian, flexible, and simple (Chambers, 1984; Warshak, 2007) but it has been criticized for being vague and for allowing judges to rely on idiosyncratic biases (Chambers, 1984; Finley & Schwartz, 2007). Because rulings are unpredictable, some argue that it fosters custody disputes (O’Connell, 2007). The major alternative proposals are the Primary Caretaker standard (Chambers, 1984; Maccoby, 1999), specifying that parents who provided the most childcare during marriage should be primary custodians; the Approximation Rule (American Law Institute, 2002; Lamb, 2007; Warshak, 2007), dictating that “the court should allocate custodial responsibility so that the proportion of custodial time the child spends with each parent approximates the proportion of time each parent spent performing caretaking functions for the child prior to the parents’ separation” (American Law Institute, 2002, p. 1); and the Joint (or Shared) Custody standard, specifying that children should reside about equally with both parents. Of these, the latter standard seems to be gaining the most ground (Fabricius, Braver, Diaz, & Velez, 2010; Parkinson, 2010), if not with feminist groups or the Bar. Critics also argue that this standard (like all the others) is insensitive to the possible effects of domestic violence, which (they argue) are more likely to continue when the parents are required to have frequent contacts, if only to transfer the children.

Processes for obtaining divorce. In addition to these substantive issues, considerable controversy surrounds the processes involved in reaching decisions. In particular, divorce settlements are guided by the courts, which promote an adversarial model of dispute resolution. Critics argue that this is inappropriate for divorce and custody disputes because the resulting conflict can harm children (Weinstein, 1997). They urge that alternative systems of dispute resolution should instead be the norm (Firestone & Weinstein, 2004).

Should children have an independent voice when disputes involve their welfare (Parkinson & Cashmore, 2009)? Should children be expected or allowed to testify (“choose between his parents”), and if so, to whom and when (in open court, to the judge in chambers, to a trained and sensitive mental health professional)? Should they have their own legal counsel or guardians ad litem to advocate for them, and/or ensure not only that children’s interests, broadly defined, but also the specific needs of individual children and families, are emphasized (Fabricius, 2003; Warshak, 2003)? These knotty questions are the focus of considerable controversy but little research.

Two issues involving professional practice standards are also controversial. As noted earlier, there is debate about whether custody evaluators base their recommendations on sound scientific grounds (Bow & Quinnell, 2002, 2004; Martin, 2005; Tippins & Wittman, 2005) and a new role of Parenting Coordinator has developed (Sullivan, 2005). Parenting Coordinators are professionals (typically psychologists and counselors) who are appointed by courts to handle ongoing visitation disputes and other parenting issues for postdivorce families. Whether judges can or should cede some of their legal authority to helping professionals who decide disputes promptly on behalf of the courts while dispensing family therapy is hotly disputed.

Methodological Issues in Studying Marital Dissolution

A number of methodological issues surface in the voluminous research on divorce, its causes and its aftermath. We focus here on two categories of issues: (1) the sources of information; and (2) the research designs.

Data sources and measures. When exploring divorce, researchers have access to some more or less objective data, including employment and labor force data, tallies of divorces, child support payments logged by administrative agencies (Braver, Fitzpatrick, & Bay, 1991) and various official court filings (Braver & Bay, 1992; Braver, Whitley, & Ng, 1993). However, most of the information sought by researchers can only be obtained by querying the family members involved, yet it is well known that such self-report data may be erroneous and biased. Unfortunately, researchers too often fail to determine whether there are valid measures of the constructs at issue, to ensure that the measures are standardized across studies (for the sake of comparability), or to document that the measures have adequate psychometric properties.

An example of a measure that has shown convincing validity is the Child Behavior Checklist (CBCL, Achenbach, 1978, 1991a; Achenbach & Edelbrock, 1979; Achenbach & Rescorla, 2007). Painstakingly developed over many decades, this measure is often thought to be the “gold standard” of children’s behavioral and mental health problems. One-week test-retest reliabilities average 0.89 (Achenbach, 1991b) and its concurrent validity using clinical referred vs. non-referred groups is excellent (Achenbach & Edelbrock, 1983). CBCL scores also correlate reasonably well with diagnoses obtained via structured interview by trained clinicians (cf. Bird et al., 1987). The original 118-item measure has been shortened to a widely used 32-item version called the Behavior Problems Index (BPI; Peterson & Zill, 1986) while Moore, Halle, Vandivere, and Mariner (2002) have successfully created an even shorter 6 item version.

At the other extreme, the amount of contact children have with their nonresidential fathers is often inconsistently—and rather poorly—measured (Smyth, 2002). Comparing six datasets, Argys et al. (2007) concluded that, “What is most striking about the reports of father-child contact … and perhaps most alarming to researchers, is the magnitude of the differences in the reported prevalence of father-child contact across the different surveys” (p. 383). In particular, quantitative scales are often more reliable and valid than ordinal items. For example, Fabricius and Luecken (2007) asked young adults four questions about the typical number of days and nights they spent with their fathers and converted those to the percentage of the child’s parenting time spent with the father. In addition, of course, researchers have argued that the construct of father involvement needs to be better conceptualized before it can be well measured (Fabricius et al., 2010; Pasley & Braver, 2003).

Researchers also need to consider who provided the information. Among others, Marsiglio, Day, and Lamb (2000) have called attention to the discrepancies among respondents’ reports. Generally, there are two sorts of informant bias in this area of research: a “self-serving” bias, with informants describing themselves in more socially desirable terms than warranted (Cialdini, Braver, & Lewis, 1974; Miller & Ross, 1975; Sicoly & Ross, 1977); and an “ex-spouse-bashing” bias (Braver et al., 1991; Sonenstein & Calhoun, 1990) characterized by exactly the opposite tendencies when informants are describing their former spouses. It is thus desirable to “triangulate” reports by obtaining matched reports from both mothers and fathers whenever feasible. Pasley and Braver (2003) also suggested that researchers obtain retrospective reports from adult children because they appear to “split the difference” between the two parents’ views.

Research design issues. Divorce processes unfold over long periods of time, but researchers typically take “snapshots” at one point in time. Correlations among variables in such cross-sectional datasets unfortunately cannot elucidate the causal processes involved. And very few studies have involved longitudinal designs, with the same family members repeatedly observed over longer periods so that causal processes can be explored. Maccoby and Mnookin (1992) and Braver and O’Connell (1998) are rare exceptions who have obtained prospective longitudinal studies, in which the families were initially observed before the divorce and were then reassessed (Braver & O’Connell, 1998; Maccoby & Mnookin, 1992). Among the causal questions that could be addressed more reliably using such data was whether fathers’ greater contact made fathers better child support payers or vice versa. In fact, Braver and colleagues found that a third variable, the experience of parental disenfranchisement, explained both of these behaviors (Braver, Wolchik, Sandler, Sheets, Fogas, et al., 1993).

Another important research design issue limits the generalizability of certain key findings, including, for example, Bauserman’s (2002) meta-analytic conclusion that children in joint custody families fare better than those in sole custody. Because virtually all families self-select or are selected into the custody arrangements they experience (custody is not assigned at random), they are likely to differ in several important ways from families with other custody arrangements. Such concomitant differences, as opposed to the arrangement per se, may account for differences in children’s outcomes. Accordingly, Emery, Otto, and O’Donohue (2005) questioned whether the same benefits would accrue if joint custody was imposed on less-than-willing families. Fabricius et al. (2010) have advocated that such important questions can better be explored using alternative research designs, such as natural experiments and using sophisticated statistical methods to control for preexisting differences and self-selection or predisposing factors (Gunnoe & Braver, 2001).

Conclusions and Implications for Practice

Divorce began to become much more common in the late 1960s, but rates have stabilized in the last 2 decades. The increased reliance on divorce in the late twentieth century was attributable to a variety of factors, with changes in the role and attitudes of women (within families and in society more generally) being particularly significant, along with changing laws that made divorce easier to attain by obviating the need to prove that either party was at fault. Although divorce proved beneficial to many of those who sought it, marital dissolution has profound influences on all the individuals involved, as several decades of research have documented. Specifically, mothers, fathers, and children are all affected emotionally, psychologically, and economically by divorce, with variations in the ways parents negotiate divorce playing a particularly important role in determining how children are affected psychologically. This realization is sometimes credited with recent declines in the rates of divorce, with at least some parents concluding that divorce should be deferred “for the sake of the children” and a number of programs around the world have been designed to forestall hasty divorces. Such efforts have also been complemented by the design of programs, exemplified by those introduced in Australia in 2008, designed to minimize the adverse effects of divorce by encouraging parents to make decisions with their children’s interests in mind. Ongoing evaluations of programs such as these will doubtless affect social policy in other jurisdictions.

One implication of such programs is the implicit recognition that divorce per se is neither good nor bad in all circumstances—the effects are extremely variable. This means that both researchers and policy makers need to undertake much more nuanced and sensitive research than might have been warranted when divorce was viewed as a similar experience for all individuals.