CENTENNIAL ESSAY Careers in Print: Books, Journals, and Scholarly ~e~utations' Elisabeth S. Clemens, Walter W. Powell, Kris McIlwaine, and Dina Okamoto University of Arizona Academic reputation rests on publication. But unlike many fields, sociology

by Pamela S. Webster, Terri L. Orbuch, James S. House
CENTENNIAL ESSAY Careers in Print: Books, Journals, and Scholarly ~e~utations' Elisabeth S. Clemens, Walter W. Powell, Kris McIlwaine, and Dina Okamoto University of Arizona Academic reputation rests on publication. But unlike many fields, sociology
Pamela S. Webster, Terri L. Orbuch, James S. House
The American Journal of Sociology
Start Page: 
End Page: 
Select license: 
Select License

Effects of Childhood Family Background on Adult Marital Quality and Perceived stability'

Pamela S. Webster

Brown University

Terri L. Orbuch and James S. House

University of Michigan

The authors examine the effect of various single-parent childhood family structures on adult marital quality and perceived stability. The authors find no important differences in marital happiness by family history. However, among adults in less than very happy marriages, both adult children of divorce and adults who never lived with their father report significantly higher chances of divorce than those from two-parent families. Among those in relatively un- happy marriages, children of divorce more often report patterns of interaction that are likely to strain a marital relationship. Children of divorce are significantly more likely than those from two-parent families to have thought their marriage may be in trouble.

A number of recent changes have occurred in the American family. In the 1960s and 1970s, divorce rates in the United States rose dramatically (Cherlin 1981). During that time, the proportion of children witnessing the breakup of their parents' marriage also grew substantially. In addi- tion, the incidence of nonmarital childbearing has greatly increased re- cently. Approximately one-quarter of all children are currently born out of wedlock (Select Committee on Children, Youth, and Families 1989).

' The first author was supported by training grant T32 AG00114-07 from the National Institute of Aging and the Institute of Gerontology at the University of Michigan while conducting this research. An earlier version of this paper was presented at the annual meeting of the Population Association of America, Cincinnati, April 3, 1993. The authors would like to thank the Institute of Gerontology and the Institute for Social Research at the University of Michigan, Arland Thornton and members of the Family Studies Seminar at the Institute for Social Research, and the AJS reviewers for valuable comments on a previous draft. Direct all correspondence to Pamela S. Webster, Population Studies and Training Center, Box 19 16, Brown University, Prov- idence, Rhode Island 02912.

O 1995 by The University of Chicago. All rights reserved. 0002-9602196/10102-0005$01.50

404 AJS Volume 101 Number 2 (September 1995): 404-32 Two decades ago, it was estimated that roughly three out of every four children born in the period from 1930 to 1960 would spend their entire childhood with both of their parents (Uhlenberg 1974; Bumpass and Sweet 1989~). Yet despite a recent leveling off (and perhaps even decline) in divorce rates, recent estimates suggest that by the age of 18, at least one-half of U.S. children will spend part of their childhood living in a single-parent family (Bumpass 1984; Bumpass and Sweet 1989a; Bianchi 1990). Our primary goal is to understand the long-term effects of these childhood experiences on adult marital quality and perceived stability.

Previous research has suggested that parental family structure substan- tially influences adult children's marital stability (i.e., risk of divorce). Adult children of divorce have long been known to exhibit higher rates of divorce than children from intact families (e.g., Bumpass and Sweet 1972; Pope and Mueller 1976; Mueller and Pope 1977; Kobrin and Waite 1984; Glenn and Kramer 1987; Wallerstein and Blakeslee 1989). In addi- tion, there is recent evidence that other nonintact parental family struc- tures may have important effects on adult children's rates of marital dissolution (Bumpass, Martin, and Sweet 1991). But few studies have examined the ways in which parental family structure influences chil- dren's subsequent marital instability (Kulka and Weingarten 1979).

Theoretical Explanations

The process whereby marital instability repeats itself in successive gener-

ations is not well understood. Several theoretical perspectives have been

proposed to explain the adverse effects of parental divorce on the well-

being of offspring (Amato 1993; Kalter et al. 1989; McLanahan 1985), although few have clear implications for the marital relationship of off- spring. First, there is the socialization perspective, which posits that the absence of one parent and concomitant decrease in parental supervision and attention lead to dysfunctional learning in single-parent families and also suggests that the absence of the same-sex parent may be especially detrimental to a child's development. This explanation would lead one to expect similar outcomes for children of divorce and for other children from single-parent families of any kind. One mechanism by which this socialization may operate is inadequate parental control, whereby paren- tal supervision over mate selection and parental support in the early years of marriage is reduced, producing similar results for those who experienced parental divorce and those who for other reasons lived in single-parent families. Moreover, one might extend the socialization per- spective, to hypothesize that those offspring who spent the least time with two parents (i.e., those who never lived with their father) should exhibit the greatest differences.

The second explanation focuses on the interparental conflict that is

thought to often precede divorce, positing that such conflict negatively

influences children's psychological adjustment. Children may react to

this hostility with negative emotions or be drawn into the conflict and

may blame themselves. This perspective would predict that the well-

being of children of divorce (assuming marital conflict precedes divorce)

and children from intact homes with high levels of parental marital con-

flict will be more adversely affected than children who grew up in single-

parent families.

A similar theoretical model developed by Caspi and Elder (1988) would lead to the same hypothesis, although this perspective emphasizes that parental conflict may influence the personality and interactional style of offspring rather than their well-being in general. This perspective posits that personality and behavior persist both over the lifecourse of individu- als and across generations. Viewing the family environment as central in linking the behavior of parents and children, the notion is that children develop problem dispositions from problematic family relationships in their parent's generation, from which children develop an interactional style that is maintained across time and that they carry into their marital relationships as adults years later (interactional continuity). If one as- sumes that divorce more often is associated with problematic family rela- tionships than other types of single-parent families, then we would pre- dict greater interpersonal (marital) difficulties among children of divorce than among those from other types of single-parent families (or intact homes).

The third explanation is that the economic hardship that frequently follows divorce for single mothers and their children is the important causal agent leading to a reduction in offspring well-being. Since eco- nomic hardship should be measured in childhood rather than adulthood, this hypothesis, while an important one to consider, cannot be tested here.

Fourth, there is the notion that the life stress and family stress associ- ated with divorce and its sequelae adversely affect offspring. Here, one would hypothesize that the effects should be greater for offspring who experienced parental divorce than for those who grew up in single-parent families throughout their childhood. However, it is unclear whether this stress should influence the marital relationship of offspring years later.

The last explanation is that the functioning of the custodial parent is key to an offspring's development (Chase-Lansdale and Hetherington 1990; Furstenberg and Cherlin 1991). The notion is that how well a parent functions will directly influence children's well-being and their relationships with others. Adverse consequences may be greater for chil- dren of divorce, because it is thought that often after a period of two

Family History

IA \

Marital Quality Marital Stability


FIG. 1.-Heuristic diagram

years divorced custodial parents adjust to their new role and function effectively in their parental role, whereas single parents who do not expe- rience the stress of a divorce (or the death of a partner) are more likely to already be functioning effectively (although they also experience sources of chronic strain). Again, it is not clear whether and how the functioning of the custodial parent during childhood is expected to influ- ence the marital relationships of offspring years later.

Because our central concern is to examine the marital relationships of adult children of divorce, only two of the above explanations are relevant and testable. First, the absence of a role model (socialization) leads to the hypothesis that both children of divorce and children from single-parent families will have lower marital quality than those with a two-parent model. Alternatively, the interparental conflict perspective predicts that children of divorce will exhibit lower marital quality and greater diffi- culty in marital interactions than children who lived in single-parent families (due to absent fathers or parental death).

Figure 1 illustrates the ways in which family history during childhood may influence marital stability in adulthood. First, the above causal mechanisms could predict that family history influences marital quality, which subsequently influences marital stability. This proposed causal linkage is illustrated in figure 1(arrows A and B). In other words, differ- entials in divorce rates may be produced by lower average marital quality among those with particular family histories. Yet family history could also influence marital stability in another way. Namely, family history may influence how marital quality affects marital stability. In other words, a statistical interaction between marital quality and family history may be present; this possibility is illustrated in figure 1 by arrow C. Wc will ask whether, at a given level of marital quality, children of divorce are more inclined to consider divorce as an alternative to marriage than their peers from intact families (Greenberg and Nay 1982; Glenn and Kramer 1987). Given that children of divorce have seen firsthand that marriages break up, they may be more prone to consider or enter into divorce in response to marital problems. Hence family history may not be important because it increases one's chances of having an unhappy marriage, but rather because it may influence how one responds to a problematic marital situation.

In order to advance our understanding of the models discussed above, this article will evaluate four questions:

  1. Does marital quality differ by family history?
  2. Does perceived marital stability differ by family history?
  3. Does the impact of marital quality on perceived marital stability differ by family history?
  4. What factors explain any greater propensity to divorce among those with particular family histories?


The data for this analysis come from the National Survey of Families and Households (NSFH), conducted in 1987 and 1988 in a nationally representative sample of households. Interviews were collected with a main sample of 9,643 respondents and with an oversample of blacks, Puerto Ricans, Mexican Americans, single-parent families, and families with stepchildren, cohabitating couples, and recently married persons, bringing the total sample size to 13,017. Because preliminary analyses suggested that family history is less strongly associated with marital qual- ity and the propensity to divorce in second, third, and higher order mar- riages, we limit our analyses to first marriages, or 6,333 (weighted N) of the sample respondents,' with a correction for selectivity (described be- low). Here, the data are weighted for individual-level analysis, using the product of a weight for the probability of selection, a screening non- response adjustment, an interview nonresponse adjustment, and a post- stratification adjustment that aligns the sample distributions of sex, age, racelethnicity, and region with the 1987 Current Population Survey

Our analyses indicate that once an individual has experienced his or her own divorce, the impact of parental family history on marital quality and stability is relatively inconsequential. Thus, our analysis of the impact of parental family history on marital relationships is confined to those in first marriages.

(CPS). A more detailed discussion of sample design and content is avail- able in Sweet, Bumpass, and Call (1988).


Family history.-Respondents were asked whether they lived with both biological parents until age 19, and, if they did not, interviewers ascertained when and why each episode occurred. From this information, a five-category family history variable was created. If by the age of 16 a respondent did not live with both biological parents due to parental sepa- ration or divorce, that respondent is coded as having divorced parents (N = 444). If by the age of 16 a respondent did not live with both biological parents because one parent had died, the respondent is coded as having a parent die during their childhood (N = 510). Another cate- gory comprises those who never lived with their biological father, most of whom were born out of wedlock (N = 199). Those who lived with both biological parents until age 16 are coded as having intact families (N = 4,978). Last, those who did not live with both biological parents for reasons other than parental divorce, separation, parental death, or never living with their father were placed in a residual "other" nonintact family history category (N = 201). This five-category variable was then recoded to create four dichotomous family history variables: parental divorce, father absent, parental death, and other (nonintact). The intact (two biological parents) variable is the omitted category. The "other" family history group represents a diverse residual group.3 This category represents an average of diverse experiences and is included for the sake of completeness rather than as an important point of comparison.

Overall subjective assessments of marital quality.-We examined two subjective measures of marital quality. Overall marital happiness is de- rived from a single measure (MHAPINES) that asked respondents to describe their relationship on a seven-point scale from very unhappy (coded "1") to very happy (coded "7"). A higher score indicates greater overall marital happiness. We also use a dichotomous recoded version of marital happiness, which distinguishes between those reporting very happy (7) marriages, now coded "1,"and all other marriages (1-6), now coded "0." We also examine a second overall indicator of marital qual-

About 25% did not live with both parents until age 16, because they left home early for school, for the military, for a job, to be on their own, or for marriage. Nearly another quarter (22%) had parents who were ill or otherwise unable to care for them. Others were removed from abusive families or left themselves, living with relatives or nonrelatives. And 24% of those in this category provided no reason why they did not live with both parents.

ity, rating how appreciated one feels in his or her spousal role (coded on

a scale from "appreciated = 1" to "unappreciated = 7").

Marital interaction.-Several variables assess relationships by asking about specific marital interactions. The summary index MEANARGU is the average frequency of open disagreements a respondent reports in seven areas: housework, money, spending time together, sex, having children, in-laws, and children (with childless couples having one less item on which their average is based). A higher score indicates a greater frequency of open disagreements. Other variables measure the style of conflict management: when spouses have a serious disagreement, how often do they (1) argue heatedly or shout at each other (SHOUT), (2) discuss disagreements calmly (CALMLY), and (3) keep their opinions to themselves (KEEPOPIN). A higher score indicates a greater frequency of shouting, calmly discussing disagreements, or keeping opinions to one- self, respectively. All of these measures were recoded into dichotomous variables, so that one can easily discern the frequency of particular be- haviors. For arguing, "1" indicates, on average, arguing several times a month or more (ARGUDMY). Shouting, calmly discussing disagree- ments, and keeping opinions to oneself were recoded (SHOUTDMY, CALMDMY, KEEPDMY) so that "1" indicates exhibiting that partic- ular behavior very often or always. Last, respondents were asked how often they end up hitting or throwing things at each other when they have a serious disagreement (EVERHIT; ever = 1, never = 0).

We examine two other marital interaction variables. ALONETIM re- ports how often a couple spends time alone with each other, talking or sharing an activity. Responses were recoded to distinguish between those who spend time alone with their spouse once per week or less (coded "0") and those who spend two or more times per week together (coded "1"; ALONEDMY). In addition, we examine perceptions of fairness in the relationship with respect to four domains: household chores, working for pay, spending money, and childcare. Those who perceive that the marital division of labor is unfair to them in any of those four areas are coded "1." We then summed these scores across the four domains, re- sulting in a scale from zero to four where a higher score indicates a greater number of domains in which the respondent feels the marital division of labor is unfair (UNFAIR). For variables that had both a continuous and dichotomous version (frequency of disagreements, shout- ing, calmly discussing disagreements, keeping opinions to oneself, and time alone), we examined both versions in data analyses but present only logistic results because OLS and logistic results are similar.

Perceived marital stability.-We examine two measures of perceived marital stability. MTROUBLE, a dichotomous variable indicative of a sense of uneasiness or concern with the marital relationship, reports

whether respondents had ever felt their marriage might be in trouble in

the past year (1 = yes). Past research has used this measure as an indica-

tor of marital instability (Booth, Johnson, and Edwards 1983; Booth

et al. 1985; Sabatelli 1988).

The second measure of perceived marital stability is the propensity to divorce, which was assessed by measuring the degree to which people are thinking about divorce. The perceived likelihood of separation or divorce (CHANCES) is based on respondents rating their chances of separation or divorce, from very low (1) to very high (5). The skewness on this variable is 2.2. Because some researchers argue against using OLS regression when the skewness is over 2.0 (Andrews 1991), we created a dichotomous version of the chances variable, so that we could also evalu- ate our models in a logistic regression format. The dichotomous version divides respondents into those who report that the chances of separation or divorce are either "very low" or "low" (coded 0) and those who say the chances are "about even," "high," or "very high" (coded 1).

Control and intervening variables.-We include in our models several background factors and indicators of social and economic conditions that prior research suggests are related either to marital satisfaction or rates of marital dissolution. Many researchers in this area argue that one's likelihood of divorce is determined not only by characteristics prior to and within one's marriage but also by the barriers to divorce and alternatives outside one's marriage (Levinger 1979; Lewis and Spanier 1979; Morgan and Rindfuss 1984; White and Booth 1991). The first group of controls includes age, gender, and racelethnicity, which are personal attributes of the individual and clearly precede other factors. Age has been shown to exhibit a curvilinear relationship to marital quality (Glenn 1990; Or- buch et al. 1992). We use a continuous version of age. Prior research also suggests blacks have higher rates of marital disruption than non-Hispanic whites (Cherlin 1981), while no differences were found between rates for Hispanic and non-Hispanic whites in a recent investigation (Bumpass et al. 1991). Hence a three-category version of racelethnicity was recoded into three dummy variables: non-Hispanic white (omitted), black, and all other racial or ethnic groups. For gender, "1" indicates female and "0" indicates male.

Our second group of variables includes education, couple's income, religion, age at marriage, cohabitation history, employment, parental status, and home ownership. Previous research indicates these variables are influenced by parental divorce and also directly influence the risk of marital dissolution. We include these variables in our models because they are potentially intervening factors and may explain why parental marital experience affects children's marriages. Prior research has shown that education is negatively related to marital dissolution (Bumpass and

Sweet 1972; Bumpass et al. 1991). Here education is a continuous mea-

sure of years of completed education. There is also a clear inverse rela-

tionship between income and other measures of socioeconomic status

and divorce (Fergusson, Horwood, and Shannon 1984; Greenstein 1985;

Martin and Bumpass 1989; South and Spitze 1986). Moreover, recent

data from a small rural sample indicated that economic pressure pro-

motes hostility in marital interactions, curtailing warm and supportive

spouse behaviors (Conger et al. 1990). The income measure bsed here is

a log of the sum of the wages of both the respondent and his or her


We examined religious affiliation with respect to several variables. We created dummy variables for the following religious affiliations: none, Catholic, Protestant, Jewish, and other. However, for the self-reported propensity to divorce variable (CHANCES), the only important variance by religious affiliation was found between those with no religious affilia- tion versus those with any religious affiliation. Hence for models of CHANCES, we employ a dichotomous religious affiliation variable (NORELIG; no religious affiliation = 1; any religious affiliation = 0).4 Previous studies have shown a strong positive relationship between age at marriage and marital stability (Kiernan 1986; Moore and Waite 1981; Teachman 1983). Moreover, research suggests children of divorce marry at younger ages than their peers from intact families (Bumpass and Sweet 1972; Glenn and Supancic 1984; Keith and Finlay 1988). We employ a continuous version of age at marriage.'

The literature also suggests that premarital cohabitation is associated with a higher probability of divorce (Balakrishnan et al. 1987; Bennett,

Our primary concern is how parental family structure influences subsequent marital quality and stability for children. Religious affiliation is only viewed as an important control variable where we first find substantial differences in the outcome variable by family history. Hence, while marital happiness was found to vary in important ways with religious affiliation, these affiliations were not of interest because no substantial differences in marital happiness by family history were found. Similar to the patterns found for CHANCES, reports of marital trouble were not found to vary in important ways between religious denominations, but rather only between those with any reli- gious affiliation vs. no religious affiliation.

We examined age at marriage in a number of alternate forms. Dividing the distribu- tion of ages at marriages into quartiles (for those in first marriages), we examined effects of these quartile dummies on the self-reported propensity to divorce (CHANCES) separately for men and women. Another test was conducted using a dummy to reflect an early age at marriage, defined as the first third of the NSFH distribution (separately by gender). No significant effects were found for either sex using these dummy variables (dichotomous version of early age at marriage or three dummies to account for the four quartiles). A linear relationship was found between age at marriage and the reported chances of divorce. Tests for the presence of an interaction between gender and age at marriage were, however, nonsignificant.

Blanc and Bloom 1988; Booth and Johnson 1988; White 1987). We in-

cluded two cohabitation measures: one is whether the respondent cohab-

ited with his or her current spouse prior to marriage (1 = yes); the other

assesses whether the respondent cohabited with someone other than his

or her current spouse prior to marriage (1 = yes).

It has been hypothesized that women's labor-force participation de-

creases marital stability (particularly encouraging wives to be less finan-

cially tied to their marriage), though the evidence on this point is equivo-

cal (White 1990). In terms of employment, we utilize a continuous version

of the average hours respondents reported spending per week in paid


Research suggests that the birth of a child appears to have a negative impact on most marriages, particularly for women (Belsky, Lang, and Rovine 1985; Belsky, Spanier, and Rovine 1983; Luckey and Bain 1970; Ryder 1973; Russell 1974), although there is some evidence this may be a duration of marriage effect rather than an effect of the transition to parenthood (Glenn 1990; McHale and Huston 1985; White and Booth 1985). Hence, in terms of parental status, we include two child variables: one indicates the presence of a child under the age of five in the household; the other indicates the overall number of children in the household. Last, we include a dichotomous indicator of home ownership (1 = homeowner), because material assets such as one's home are thought to influence marital cohesiveness (Levinger 1979).

Other than the control variables age, racelethnicity, and gender, which clearly precede other factors, most of the background factors dis- cussed above are not likely to be sources of spuriousness in the relation- ship between parental divorce and marital success (although religion, e.g., could be). Rather, variables such as cohabitation may help to ex- plain why and how parental divorce increases the risk of divorce in offspring.

Selectivity Correction

Of course, marital quality can only be assessed among those that are married (and here analyses are further restricted'to those in first mar-

We also examined a three-category version of employment status: working for pay full-time (30+ hourslweek, omitted); working part-time (1-29 hourslweek), or not employed for pay. We found that the continuous version of hours explained more variance in the self-reported propensity to divorce than the categorized version of employment status (full-time, part-time, none). We also examined the effect of hours in paid work separately by gender and only found a significant (positive) relationship between hours and chances of divorce for women. Again, tests for the presence of a statistical interaction between labor-force participation and gender were nonsignificant.

riages). Hence some observations (those individuals who got divorced)

have been excluded in a systematic manner, producing a nonrandom

sample that truncates the true distribution of marital quality and biases

the coefficients of interest. To address these concerns, we have followed

the procedures suggested by Heckman (1979) and others (e.g., Berk


First, we used a probit model to estimate the probability that an ever- married individual was in his or her first marriage. We estimated the probability using factors that have been shown to influence marital disso- lution: racelethnicity, education, age, age at marriage, childlessness, co- habitation prior to marriage, and childhood family structure.

Second, we used results from the selection model (the maximum likeli- hood coefficients and intercept) to assign each respondent a probability of being in his or her first marriage. From this variable, multiplied by -1 to capture the likelihood that one is no longer in his or her first marriage and to indicate the chance of being excluded from our analyses, we computed a lambda coefficient or hazard rate. Lambda was then entered into our substantive equations of interest (in tables 5 and 6) in order to correct for sample selection bias. It should be noted that with only one exception (noted below) including lambda in our models did not change the direction (positive or negative) of any significant predictor variables.


Diferences in background variables by family history.-Consistent with previous research, our examination of these data reveals that adults from any type of single-parent family are generally lower in socioeco- nomic status than those who grew up with two parents (results not shown). On average, the youngest groups are children of divorce and those who never lived with their father, while those who had a parent die are the oldest on average, consistent with recent increases in divorce and nonmarital childbearing and with decreases in mortality.

In two respects, the background of children of divorce appears unique. First, the proportion reporting no religious affiliation is twice as high among children of divorce as it is for any other family history category examined. Second, controlling for age, children of divorce are about twice as likely as those from two-parent families to have cohabited with someone other than their spouse (P < .001) or to have cohabited with their spouse (P< .001) prior to marriage, consistent with earlier research (Bumpass and Sweet 1989b; Thornton 1991; Furstenberg and Teitler 1994). The cohabitation rates of those who never lived with their father

are not significantly higher than rates of those from two-parent families

in models controlling for age.

Marital quality.-Table 1 shows global subjective assessments of re- spondents' marital relationship by family history. The data reveal no substantial differences in marital quality by childhood family history. There are no significant differences in average overall marital happiness by family history for the continuous version of the scale. This result is not changed by controlling for differences in age, socioeconomic status, and racelethnicity (results not shown). A dichotomized version of marital happiness, separating those who rated their marital happiness as "6" or "7" from all others, reveals a slight tendency for children of divorce and those who never lived with their father to report less happy marriages, but the differences are small. Similarly, average ratings of how appreci- ated respondents felt in their spousal role do not differ by family history. These results are similar to those reported by Amato and Booth (1991)' who found no significant differences between adult children of divorce and those from intact families when examining a marital happiness index based on happiness in 11 domains. Hence, one striking finding of this study is that we find no substantial differences in overall marital happi- ness or in feelings of appreciation by family history. Overall subjective assessments of the marital relationship do not appear to vary by parental family structure.

Perceived marital instability.-By contrast, we find important differ- ences in perceptions of marital stability by family history. First, the fre- quency of reporting marital trouble is strikingly higher for one family history group in particular (see table 1). The unadjusted figures suggest children of divorce are 70% more likely than those from two-parent fami- lies to fear that their marriage was in trouble in the past year. We do not find similarly elevated reports among those who never lived with their father, which is consistent with predictions from the interparental conflict perspective. It is also noteworthy that those who had a parent die in childhood are significantly less likely than others to think their marriage may be in trouble. We will return to these findings at a later point in our discussion.

In addition, we find considerable evidence of substantial and signifi- cant differences in the propensity to divorce by family history (see ta- ble 1). Continuous and dichotomous measures of the self-reported chances of divorce are significantly higher both for children of divorce and those who never lived with their father relative to adults from two- parent families. These results support the socialization perspective, which stresses the importance of one absent parent. These findings sug- gest there are important differences in the perceived stability of one's marriage by parental family structure.


Childhood Family Other Family Intact Parents Divorced Father Absent Parent Died Historya Indicators (1) (2) (3) (4) ($1

Marital happiness and appreciation:
Marital happinessb ...................................................... 6.00 5.90 5.89 5.99 5.94
% quite happy in marriagec ........................................... 75.8 70.8* 69.3* 74.6 75.4
Appreciation in spousal roled ......................................... 2.39 2.54 2.28 2.46 2.50
% quite appreciated in spousal rolee ................................ 62.0 58.4 63.6 61.7 59.3
Marital Trouble: Thought marriage in trouble in past year ......................... 18.6 31.7*** 23.1 12.9** 20.3 Chances of separation or divorcei ................................... 1.29 1.46*** 1.54*** 1.26 1.30 % whose chances of separation or divorce are about even or higher ......................................................... 6.7 14.5*** 15.5*** 4.9 6.1
N ............................................................................... 4,978 444 199 5 10 201

No~~.-Data for first marriages only; all figures are unadjusted. Significance tests reflect results relative to the intact family history category.
a Category includes those who, as children, did not live with both biological parents for reasons other than those included in cols. 2-4.
MHAPINES; coded on a scale from 1 (very unhappy) to 7 (very happy).
' MHAPINES = 6 or 7.
SPAPPREC; coded on a scale from 1 (appreciated) to 7 (unappreciated).
SPAPPREC = 1 or 2.
'CHANCES; coded on a scale from 1 (very low) to 5 (very high).

* P < .05.

** P < .01.

*** P < .001.


Childhood family intact ............................ 5.3 18.2 39.3 46.9
Parent died ......................................... 2.7 6.5** 21.8 58.4
Other family history ................................. 6.3 16.1 43.3 55.9
Father absent ......................................... 6.0 17.7 35.7 64.5
Parents divorced ..................................... 12.2*** 29.5** 49.6 75. I***

NOTE.-Data for first marriages only; all figures are unadjusted. Significance tests reflect results

relative to the intact family history category. Sample size for each classification is reported in table 3.

" MHAPINES; coded on a 7-point scale with 1 = very unhappy and 7 = very happy.

* P < .05.

** P < .01.

*** P < ,001.

Impact of marital quality on marital stability.-Our results indicate greater perceived marital instability among certain single-parent family history groups, despite few differences in overall marital quality. The next issue we address is whether this perceived instability appears regard- less of marital quality or only under conditions of poor marital quality.

Marital trouble.-Table 2 shows that at three of the four marital hap- piness levels (including both ends of the scale) adult children of divorce are significantly more likely than adults from two-parent families to fear their marriage may be in trouble. Thus, children of divorce are more likely to perceive their marriages as troubled regardless of their overall level of marital happiness.

Propensity to divorce.-By contrast, results in table 3 indicate there is an interaction between family history and marital quality, whereby the increase in the chances of divorce associated with certain family histories depends on one's marital quality. Table 3 reports the percentage of respondents who estimated their chances of divorce as about even or higher, by marital quality and family history. These data show that when marriages are reported to be very happy (as 45.4% of them are) no family history is associated with an increased propensity to divorce. Yet at lower levels of marital happiness, the proportion reporting that their chances of divorce are 50150 or higher is significantly greater both for children of divorce and those who never lived with their father. For example, among those in very unhappy marriages (see table 3) only about one in five adults from intact families rate their chances of divorce as about even or higher, compared with about one in two adults from divorced or father-absent backgrounds. Using OLS regression, conventional tests for


Childhood family intact .................... 1.1 (1,927) Parent died .................................... .8

(218) Other family historya ....................... .O


Father absent



Parents divorced ............................. 1.6

N .............................................. 2,733

NOTE.-Unadjusted data; first marriages only. Significance tests reflect results relative to the intact family history category. Sample size for each classification is reported in parentheses.

" MHAPINES; coded on a ?-point scale with 1 = very unhappy and 7 = very happy.

* P < .05.

** P < .01.

*** P < ,001.

the presence of a statistical interaction between family history and marital happiness are consistent with the above conclusions (F,,,,,, = 19.5; P < .001, for the entire sample of first marriages). Hence, consistent with arrow C in figure 1, these data indicate that the variation in perceived chances of divorce by family history is limited to those who are not extremely satisfied with their marriage (marital happiness = 1-6). While marital happiness is a fairly strong correlate of the propensity to divorce (r = -.41), family marital history markedly increases the propensity to divorce only among those in less happy marriages. Though prior research has treated both the marital trouble and chances of divorce question as measures of marital instability, both do not exhibit an interaction with overall marital quality (arrow C in fig. l), suggesting the two measures may not measure a single concept.

Marital interactions.-What explanations can account for the greater perceived chances of divorce among children of divorce and those who never lived with their father? The interparental conflict and intergenera- tional transmission perspectives discussed earlier (second explanation) lead us to ask how spouses relate to one another and whether these marital interactions vary by parental family structure. One consequence of observing interparental conflict may be that children learn interactions that escalate conflict and reduce communication in their own relation- ships. We examined differences by family history in the frequency of certain marital interactions, separately for those in the happiest marriages

(7) and all others (1-6), because the greater propensity to divorce for particular family histories is confined to marriages that are less than very happy. Among those in the happiest marriages, we find that the fre- quency of few marital interactions significantly differs by family history (results not shown).

In contrast, among those in less than very happy marriages, results reveal striking differences in marital interactions by family history, differ- ences that appear consistently for children of divorce. Table 4 presents results of logistic regressions of marital interaction variables on parental history, with and without controls for age, race, gender, and education.' First consider the results without controls. Among those in less than very happy marriages, children of divorce are more than twice as likely to argue frequently, shout when arguing, and hit when arguing. Also, adult children of divorce are significantly more likely to perceive the marital division of labor as unfair. In addition, children of divorce report calmly discussing disagreements significantly less often and have significantly less time alone as a couple. When controls (for age, race, gender, and education) are included, significant differences among adult children of divorce remain for all but three variables (frequency of arguments, time alone, and hitting). Hence, it appears that when children of divorce are less than very happily married, they more often escalate conflict and reduce communication with their spouse.' These results support the inter- parental conflict explanation and are consistent with conclusions from one recent study that children of divorce experience more marital dis-

'We examined both OLS and logistic regression models. One dichotomous variable examined, reports of ever hitting or throwing things, was not appropriate to examine in OLS given its frequency distribution and skewness. Because OLS results for the other variables are comparable to the results in table 4, we present only the results of logit models.

We assessed whether the impact of marital interactions on the propensity to divorce differs by family history. This assessment involved creating four interaction variables, which are the products of the four parental history dummies by each marital interac- tion variable. We then computed an F-test based on the improvement in chi-square when these four interactions were added to our base models (which included marital interaction and family history variables only). Our analyses provide no strong evidence that the impact of marital interactions on the propensity to divorce (CHANCES) differs by family history. Nevertheless, the hypothesis that particular marital interac- tions have a different impact on the propensity to divorce among those with different family histories should be considered in the future. Statistical interactions are difficult to detect in logistic regression, and, using our data, the sample size for particular cells became very small.



Parent Other Family Father Parents Marital Interactions (Endogenous) Died History Absent Divorced

Frequency of disagreements: Zero-order ..................................... With controlsc ..................................

Shout when arg~ing:~

Zero-order .......................................

With controls ...................................

Keep opinions to self:d

Zero-order .....................................

With controls ................................... Calmly discuss disagreement^:^ Zero-order ................................... With controls ................................... Feel division of labor unfair:'

Zero-order .......................................

With controls ................................... Time alone as a couple:' Zero-order ..................................... With controls ...................................

Hit when arg~ing:~

Zero-order ....................................

With controls ...................................

NOTE.-Unweighted N ranges from 2,594 to 2,732. Results shown are exponents of the logistic regression coefficients, which express the ratio of the odds for those who grew up in an intact family vs, single-parent categories.


On average, argue several times per month or more per area (e.g., housework, money, etc.).

' Age, race, gender, and education were included as controls.

Very often or always.

Feel division of labor is unfair to self on any of four fairness items.

'Two or more times per week.

8 Response other than "never hits or throws things when arguing."

* P < .05

** P < .01.

*** P < ,001.

agreements and marital problems than their peers from intact families (Booth and Edwards 1990).

Explanations for the Higher Perceived Marital Instability

Marital trouble.-Table 5 examines reports of marital trouble in a multivariate logistic format for all respondents in first marriages, because table 2 suggests differences appear at all levels of marital happiness. The


Zero-order ............... .89*** 2.42 .41* 1.51 -.24 .78 .44* 1.56
With controlsa .. . . . . .. . . .60*** 1.83 .17 1.18 -.13 .88 .24 1.27

NOTE.--Data include first marriages only; N = 4,689. "Intact" category omitted; all models include

a term to correct for sample selection bias.

" Age, race, gender, and education were included as controls.

* P < .05.

** P < .01.

*** P < ,001.

zero-order results suggest that relative to adults from two-parent families, children of divorce are more than twice as likely to say that in the past year they thought their marriage might be in trouble. When we control for age, race or ethnicity, gender, and education (see table S), the parental divorce coefficient is reduced by about one-third. Yet even with these controls included, children of divorce are still 83% more likely than those from two-parent families to fear their marriage may be in trouble. Even if one controls for other social and economic conditions (income, home ownership, age at marriage, cohabitation history, religious affiliation, employment, and parental status) and for marital interactions, children of divorce are still significantly (80%) more likely than those from intact families to fear their marriage may be in trouble (results not sh~wn).~

It is not clear from these analyses that any of the intervening variables considered here are important mediators in these processes. Since chil- dren of divorce are the only single-parent family structure who report marital trouble significantly more often, these results support the inter- parental conflict perspective over the socialization hypothesis. The ques- tion of thinking one's marriage is in trouble appears to capture a unique and ubiquitous legacy of parental divorce.

Propensity to divorce.-Last, we seek to identify what factors may explain the greater self-reported chances of divorce among adult children of divorce and those who never lived with their father. As explained previously, the models of the chances of divorce shown in table 6 will only examine respondents in less than very happy marriages. Table 6

The model from which this figure is taken also includes a lambda term to correct for sample selection bias.

Parents divorced ...........................................
Father absent ...............................................
Parent died ..................................................
Other family history ......................................
Age ............................................................
Gender .......................................................
Black .........................................................
Other racelethnicity .......................................
Education ....................................................
In Income ...................................................
Age at marriage ............................................
Cohabitation with spouse ................................


can be viewed in two ways. Some would argue that the self-reported chance of divorce is yet another subjective assessment of marital quality, which cannot be thought of as endogenous to other measures of marital quality. In that case, one would only want to interpret results in models 1-3, which examine CHANCES in a multivariate format controlling for a number of background factors and non-marital social and economic conditions.

Alternatively, one can view CHANCES as a measure of the propensity to divorce, which is influenced by assessments of one's marital relation- ship. We adopt this view because the propensity to divorce and marital quality variables are conceptually distinct and manifest different empiri- cal relationships to parental family structure and other variables. Grant- ing that propensity to divorce and marital quality are distinct variables, it also makes sense to view marital quality as determining propensity to divorce. If one accepts this viewpoint, as we do in table 6, then the empirical evidence primarily supports two explanations for the higher self-reported propensity to divorce among children of divorce in less than very happy marriages: demographic risk factors (age and race) and more frequent behavioral marital interactions that escalate conflict and reduce communication. However, these features of marital relationships do not explain the higher self-reported chances of divorce among those who never lived with their father.

First, consider logistic regression results for children of divorce (see table 6). The data show that children of divorce are disproportionately young and black, which increases the likelihood that they perceive their chances of divorce as about even or higher. The black coefficient was the one that became significant when the correction for sample selection bias was added (increasing from .42 to .5 1). When we adjust for composi- tional differences in age and racelethnicity (shown in model 2), these factors reduce the parental divorce coefficient by more than one-third (39%, from 1.13 to .69). Other background factors, such as religious affiliation, cohabitation history, age at marriage, income, home owner- ship, employment, and parental status (added in model 3), barely reduce the parental divorce coefficient further (to .62).1° Thus, these data do not support the notion that these variables intervene in the relation- ship between parental family structure and children's marriages. Hence, about one-third of the greater perceived chances of divorce by children of divorce is primarily explained by their age and racelethnicity. l1 When

lo In OLS models, the lack of religious affiliation and higher rates of cohabitation among children of divorce (particularly with their spouse before marriage) also contrib- ute to their higher perceived chances of divorce.

" While previous research has found some differences in the effects of parental divorce by gender, we identify no main effects of gender. Moreover, we also examined interac- tions between gender and family history, but none appeared significant and robust.

we control for background factors and nonmarital social and economic

conditions, children of divorce are still 86% more likely than those from

intact families to report that their chances of divorce are about even or


More important, the data also suggest that children of divorce more often engage in styles of marital interactions that are associated with a greater perceived chance of divorce. When we add styles of marital interaction, namely arguing frequently, escalating arguments (shouting, hitting or throwing things, not discussing disagreements calmly), keeping one's thoughts to oneself, not creating time alone as a couple, and feeling the household division of labor is unfair (model 4), the parental divorce coefficient is reduced again by more than two-thirds (. 19) to nonsignifi- cance.12 These results provide strong support for the explanation that when their marriages are less than very happy, children of divorce have a greater tendency than those with other family histories to adopt patterns of marital interaction that decrease the perceived stability of their mar- riages. These results are consistent with the hypothesis that the important intervening variable between parental divorce and the perceived marital stability of offspring may be the interactional style that offspring learn (Caspi and Elder 1988).

It should also be noted that controlling for marital interactions and characteristics in our model has an important effect on the coefficients associated with background factors and social conditions. For example, the higher perceived chances of divorce among blacks is cut in half (from .40 to .21) by controlling for styles of marital interaction. Also, the effect of cohabiting with someone other than one's spouse prior to marriage, while nonsignificant, declines from .36 to nearly zero, when measures of marital interactions are included. Moreover, the effect associated with number of children is reduced to nonsignificance when these patterns of relating to one's spouse are included. These results suggest that marital interactions may be important intervening variables in the relationship between background factors or other marital characteristics and divorce.

Next, consider results in table 6 for those who never lived with their father. Relative to children of divorce, the father-absent coefficient is slightly larger and somewhat more difficult to explain. Only one group of control variables substantially reduces the father-absent coefficient: background factors. The father-absent coefficient is reduced by more than one-third (from 1.25 to .74 from model 1 to 2) when we account

l2 Moreover, children of divorce more often have spouses who abuse drugs and alco- hol, which may indicate these spouses lack other coping mechanisms. Among those in less than very happy marriages, children of divorce are more than twice as likely as those from intact families to report that their spouse abuses drugs or alcohol.

for compositional differences in age and race (those with an absent father are disproportionately young and black). When we include other back- ground factors in the model, these factors reduce the father-absent coef- ficient only slightly (11%), which is similar to the ability of background factors to explain the higher propensity to divorce among children of divorce. Since we found few differences in marital interactions (table 3) for this group, it is not surprising that including marital interactions in the model did not reduce the father-absent coefficient much further. However, adding those variables (model 4) reduces the coefficient by 11% (from .66 to .59), which brings the father-absent coefficient to non- significance. Yet this reduction is much less than we find for adult chil- dren of divorce. In sum, a model that controls for background factors, social and economic conditions, and styles of marital interaction is able to explain the higher self-reported chances of divorce by those who never lived with their father. Race is the most important factor here, as prelimi- nary analyses showed that among respondents who never lived with their father, about one in four identify themselves as black, a much higher proportion than those in other family structure categories, reflecting the higher rates of nonmarital childbearing among African-Americans.


This study sought to improve our understanding of the long-term effects of childhood family structure on adult marriage, using the 1987188 Na- tional Survey of Families and Households (NSFH). This research extends previous work by comparing the marital quality and perceived marital stability of adults with several different types of single-parent family histories.

We find no substantial differences in overall subjective assessments of marital quality by family history. By contrast, measures of perceived marital stability vary significantly by family history, with children of divorce and those who never lived with their father most often expressing doubts about the stability of their marriage. In addition, the empirical evidence reveals an important finding about the conditions under which divorce is contemplated. The data suggest it is only when marriages become less than very happy that children of divorce and those who never lived with their father give divorce more serious consideration than their peers from intact families. These analyses show that the response to marital unhappiness appears to differ by parental family structure. Our results contradict the idea that persons from single-parent families universally consider divorce more often.

The evidence supports two explanations for the higher perceived chances of divorce among children of divorce. First, compositional differ-

ences in the age and racelethnicity of children of divorce contribute to

their greater marital instability. Second, the data indicate that children

of divorce report significant differences in marital interactions or patterns

of relating which suggest that they tend to escalate conflict and reduce

communication, thereby increasing the perceived chances of divorce.

This style of marital interaction appears to also partially explain the

relationship between other background factors (e.g., cohabitation, race,

and number of children) and one's propensity to divorce. The younger

average age and greater proportion of blacks among those who never

lived with their father also contributes to their greater reported chances

of divorce.

Comparisons across different single-parent family backgrounds pro- vide evidence about which causal mechanisms are most likely to be im- portant. Generally, across several measures of marital quality and stabil- ity, we found that adults who lost a parent to death do not differ from those from intact families, while those who never lived with their father most closely resemble children of divorce. It appears that under some conditions (e.g., when a parent dies) living with a single parent is not associated with any later marital problems for the offspring. Because our analyses reveal important differences among those who spent time in different types of single-parent families, we conclude that the socializa- tion hypothesis does not receive much empirical support.

One other pathway by which family history may influence adult mar- riage is the inappropriate modeling of spousal roles (Pope and Mueller 1976). The hypothesis that observing a poor parental marriage or inter- parental conflict may adversely affect the marital success of offspring is supported by these data. Children of divorce, more than those from any other single-parent family type, express the most doubts about their mari- tal stability (in addition to the greater perceived chances of divorce, they more often report marital trouble even when very happily married). Moreover, among those in less than very happy marriages, children of divorce are more likely than those with any other single-parent history to escalate conflict and reduce communication with their spouse, and those marital interactions are associated with an increased self-reported risk of divorce. Having lost a parent to death or never living with one's father are not associated with the same long-term effects.

We speculate that thinking one's marriage is in trouble may capture a unique legacy of parental divorce, which might be described as an under- lying insecurity or fear of repeating the experience of the previous genera- tion. Glenn and Kramer (1987) hypothesize that children of divorce have higher rates of divorce not because their marriages are less happy but rather because of a higher propensity to divorce that stems from a reluc- tance, given their fear of failure, to fully commit to marriage. Children of

divorce appear more likely to entertain reservations or experience anxiety

about their marriage, even when their average marital happiness is no

different from those from two-parent families. Future research needs to

explore the relationship between specific marital interactions and mea-

sures of overall subjective evaluations of marital quality by family


Despite our robust findings, there are several limitations to this study. First, our data cannot assess whether the marriages of parents from children of intact families were happy or conflicted. Because our findings imply that the inappropriate modeling of spousal roles may be important, we would predict that adults with intact family histories yet inappropri- ate models of parental marriage would also report differences in marital interactions and perceived marital stability (this hypothesis is supported by the research of Amato and Booth [1991]). Future research should also pursue how remarriage, role models provided by stepparents, and the affective nature of a second (or higher order) parental marriage influence the marital quality and marital stability of adult children.

In addition, we are attempting to understand a process that occurs over time (a process leading to divorce), using data that are cross-sectional. In the future, we hope to gain a better understanding of the process leading to divorce and the intergenerational transmission of divorce proneness by using longitudinal data. Self-report measures are subject to response bias. However, the Bumpass et al. (1991) study that examined marital dissolution with these data reports an elevated risk of divorce among persons with the same family backgrounds one would predict from our analyses, namely children of divorce and those who never lived with their father. Hence, while the question of how well self-reported propensity to divorce predicts actual marital dissolution will have to await future re- search, there is reason to believe self-reports may be important as pre- dictors of divorce.

One difficulty in studying marital relationships and the process leading to divorce is that those with the greatest propensity to divorce will do so, eliminating themselves from the group of married individuals under study. This process, known as selection, led us to correct for sample selection bias. For the tables that did not include such corrections, one should keep in mind that selection biases are likely to produce a rosier picture of the marriages of children of divorce in cross-sectional data like those here than is in fact correct, because selection is likely to remove a greater proportion of poor marriages among those with family histories that increase the risk of divorce. Selection may be less important for studies of family structure in higher-order marriages, because we found that the effects of parental family structure were considerably weaker among those in second and higher order marriages and appear to dimin- ish with time.

Moreover, our study cannot evaluate all potential sources of spurious- ness. It may be that something about the parental home or parental behavior before divorce causes both parental divorce and a heightened propensity to divorce among the offspring of the divorced couple. For example, values from one's family of origin or parental personality (as Caspi and Elder [I9881 hypothesize) may influence both parental divorce and the offspring's marriage.

As the proportion of adults who spent part of their childhood living in a single-parent family grows, the relations of childhood family history to adult experiences of marriage and divorce, and explanations of these relations, should become more central to research on marriage and the family.


Amato, Paul. 1993. "Children's Adjustment to Divorce: Theories, Hypotheses, and Empirical Support. " Journal of Marriage and the Family 5 5:23-38. Amato, Paul, and Alan Booth. 1991. "Consequences of Parental Divorce and Marital Unhappiness for Adult Well-Being." Social Forces 69395-914. Andrews, Frank M. 1991. Lecture in Sociology/Psychology 726, "Multivariate Analy- sis Methods." University of Michigan, Ann Arbor.

Balakrishnan, T. R., K. Vaninadha Rao, Evelyne Lapierre-Adamcyk, and Karol J. Krotki. 1987. "A Hazard Model Analysis of the Covariates of Marriage Dissolution in Canada." Demography 24:395-406.

Belsky, Jay, Mary E. Lang, and Michael Rovine. 1985. "Stability and Change in Marriage across the Transition to Parenthood: A Second Study." Journal of Mar- riage and the Family 47:855-65.

Belsky, Jay, Graham B. Spanier, and Michael Rovine. 1983. "Stability and Change in Marriage across the Transition to Parenthood." Journal of Marriage and the Family 45567-77.

Bennett, Neil G., Ann Klimas Blanc, and David E. Bloom. 1988. "Commitment and the Modern Union: Assessing the Link between Premarital Cohabitation and Subsequent Marital Stability." American Sociological Review 53: 12 7-38.

Berk, Richard A. 1983. "An Introduction to Sample Selection Bias in Sociological Data." American Sociological Review 48:386-98. Bianchi, Suzanne M. 1990. America's Children: Mixed Prospects. Population Bulle- tin, vol. 45, no. 1.

Booth, Alan, and John N. Edwards. 1990. "Transmission of Marital and Family Quality over the Generations: The Effects of Parental Divorce and Unhappiness." Journal of Divorce 13:41-58.

Booth, Alan, and David Johnson. 1988. "Premarital Cohabitation and Marital Suc- cess." Journal of Family Issues 9:255-72. Booth, Alan, David Johnson, and John N. Edwards. 1983. "Measuring Marital Insta- bility." Journal of Marriage and the Family 45:387-93. Booth, Alan, David R. Johnson, Lynn K. White, and John N. Edwards. 1985. "Pre- dicting Divorce and Permanent Separation." Journal of Family Issues 6:331-46.

Bumpass, Larry L. 1984. "Children and Marital Disruption: A Replication and Up- date." Demography 2 1:7 1-82.

Bumpass, Larry L., Teresa Castro Martin, and James A. Sweet. 1991. "The Impact of Family Background and Early Marital Factors on Marital Disruption." Journal of Family Issues 12:22-42.

Bumpass, Larry L., and James A. Sweet. 1972. "Differentials in Marital Instability: 1970." American Sociological Review 37:754-66. . 1989a. "Children's Experience in Single-Parent Families: Implications of cohabitation and Marital Transitions. Family Planning Perspectives 21:256-60. . 1989b. "National Estimates of Cohabitation." Demography 26:615-25.

Caspi, Avshalom, and Glen H. Elder, Jr. 1988. "Emergent Family Patterns: The Intergenerational Construction of Problem Behavior and Relationships." Pp. 218-40 in Relationships within Families, edited by Robert A. Hinde and Joan Stevenson-Hinde. New York: Oxford University Press.

Chase-Lansdale, P. Lindsay, and E. Mavis Hetherington. 1990. "The Impact of Divorce on Life-Span Development: Short and Long-Term Effects." Pp. 105-51 in L$e Span Development and Behavior, edited by David L. Featherman and Richard M. Lerner. Hillsdale, N.J.: Lawrence Erlbaum Associates.

Cherlin, Andrew J. 1981. Marriage, Divorce, Remarriage. Cambridge, Mass.: Har- vard University Press.

Conger, Rand D., Glen H. Elder, Jr., Frederick 0. Lorenz, Katherine J. Conger, Ronald L. Simons, Les B. Whitbeck, Shirley Huck, and Janet N. Melby. 1990. "Linking Economic Hardship to Marital Quality and Instability." Journal of Mar- riage and the Family 52:643-56.

Fergusson, D. M., L. J. Horwood, and F. T. Shannon. 1984. "A Proportional Hazards Model of Family Breakdown." Journal of Marriage and the Family 46: 539-49.

Furstenberg, Frank F., Jr., and Andrew J. Cherlin. 1991. Divided Families. Cambridge, Mass.: Harvard University Press.

Furstenberg, Frank F., and Julien 0. Teitler. 1994. "Reconsidering the Effects of Marital Disruption: What Happens to Children of Divorce in Early Adulthood?" Journal of Family Issues 15:173-90.

Glenn, Norval D. 1990. "Quantitative Research on Marital Quality in the 1980s: A Critical Review." Journal of Marriage and the Family 52:818-31. Glenn, Norval D., and Kathryn B. Kramer. 1987. "The Marriages and Divorces of the Children of Divorce." Journal of Marriage and the Family 49:811-25.

Glenn, Norval D., and Michael Supancic. 1984. "The Social and Demographic Corre- lates of Divorce and Separation in the United States: An Update and Reconsidera- tion." Journal of Marriage and the Family 46563-76.

Greenberg, Ellen F., and W. Robert Nay. 1982. "The Intergenerational Transmission of Marital Instability Reconsidered." Journal of Marriage and the Family 44: 335-47.

Greenstein, Theodore N. 1985. "Occupation and Divorce." Journal of Family Issues 6:347-57. Heckman, James J. 1979. "Sample Selection Bias as a Specification Error." Econometric~47:153-61.

Kalter, Neil, Amy Kloner, Shelly Schreier, and Katherine Okla. 1989. "Predictors of Children's Postdivorce Adjustment." American Journal of Orthopsychiatry 59: 605-18.

Keith, Verna M., and Barbara Finlay. 1988. "The Impact of Parental Divorce on Children's Educational Attainment, Marital Timing, and Likelihood of Divorce." Journal of Marriage and the Family 50:797-809.

Kiernan, Kathleen E. 1986. "Teenage Marriage and Marital Breakdown." Population Studies 40:35-44.

Kobrin, Frances E., and Linda J. Waite. 1984. "Effects of Childhood Family Struc- ture on the Transition to Marriage." Journal of Marriage and the Family 46: 807-16.

Kulka, Richard A,, and Helen Weingarten. 1979. "The Long-Term Effects of Paren- tal Divorce in Childhood on Adult Adjustment." Journal of Social Issues 35:50


Levinger, G. 1979. "A Social Psychological Perspective on Marital Dissolution." Pp. 36-70 in Divorce and Separation, edited by G. Levinger and 0. C. Moles. New York: Basic.

Lewis, Robert A,, and Graham Spanier. 1979. "Theorizing about the Quality and Stability of Marriage." Pp. 268-94 in Contemporary Theories about the Family, vol. 1. Edited by W. Burr, R. Hill, F. I. Nye, and I. Reiss. New York: Free Press.

Luckey, Eleanore Braun, and Joyce Koym Bain. 1970. "Children: A Factor in Marital Satisfaction." Journal of Marriage and the Family 32:43-44. Martin, Teresa Castro, and Larry L. Bumpass. 1989. "Recent Trends in Marital Disruption." Demografihy 26:37-5 1.

McHale, Susan M., and Ted L. Huston. 1985. "The Effect of the Transition to Parenthood on the Marriage Relationship: A Longitudinal Study." Journal of Fam- ily Issues 6:409-34.

McLanahan, Sara. 1985. "Family Structure and the Reproduction of Poverty." American Journal of Sociology 90373-901. Moore, Kristin, and Linda J. Waite. 1981. "Marital Dissolution, Early Motherhood, and Early Marriage." Social Forces 60:20-40. Morgan, S. Philip, and Ronald R. Rindfuss. 1984. Marital Disruption: Structural and Temporal Dimensions." American Journal of Sociology 90: 1055-7 7.

Mueller, Charles W., and Hallowell Pope. 1977. "Marital Instability: A Study of Its Transmission between Generations." Journal of Marriage and the Family 39: 83-93.

Orbuch, Terri L., James S. House, Rich P. Mero, and Pamela S. Webster. 1992. "Marital Quality through the Lifecourse." Paper presented at the annual meeting of the American Sociological Association, Pittsburgh, Pennsylvania.

Pope, Hallowell, and Charles W. Mueller. 1976. "The Intergenerational Transmission of Marital Instability: Comparisons by Sex and Race." Journal of Social Issues 32149-66.

Russell, Candyce Smith. 1974. "Transition to Parenthood: Problems and Gratifica- tions." Journal of Marriage and the Family 36:294-302. Ryder, Robert G. 1973. "Longitudinal Data Relating Marriage Satisfaction and Hav- ing a Child." Journal of Marriage and the Family 35:604-06.

Sabatelli, Ronald M. 1988. "Measurement Issues in Marital Research: A Review and Critique of Contemporary Survey Instruments." Journal of Marriage and the Fam- ily 50:891-915.

Select Committee on Children, Youth, and Families. 1989. U.S. Children and Their Families: Current Conditions and Recent Trends, 1989. Washington, D. C. : U.S. Government Printing Office.

South, Scott, and Glenna Spitze. 1986. "Determinants of Divorce over the Marital Lifecourse." American Sociological Review 51:583-90.

Sweet, James A,, Larry L. Bumpass, and Vaughn Call. 1988. "The Design and Content of the National Survey of Families and Households." NSFH Working Paper no. 1. Center for Demography and Ecology, University of Wisconsin, Madison.

Teachman, Jay D. 1983. "Early Marriage, Premarital Fertility, and Marital Dissolu- tion: Results for Blacks and Whites." Journal of Family Issues 4:105-26. Thornton, Arland. 1991. "Influence of the Marital History of Parents on the Marital

and Cohabitational Experiences of Children." American Journal of Sociology 96: 868-94. Uhlenberg, Peter. 1974. "Cohort Variations in Family Life Cycle Experiences of U.S. Females." Journal of Marriage and the Family 36:284-92. Wallerstein, Judith S., and Sandra Blakeslee. 1989. Second Chances: Men, Women and Children a Decade after Divorce. New York: Ticknor & Fields. White, James. 1987. "Premarital Cohabitation and Marital Stability in Canada." Journal of Marriage and the Family 49:641-47. White, Lynn K. 1990. "Determinants of Divorce: A Review of Research in the Eight- ies." Journal of Marriage and the Family 52:904-12. White, Lynn K., and Alan Booth. 1985. "Transition to Parenthood and Marital Quality." Journal of Family Issues 6:43S-SO. . 1991. "Divorce over the Life Course: The Role of Marital Happiness." Journal of Family Issues 12:s-2 1.

  • Recommend Us