hypothesis about divorce

The Causal Effects of Parental Divorce and Parental Temporary Separation on Children’s Cognitive Abilities and Psychological Well-being According to Parental Relationship Quality

• Original Research
• Open access
• Published: 27 July 2020
• Volume 161 , pages 963–987, ( 2022 )

Cite this article

You have full access to this open access article

• Anna Garriga 1 &
• Fulvia Pennoni   ORCID: orcid.org/0000-0002-6331-7211 2  

52k Accesses

14 Citations

12 Altmetric

Explore all metrics

We explore the effects of parental divorce and parental temporary separation on well-being of children at a specific stage of their development according to the parental relationship quality. Despite the importance of this subject, among previous studies only few consider very young children and are based on statistical methods properly tailored to enhance causal evaluations. We attempt to establish the effects on both cognitive abilities and psychological dimensions of children at age five by using data drawn from the first three waves of the UK Millennium Cohort Study. Using an appropriate imputation method, we apply the augmented inverse propensity treatment weighted estimator to infer causality. Overcoming some of the limitations of previous research, we find that the dissolution of high-quality parental unions has the most harmful effects on children, especially concerning conduct problems. We demonstrate the substantial variation on consequences of parental divorce depending on the level of parental relationship quality. We show that parental temporary separation is a type of family disruption that has significant negative effects on young children. In fact, we infer that they have more conduct and hyperactivity problems than children from stable or divorced families. Our results also suggest children to be targeted with appropriate policies aimed to reduce the adverse effect of family disruption.

Similar content being viewed by others

Parental breakup and children’s development: the role of time and of post-separation conditions

Chiara Pronzato & Arnstein Aassve

Floor Effects or Compensation of Social Origin? The Relation Between Divorce and Children’s School Engagement According to Parents’ Educational Level

Childhood behavioral problems are associated with the intergenerational transmission of low education: a 16-year population-based study.

Adam Vanzella-Yang, Francis Vergunst, … Sylvana Côté

Avoid common mistakes on your manuscript.

1 Introduction

Parental divorce and union dissolution is an increasingly common experience for children in all developed countries. It has raised the debate on whether parental divorce is damaging for children’s well-being and to what extent parents should remain together for the sake of the children. In accordance with this social concern, one of the most extensively discussed topics in the literature has been the average effects of divorce on children well-being. Many social surveys have been considered and various statistical methods have been used putting special emphasis on controlling for parental relationship quality and conflict prior to separation but often without considering that parental conflict does not always precede separation. In fact, a large percentage of low-distress couples divorce, a phenomena that has increased substantially in recent decades (Gähler and Palmtag 2015 ).

For this reason, some studies in the last two decades have offered a more nuanced explanation of the causality of divorce that focuses on the heterogeneity of divorce effects by parental relationship quality (Amato et al. 1995 ; Jekielek 1998 ; Hanson 1999 ; Strohschein 2005 ). They take into account in which way divorce affects different children, either positively or negatively, instead of concentrating on the average causal effect of divorce across the board (Amato 2010 ). These studies suggest that divorce may be a positive experience for children from high-distress marriages, while the dissolution of low-distress marriages may have opposite effects (Amato et al. 1995 ; Booth and Amato 2001 ). Despite the significant ramifications of these findings, to the best of the author’s knowledge only few studies have examined the heterogeneity of the consequences of parental divorce by the level of parental relationship quality. These studies (for instance, see: Amato et al. 1995 ; Jekielek 1998 ; Hanson 1999 ; Morrison and Coiro 1999 ; Booth and Amato 2001 ; Strohschein 2005 ; Fomby and Osborne 2010 ; Yu et al. 2010 ; Kalmijn 2015 ) present the following characteristics: a ) only two are based on non-US data, and only few use nationally representative samples or methods to infer causality; b ) they have mainly analyzed children’s psychological well-being while evidence on other children’s outcomes such as cognitive development is scarce; c ) the most of them focus solely on children in middle childhood or older. Amato ( 2010 ), in his most recent review of the literature, encourages more research concerning this issue. We use a national large sample with several cases of divorce and temporary separations among parents and cohabiting couples representing the current trend in the Western societies. We explore the effects both on the psychological well-being and cognitive development of the children. We investigate if and how these effects are different according to parental relationship quality. The latter is measured in a way to capture not only parental conflicts before separation but also communication, affection and emotions among the couple. We use the first three waves of the Millennium Cohort Study (MCS) which is a nationally representative longitudinal study of a cohort of British children born from 2000 to 2002 in the UK. We move forward from previous work and contribute in respect to the analysis of the interrelationships between family disruption, parental relationship quality and children’s psychological well-being and cognitive development. First, we test whether the hypothesis of heterogeneity of divorce effects by parental relationship quality is also true for children from UK. Second, we aim to assess whether this hypothesis is also valid for young children since only Fomby and Osborne ( 2010 ) account for very young children and do not find evidence of heterogeneity of divorce by parental relationship quality. We have specifically focused on a salient period of children’s lives, namely the transition to school. It is well-demonstrated that children who enter school without the necessary cognitive or socio-emotional skills have greater academic and behavioral difficulties during their school years and beyond than their more “school-ready” counterparts (Romano et al. 2010 ). Third, we aim to assess the heterogeneity hypothesis by improving and extending the methodological and analytical approach proposed in the literature. We focus on many of children’s outcomes rather than just on one or two since we consider the following multiple dimensions of children’s school readiness: three different cognitive abilities (verbal, problem-solving and spatial abilities) and five psychological dimensions (conduct, hyperactivity, internalizing and peer problems, and pro-social behavior). Unlike previous research on parental divorce, we use the Augmented Inverse Propensity Treatment Weighted (AIPTW, Robins et al. 2000 ) estimator in order to yield robust estimates of the effects of interest and an imputation method based on the statistical methodology of chained equations (Raghunathan et al. 2001 ), which allows us to jointly impute missing data for different types of variables. Furthermore, in previous research on the interplay between parental divorce, parental relationship quality and children’s outcomes, the fact that a significant proportion of parents separate only temporarily was not considered. Little is known about the level of relationship quality of these parents before separation and the risks children experience when they face this type of family disruption (Kiernan et al. 2011 ; Nepomnyaschy and Teitler 2013 ).

In sum, by using cohort data similar to that used by Fomby and Osborne ( 2010 ), we aim to test the following three hypotheses: i ) parental relationship quality and family disruption are unrelated processes that have independent effects on children (the independent hypothesis); ii ) the negative association between family disruption and children’s well-being may be spurious because poor relationship quality is related to both family disruption and poor children well-being (the selection hypothesis); iii ) the consequences of family disruption on children are contingent on the level of parental relationship quality experienced prior to this event (the heterogeneity hypothesis). To test this third hypothesis about the heterogeneity of the effects of family disruption by parental relationship quality is the main contribution of our study.

The paper is structured as follows. In Sect.  2 we provide a background section considering the effects of family disruption on children well-being, then we focus on the conceptual framework. In Sect.  3 we describe data and methods. In Sect.  4 we show the main results. In Sect.  5 we provide a discussion. The supplementary material provides further details on data, missing data imputation and additional tables with results.

2 Background

2.1 family disruption on children well-being.

Studies on the effects of parental divorce on children’s well-being that use ordinary least squares (OLS) and logistic models show that part of this effect is spurious and it is only partially explained by parental relationship quality (Hanson 1999 ). Since the late 1990s, several studies have used more innovative research designs to identify the independent effects of parental divorce and father absence such as lagged dependent variable models, growth curve models, individual and sibling fixed effects models, natural experiments and instrumental variables, and propensity score matching. McLanahan et al. ( 2013 ) review these studies and find consistent evidence that parental divorce exerts negative effects on the well-being of offspring. They also show that this evidence is stronger for children’s socio-emotional development, especially in externalizing problems, than for children’s cognitive ability. Nevertheless, they present the following features: a ) most studies of the effect that of parental divorce on cognitive and psychological development are based on US samples b ) very few studies focus on children who experience parental divorce in early childhood, and c ) only one (Strohschein 2005 ) explores the heterogeneity of divorce effects by the quality of the parental relationship prior to separation.

A weakness of existing research is that it does not consider parents who separate only temporarily. Recent studies observe that a non-negligible proportion of parents separate for a short time period and then re-partner with the same person (Kiernan et al. 2011 ; Nepomnyaschy and Teitler 2013 ). However, as stated by Nepomnyaschy and Teitler ( 2013 , 3) “in most studies, this family ‘type’ is usually classified as either intact or separated (depending on when cohabitation status is ascertained), but it may differ in many respects from both of those groups”. The reason this type of family disruption is scarcely considered in previous research is that most studies only use two waves of survey data, and at least three waves are necessary to detect it. The existing research on the characteristics of parents who only separate temporarily show that such couples have a more disadvantaged socio-demographic background than with continuously intact relationships (Kiernan et al. 2011 ; Nepomnyaschy and Teitler 2013 ). Despite that, the only two studies that analyze the consequences of temporary separation on children well-being find evidence of a negative effect, even when controlling for several socio-demographic characteristics (Kiernan et al. 2011 ; Nepomnyaschy and Teitler 2013 ). However, no study controls for the relationship quality of parents before separation and some evidence suggest that couples who separate temporarily have a lower relationship quality than stable couples long before separation occurs (Vennum et al. 2014 ). Thus, if poor parental relationship quality may cause temporary separation, then it is difficult to rule out the possibility that the negative association between parental temporary separation and children’s outcomes may be due to relationship quality rather than this event per se.

2.2 Heterogeneity of the Effects of Parental Divorce by Parental Relationship Quality

2.2.1 conceptual framework.

Two main explanations are provided regarding the heterogeneity of the effects of parental divorce Footnote 1 by parental relationship quality. One is the stress relief hypothesis (Wheaton 1990 ) which concerns the consequences of transitions in life roles. Wheaton ( 1990 , 210) stated that “…instead of being stressful, life events may at times be either non-problematic or even beneficial, offering escape from a chronically stressful role situation, creating the apparent paradox of more ‘stress’ functioning as stress relief”. According to this perspective, the stressful event of parental divorce may be beneficial for children whose parental relationship prior to divorce has been poor, as it takes them away from an aversive and stressful home environment. After divorce, these children should enjoy an improvement in their well-being since they no longer experience the parental conflict (Booth and Amato 2001 ; Strohschein 2005 ).

By contrast, the dissolution of low-distress parental relationships may be detrimental to children’s development. Children from relatively harmonious families may not benefit from divorce, since it is unlikely that they experience this event as stress relief. For these children, divorce may instead give rise to stressful situations such as a decline in their standard of living, moving to a poorer neighborhood, changing schools, and losing contact with the non-custodial parent (Amato 2010 ). Children from non-dysfunctional families may also begin to experience parental discord after separation, since issues such as custody, childrearing, visitation, and child support are potentially conflictual (Booth and Amato 2001 ).

In addition to changes in stress, children’s understanding and perceptions of divorce depend on the level of their parents’ pre-divorce relationship problems, another factor related to children’s adjustment after separation. Children who have witnessed parental disputes may anticipate their parents’ divorce and attribute it to external reasons, such as parental conflict, as argued by Booth and Amato ( 2001 ). For children from low-distress families, by contrast, divorce might come as more of a surprise and they might see divorce as a threat to their happiness. Booth and Amato ( 2001 ) give possible reasons as to how an unexpected divorce may adversely impact on children. First, for these children, it is more difficult to comprehend and accept the reasons for their parents’ separation. As Maes et al. ( 2012 , 276) state: “if children do not understand why their parents have divorced, they make up their own story around things they do know, increasing the danger that children will blame themselves”. Second, children who do not anticipate parental divorce may feel that they have little control over events in their lives (Booth and Amato 2001 ). Children’s self-blame and locus of control are, in turn, negatively related to their adjustment after divorce (Bussell 1996 ; Kim et al. 1997 ).

Are these useful in explaining the heterogeneity of parental divorce for infants and very young children? The explanation of children’s understanding and perceptions of divorce is unlikely to be valid for very young children due to the kind of reasoning needed for children to be able to anticipate this event and blame themselves for it. The stress relief explanation developed for older children and adults, however, can also be applied to infants and very young children. There is a growing and consistent body of research documenting that the exposure to poor parental relationship quality during infancy affects children’s well-being cross-sectionally during infancy and longitudinally during their pre-school years (Fitzgerald 2010 ; Graham et al. 2013 ; Zhou, Cao and Leerkes 2017 ). For example, a possible mechanism is that parental conflict experienced by infants is associated with neural responses to emotional tone of voice, particularly very angry speech (Graham et al. 2013 ). With the existing evidence, it is reasonable to assume that if parental conflict produces stress in infants, then when parental divorce occurs this source of stress will disappear and their well-being will improve.

The second explanation does not focus on the consequences of direct exposure of infants to parental conflict but highlights an indirect pathway through parental well-being: parental relationship quality moderates the effect of parental divorce on very young children because parental relationship quality also moderates the effect of divorce on parent’s well-being. To our knowledge, this explanation has not been mentioned by previous research and is based on two main premises. First, it has been largely demonstrated that parents’ emotional adjustment after divorce is an important predictor of children’s well-being (Amato 1993 ) and that parents’ emotional problems are also clearly associated with adverse children’s outcomes during infancy and early childhood (Petterson and Albers 2001 ; Kiernan and Huerta 2008 ). In addition to that, few empirical studies that have focused on this topic predominantly show that people who enjoyed a high relationship quality prior to divorce suffer the most harmful negative effects on their emotional well-being (for instance, see Wheaton 1990 ; Booth and Amato 2001 ; Williams 2003 ; Waite et al. 2009 ; Ye et al. 2017 ). For people with low levels of relationship quality, the findings are mixed. Some studies give support to the hypothesis that divorce is beneficial for the emotional well-being of people in highly conflictual or unsatisfactory relationships (for instance, see: Wheaton 1990 ; Williams 2003 ; Amato and Hohmann-Marriott 2007 ; Ye et al. 2017 ). Others find evidence that when people divorce from an unsatisfactory relationship, they experience a decrease in their emotional well-being but to a lesser extent than those who divorce from satisfactory relationships (Kalmijn and Monden 2006 ; Waite et al. 2009 ). For these reasons, it seems plausible to hypothesize that if divorce has the most harmful effects on parents who enjoyed a high level of relationship quality, their children would also experience the most harmful effects of this event.

2.2.2 Previous Research

To the extent of our knowledge the current research is based on the possible data according to the characteristics of the sample at the time and on the best method of analysis. Confidence in research findings increases when studies are based on a nationally representative sample with a large sample size. Some studies have less than 300 cases in the divorce group, and only three (Hanson 1999 ; Strohschein 2005 ; Kalmijn 2015 ) use nationally representative surveys. The majority of samples are based on American children, with the exception of Kalmijn ( 2015 ) and Strohschein ( 2005 ), and there is not enough evidence to conclude the hypothesis of heterogeneity of divorce effects is valid in all Western countries or if this hypothesis is country-specific. With exception of Fomby and Osborne ( 2010 ), all relevant studies examine only children whose parents are married; they exclude the large and increasing proportion of children who are living with their biological cohabiting parents (Kiernan et al. 2011 ).

Concerning the characteristics of the outcomes and focal variables we observe that seven of the nine studies in this field used the psychological well-being of offspring; there is less consistent evidence of variation in divorce effects in other important outcomes. Among studies concerning the heterogeneity of divorce, only one focus on educational achievement (Hanson 1999 ). For this reason, with the existing research, it is not possible to say whether the hypothesis about the heterogeneity of divorce effects is valid for most children’s outcomes, or only for psychological ones.

In addition, existing research does not focus on a specific stage of children’s development. Instead, samples are used with great variation in the children’s ages at the time of divorce, and the age when response variables are measured. Most studies look at children who experienced parental divorce over a wide range of ages (Booth and Amato 2001 ; Hanson 1999 ; Kalmijn 2015 ). In some of them, divorce occurred any time from when the children were born to when they were adults. Only Fomby and Osborne ( 2010 ) focus on a specific stage of children’s development namely parental divorce that occurs before age 3, and the response variable is measured at age 3. Second, as mentioned, studies finding evidence in favor of the heterogeneity hypothesis analyze children’s outcomes measured during middle childhood and/or adolescence (Hanson 1999 ; Jekielek 1998 ; Morrison and Coiro 1999 ; Strohschein 2005 ) or adulthood (Amato et al. 1995 ; Booth and Amato 2001 ; Yu et al. 2010 ; Kalmijn 2015 ) and the only paper that does not support this hypothesis focuses on outcomes in very young children (Fomby and Osborne 2010 ). These contradictory results may suggest that the effects of divorce only vary by parental relationship quality for children in middle childhood or older. However, with only one study on very young children, there is not enough evidence to conclude whether divorce effects are heterogeneous depending on the age of the child at the time of divorce and/or the age when the outcomes were measured.

3 Materials and Methods

The data correspond to the first three waves of Millennium Cohort Study (MCS) which is a high-quality profile survey representative for the UK (Plewis et al. 2000 ; Hansen and Joshi 2007 ; Plewis 2007 ; Hansen et al. 2012 ). The first sweep was carried out between September 2000 and January 2002. It contains information on 18,819 babies from 18,533 families, collected from the parents when the babies were 9–11 months old. The families were contacted again when the children were aged 3 and 5 years. The response rates achieved for the second (2004/05) and third (2006) waves were 78% and 79% of the target sample, respectively. More than two-thirds of the sample (around 69% representing 13,234 families) responded in all three waves (Ketende 2010 ). The MCS sample design allowed for over-representation of families living in areas with high rates of child poverty and/or high proportions of ethnic minorities. Survey methods were used to take account of the initial sampling design, and adjustments were made for non-response in the recruitment of the original sample and sample attrition over the follow-up period to age five. Footnote 2

We consider children whose family structure is available for all the first three waves of the MCS. The sample includes only singleton children and families where the mother is the main respondent at the first wave. More details on the data and on how we handle missing values are available in Section A1 of the supplementary material.

3.2 Variables

3.2.1 response variables.

The variables of interest for school readiness are measured when children are 5 years old, at the third wave. The Strengths and Difficulties Questionnaire (SDQ) (Goodman 1997 ) assesses children’s behavioral adjustment and is answered by the mother. The SDQ is made up of five subscales assessing emotional symptoms, conduct problems, hyperactivity or inattention problems, peer problems, and pro-social behavior. Each subscale has five items with scores ranging from 0 to 2. Children’s cognitive development is assessed by using the British Ability Scales (BAS II) (Elliott et al. 1997 ). The following BAS subscales were used to measure different domains of cognitive development: the naming vocabulary test, which assesses expressive language; the picture similarities test, which measures pictorial reasoning; and the pattern construction test, which assesses spatial ability. These were conducted by an interviewer at home. The three tests assess the three most significant information-processing skills: verbal reasoning, non-verbal reasoning and spatial abilities (Hill 2005 ). A standardized score is computed for each cohort member according to his/her age band considered every three months. Table  1 shows the average scores for the response variables stratified according to the family situation at age 5 as defined in the following section. Children experiencing parental temporary separation or parental divorce shows lightly more psychological problems and lower scores for cognitive development with respect to children with stable family.

3.2.2 Focal Variables

We use the first three waves of the survey to create the following main family situations: children that experience parental divorce are those whose parents were together (married or cohabiting) until they were at least 9 months old, but who divorced when they were aged between 9 months and 5 years (N = 1177); children that experience parental temporary separation are those whose parents were together (married or cohabiting) when they were born and when they were 9 months and 5 years old (N = 277); however, on one or more occasions, their parents spent more than one month living apart; children in stable families are those whose parents remained in stable married or cohabiting unions from their birth until age 5 (N = 9001).

Partnership quality was derived from the Golombok Rust Inventory of Marital State (GRIMS, Rust et al. 1990 ) which is a psychometric instrument for the assessment of marital discord and the overall quality of a couple’s relationship. We only used the GRIMS scale for responses from the mother, as the fathers’ questionnaire showed a high percentage of missing cases. We use this scale at the first wave (9 months) since it has seven items, as opposed to four items in the subsequent waves.

The following four items, the responses to which were collected at the first wave assess the negative aspects of relationship quality: (1) “my partner doesn’t seem to listen to me”; (2) “sometimes I feel lonely even when I am with my partner”; (3) “I wish there was more warmth and affection between us”; and (4) “I suspect we may be on the brink of separation”. The other three items assess the positive aspects of relationship quality: (1) “my partner is usually sensitive to and aware of my needs”; (2) “our relationship is full of joy and excitement”; and (3) “we can always make up quickly after an argument”. The item responses consist of the following: strongly agree (0); agree (1); neither agree nor disagree (2); disagree (3); strongly disagree (4) and can’t say (5). “Can’t say” responses were considered as missing information. To create an ordinal scale, we included both the positive and the negative items, which involved reversing the answers to the positive items. For these items the answers were: strongly disagree (0); disagree (1); neither agree nor disagree (2); agree (3); strongly agree (4). We then added up respondents’ answers to the seven items, which produced a scale with a minimum of 0 and a maximum of 28.

Most studies, also due to few observed divorced couples, consider the heterogeneity of divorce by accounting for an interaction between parental divorce and the continuous variable measuring parental relationship quality. They assume that the magnitude and sign of the interaction effect is the same across any value of the relationship quality and they do not allow the extent to which the effect of parental divorce diverges in according to the intensity of the relation to be examined. Only Fomby and Osborne ( 2010 ) use a binary variable to identify couples with a low relationship quality if the reported value is below the 25th percentile of the sample distribution. We consider the quartiles of the empirical distribution and we account for the following ordered categories of decreasing union quality: very good, good, poor and very poor. We choose this specification to obtain a more accurate portrayal of children who experience especially poor and very poor parental relationship quality.

Table  2 shows the descriptive statistics of the family situation according to parental relationship quality. In the sample 86% belong to stable family, 11% experience parental divorce and around 3% experience parental temporary separation. The percentages of those reporting different levels of relationship quality are quite similar: 28% and 20% reported high and low quality relationships, respectively. The data reveal that parents who remained together from wave 1 (children were 9 months old) to wave 3 (children were 5 years old) had better relationship quality on average than those who divorced or experienced some period of separation. Comparing the two types of family disruption, parents who subsequently divorce exhibit worse relationship quality than those who only temporarily separate. At wave 1, around 18% of parents in stable family reported the lowest relationship quality compared with 32% of those who later separated temporarily and 39% of those who later divorced. Hence, in accordance with the selection hypothesis, a large number of children with divorced parents were exposed to poor union quality before parental separation. However, contrary to this hypothesis, Table  2 also shows that a considerable proportion of parents who divorced had not experienced poor relationship quality prior to ending their relationship. Among children whose parents divorced, around 17% and 22% belonged to families with the highest (q 1 ) and high (q 2 ) relationship quality, respectively. It is important to acknowledge that children whose parents had the highest relationship quality at wave 1 could experience poor parental relationship quality after this wave and prior to their parents’ divorce since this event occurs between wave 1 and wave 3 of the survey.

Although the percentages show that a large proportion of divorced parents reported the lowest level of relationship quality before separation, the row percentages demonstrate that the majority of parents with poor-quality relationships do not separate. Approximately three-quarters (73%) of mothers with the lowest level of relationship quality at wave 1 remained in a relationship with the father of their child four years later.

Overall, considering only values from Tables  1 and 2 , we cannot say whether the observed differences on school readiness between children from different family situations are explained by differences in parental relationship quality pre-dating the experience of family disruption.

3.2.3 Control Variables

The control variables illustrated in Table 3 are measured when children were 9 months old (wave 1), namely before parental separation took place, and took into account several socio-demographic characteristics related to family disruption and children’s well-being (Booth and Amato 1991 , 2001 ; Wilson and Waddoups 2002 ; Amato and Hohmann-Marriott 2007 ; Kiernan and Huerta 2008 ; Brown 2004 ; Kiernan and Mensah 2009 ; Brooks-Gunn et al. 2010 ; Muluk et al. 2014 ; Idstad et al. 2015 ; Karraker and Latham 2015 ; Oláh and Gähler 2014 ; Sabates and Dex 2015 ).

Concerning the social exchange theory (Levinger 1976 ) we include into the rewards and costs the following variables to control for the selection into family disruption: family income, housing tenure, mother’s educational attainment and ethnicity, mother’s health (depression and longstanding illness) and the presence of half- or step-siblings at home. We consider the following variables as barriers to family disruption: paid work status of the mother; whether the mother lived with someone else as a couple before living with the father of the child; type of parental union (married directly, cohabitation before marriage, or cohabitation); year that parents began living together as a couple; whether parents grew up in a non-stable family and mother’s attitudes to single-parent upbringing. Instrumental support is measured by the following response provided by the mother: “If I had financial problems, I know my family would help if they could”. Finally, another group of control variables is related to the division of unpaid work, which is associated with the probability of divorce: who is mostly responsible for household tasks and who is generally with and looking after the children.

3.3 Methods and Analytical Strategy

The effect of parental divorce (or parental temporary separation) on children’s outcomes is evaluated under the framework of counterfactual reasoning or Potential Outcomes (POs) (Rubin 1974 , Holland and Rosenbaum 1986 ). In this context, we are interested to estimate the Average Treatment Effect (ATE) that is conceived as the difference between the expected values of the POs of the children in the treated and untreated condition. We refer to the POs of individual exposed to treatment as \( Y_{i}^{\left( 1 \right)} \) and not exposed as \( Y_{i}^{\left( 0 \right)} \) for each i , i  = 1,…, n . The treatment is provided only to a fraction of the units, and it is denoted by a binary variable Z i that is equal to 1 if the individual i is treated and to 0 if he/she is not treated. The treatment effect is defined as the difference \( Y_{i}^{\left( 1 \right)} - Y_{i}^{\left( 0 \right)} \) and the same value over the population is the ATE defined as the expected value of the POs as follows

The realized outcome for individual i is given by

In this framework, the outcomes of the children whose parents are divorced or temporary separated are only observed in the presence of the treatment conditions and the outcomes of the children in stable families are only observed in the absence of treatment. That is, a child can experience parental divorce or can live in a stable family from 9 months to age 5. As recently stated by Kim ( 2011 ) the ATE joints the realized developmental outcomes for children had experienced divorce or temporary separation and the counterfactual outcomes for these children had their parents remained together. The effects of family disruption on children’s outcomes can be assessed only on average in this non-experimental study, since each child belongs to one of the treatment or to the control group and one PO is always not realized. Children experiencing family disruption may be not randomly selected, and the family characteristics that determine the disruption may also affect the child’s well-being through other pathways (McLanahan et al. 2013 ).

The Propensity Score (PS, Rosenbaum and Rubin 1983 ) is a multivariate statistical matching method proposed for data collected in non-experimental contexts aimed to reduce the bias of the estimator of the treatment effect by considering the observed pre-treatment covariates (Rosenbaum 2020 ). The PS concerns the conditional probability of the treatment (the probability of experiencing parental divorce or parental temporary separation) given the observed pre-treatment covariates. This aims to mimic an experimental context especially when the observational data are rich as in the context of this application where similar questionnaires are administered to the participants. For estimating the ATE the weighted regression estimator (Rosenbaum 1987 ) defined as the Inverse Propensity Treatment Weighted (IPTW) estimator (Robins et al. 2000 ) weights each unit according to the estimated inverse probability of receiving the treatment actually received. The weights are obtained as the inverse of the estimated PS and in this context they allow us to compare children exposed to divorce despite their low probability of exposure and children not-exposed to divorce. We use observable pre-treatment covariates collected into the column vector \( \varvec{X}_{i} \) whose realized values are denoted with \( \varvec{x}_{i} \) for i  = 1, …,  n . We assume that conditional to the pre-treatment covariates the average outcomes in the treated and control groups in the absence of treatment would be the same. Another assumption is defined as strong ignorability (Rosenbaum and Rubin 1983 ) and postulates that given the pre-treatment covariates the treatment choice is independent of the POs. The positivity assumption is also required meaning that each treatment level occurs with some positive probability. This is also defined overlap assumption since it implies that the support of the conditional distribution of the covariates x i given Z i = 0 overlaps completely with the conditional distribution of x i given Z i = 1 (Imbens and Wooldridge 2009 ).

Disposing of a sample of n independent units, the IPTW estimator uses weights estimated through the maximum likelihood estimates of the parameters of the multiple logistic regression model given by

In this way it is possible to mimic a pseudo-population in which the covariates are balanced between treated and untreated individuals. The Augmented IPTW (AIPTW) estimator has the smallest asymptotic variance among the class of the IPTW estimators (Robins et al. 1994 ) and it is obtained according to the proposal of Lunceford and Davidian ( 2004 ) as follows:

where \( \hat{f}\left(Y_i= {y_{i} |\varvec{v}_{i} } \right) \) denotes the multiple linear regression model estimated for the observed responses by using ordinary least squares or robust inferential methods with \( \varvec{v}_{i} \) denoting the vector of the observed covariates. All the relevant covariates should be included in the sets \( \varvec{x}_{i} \) and \( \varvec{v}_{i} \) . We propose to apply the AIPTW estimator since it corrects for possible mis-specifications in the PS model or in POs model and it is statistically more robust with respect to other methods. The so-called “double robustness” property (Bang and Robins 2005 ; Neugebauer and van der Laan 2005 ) implies that the estimator remains consistent if the POs or PS model are incorrectly specified, see among others, Cao et al. ( 2009 ) and Glynn and Quinn ( 2010 ). When the estimated weights are too large Robins et al. ( 2000 ) propose to truncate such weights up to a specified threshold, preventing to some units being highly influential.

The analytical strategy we follow is to consider the three hypotheses illustrated in the introduction and to show the results according to the following steps. First, we estimate the ATEs of parental divorce and parental temporary separation including the available covariates in the PS model shown in Table  3 and accounting also for the overall weights to consider attrition and the initial sampling design. We estimate the POs model by considering all the variables most directly related to the children’s living conditions selected according to the knowledge in the field. They are the following: sex of the child; number of children at home; mother’s education, ethnicity and labor force participation; household income; housing tenure; mother’s longstanding illness and depression, and type of parental union. Second, to evaluate the selection hypothesis we estimate the ATE including the parental relationship quality in the POs and PS models. Third, to evaluate the heterogeneity for parental divorce we estimate the ATE by considering the quartiles of the variable relationship quality in the POs and PS models. The heterogeneity hypothesis is not considered for children experiencing parental temporary separation. The three steps are repeated for every outcome by considering each time five imputed datasets in order to account for missing values. More details are provided in Section A1 of the supplementary material.

4.1 Average Effects of Family Disruption on Children’s Well-being

In order to evaluate i ) the independent and ii ) the selection hypotheses, we compare the following two models: Model 1, which only includes control variables, and Model 2, which also considers parental relationship quality in both outcome and treatment models. Table  4 reports the results for each psychological and cognitive dimension. As expected, this covariate is significant in predicting the probability of parental divorce and parental temporary separation for each dimension. We refer to Section A2 of the supplementary material for some additional results for conduct problems where we show the estimated regression coefficients of the covariates included in the PO and PS models (Table A1 and A2, respectively). Concerning the results showed in Table  4 we notice that the estimated ATE of parental divorce is significant for all the psychological dimensions and except for the picture similarity test in Model 1 for all the other cognitive dimensions. However, when parental relationship quality is introduced among the control variables (Model 2), the effect of parental divorce is not significant for internalizing problems and peer problems. For conduct and hyperactivity problems, the magnitude of the effect of parental divorce is still significant but is considerably reduced. For conduct problems, parental divorce increases the average score of 0.244 points (Model 1) with respect to the score of children in stable family but this average score decreases to 0.162 when parental relationship quality is included (Model 2). For hyperactivity, the effect of parental divorce is 0.407 in Model 1 and 0.241 in Model 2. By considering parental relationship quality the ATE is reduced of around 34% for conduct problems and around 41% for hyperactivity problems Footnote 3 . Unexpectedly, the effect of parental divorce on pro-social behavior becomes significant in Model 2. For the cognitive dimension, the estimated effect of parental divorce in Model 1 is significant for all cognitive variables with the exception of the picture similarity test. Unlike the results for the most psychological variables, when parental relationship quality is included (Model 2), the effect of parental divorce does not decrease for the pattern construction test, and even increases slightly for the vocabulary test.

The effect of parental temporary separation is not significant in any model for internalizing, peer problems and pro-social behavior. In contrast, parental temporary separation has a significant negative effect on children’s hyperactivity and conduct in both models. For parental divorce, parental relationship quality does not reduce the effect of parental temporary separation in any of these psychological dimensions. It is also important to point out that for conduct and for hyperactivity problems, the magnitude of the effect of parental temporary separation is greater than the effect of parental divorce. The results of Model 2 show that the estimated PO mean for the conduct scores of children in stable family is 1.287. Parental temporary separation increases this score by an average of 0.384 while parental divorce increases it by an average of 0.162. In other words, the effect of parental temporary separation increases conduct problems by around 30% while parental divorce only increases conduct problems by around 16%. Similar differences result for hyperactivity problems. Turning to cognitive variables, the effect of parental temporary separation is not significant in any model for the pattern construction and vocabulary tests and this effect is only significant in Model 1 for the picture similarity test.

4.2 Heterogeneous Effects of Parental Divorce According to Parental Relationship Quality

The third hypothesis is evaluated according to the results showed in Table  5 reporting the ATE estimated according to the relationship quality. With regard to the psychological dimension, the effect of parental divorce on conduct problems is only significant for children that experienced extreme levels of parental relationship quality. Among children whose parents reported the highest relationship quality (q 1 ), the PO mean in stable family is 0.960, with parental divorce increasing it by 0.349. In other words, children with parental divorce experiencing good relationship quality show more conduct problems than children in stable family. The difference in percentage is lower among children whose parents had a very poor relationship quality (q 4 ).

For hyperactivity problems, the effect of parental divorce is significant only for those whose parents had a very poor relationship. Children with parental divorce experiencing very poor relationships among parents, have a higher probability of reporting hyperactivity problems compared to children with a stable family; the difference in percentage is around 11%.

As it can be seen, for internalizing problems the average effect of parental divorce is not significant once parental relationship quality is considered. However, when this effect is analyzed according to the quartiles of parental relationship quality, we get similar results to those obtained for conduct problems. The effect of parental divorce is significant in the extreme level of the relationship quality: within the group of children whose parents showed very good relationships, those who experience parental divorce have a higher probability of manifesting internalizing problems compared to children from stable family; the difference in percentage is around 20% Footnote 4 . Within the group of children whose parents had very poor relationships, the difference in percentage is lower, at around 12%. For peer problems, the effect of parental divorce is only significant for children whose parents had a good relationship (q 2 ) and, for pro-social behavior, the effect is only significant for those with bad relationships (q 3 ).

With regard to the cognitive dimension of children’s school readiness, although the average effect of parental divorce on the picture similarity tests is not significant Table  4 , the results are different in Table  5 . The effect of parental divorce is significant and equal to − 1.326 among children with parents reporting very poor relationships (q 4 ). The effect for those experiencing very good relationships (q 1 ) among parents is not significant. For the vocabulary test, it is interesting to note that the estimated ATE is significant and negatively large for those children exposed to a very good relationship (q 1 ) among parents. In this category, parental divorce decreases the score of the vocabulary test by an average of 3.319 points. It is also significant but lower in magnitude for those whose parents had a bad relationship (q 3 ). For the pattern construction test the estimated ATE is significant and negative for children experiencing poor and very poor relationships.

5 Discussion

This work is an attempt to elucidate the interrelationships between family disruption and parental relationship quality by testing the following three main hypotheses: i ) parental relationship quality and family disruption are unrelated processes that have independent effects on children (the independent hypothesis); ii ) the apparent effect of family disruption is explained according to the parental relationship quality (the selection hypothesis); iii ) the effect of family disruption on children depends on the quality of the parental relationship (the heterogeneity hypothesis).

We advance previous research in several ways. First, we evaluate the importance of these hypotheses using a comprehensive view of child development rather than focusing on a single outcome. We analyze multiple domains of children’s school readiness: cognitive and psychological well-being. Second, we focus on very young children who are at a key point in their development, namely the transition to school, while most research focuses on children in middle childhood or older. Third, we also analyze parental temporary separation which is a type of family disruption that is only scarcely covered in previous literature. Fourth, unlike most previous research, our study examines the heterogeneity of divorce effects by parental relationship quality outside of the US by using a UK nationally representative sample. Fifth, we use a proper multivariate method to impute the missing values and we employ the augmented inverse propensity treatment weighted estimator to infer the causal effects in each imputed dataset and we combine the results. Up to our knowledge this estimator has not been used by previous studies to assess the effect of parental divorce on children well-being.

We find mixed support for the i ) independent and the ii ) selection hypotheses, obtaining a different pattern for each outcome and type of family disruption. The selection hypothesis is supported by the potential outcome models regarding the average effect of parental divorce on pro-social behavior, internalizing and peer problems. Nevertheless, there is evidence in favor of the independent hypothesis in five of the eight outcomes.

Parental temporary separation only has a significant effect on conduct and hyperactivity problems; however, the magnitude of the effect of this type of family disruption is greater than the magnitude of the effect of divorce. These results indicate that, although children experiencing parental temporary separation have been invisible in most previous research and family policies, they are also at risk, and more research on this type of family disruption is needed (Nepomnyaschy and Teitler 2013 ; Halpern-Meekin and Turney 2016 ).

With regard to the third hypothesis related to the heterogeneity of divorce effects, this study shows that the average independent effects mask the substantial variation of the effect of parental divorce. First, we find that a non-negligible proportion of children from divorced families did not experience parental relationship problems. For this group of children, the idea that the negative effects of parental divorce are explained by parental relationship quality is not valid. In addition, our findings clearly support the hypothesis that the dissolution of high-quality parental unions has the most harmful effects on children’s lives. We find that among children whose parents had a very good relationship quality, there are substantial differences between those whose parents divorce and those that remain together in six of the eight analyzed dimensions. In four outcomes, the effect of parental divorce is greater for them with respect to the others.

Our findings for children of non-distressed families are in accordance with the existing literature on the heterogeneity of divorce effects based on children in middle childhood, adolescence or adulthood using US and Canadian data. However, it is important to point out that our results based on children at age 5 clearly diverge from those obtained by Fomby and Osborne ( 2010 ) with children at age 3, which find that parental divorce is not harmful for children in high- and low-conflict families. This discrepancy is probably due to the fact that these authors categorize three-quarters of the unions as high-quality, while we consider of very good quality only those unions above the 75th percentile (highest relationship quality) of the sample distribution. This result is consistent with research on the heterogeneity of divorce effects in adults which shows that divorce has the most harmful emotional effects among those who had satisfactory relationships prior to separation. Therefore, it is reasonable to assume that the children of these parents, especially the very young, are also affected more strongly.

We do not find any evidence that corroborates the hypothesis that the effect of parental divorce is positive for children who experienced poor parental relationship quality (see also Booth and Amato 2001 ; Hanson 1999 ). It is also important to acknowledge that we do not expect to obtain this finding in a country such as the UK where fewer children are living in poor families compared to the US (OECD 2017 ). Comparative research on the heterogeneity of the effects of parental divorce is needed in order to determine to what extent this hypothesis varies by country and the mechanisms that may explain this variation such as the generosity of family policies.

In addition to that, we explain why we find a divergence between our result and the one obtained by previous studies. First, as mentioned, these studies have used a continuous measure of parental relationship quality and they consider the interaction effect between parental divorce and parental relationship quality instead we evaluate to what extent the effect of divorce differs between very good, good, poor and very poor relationship quality. This allow us to detect a non-linear pattern in some of our outcomes: children whose parental relationship lay in the extremes (very good or very poor) are those most affected by divorce while children whose parental relationship was moderately good or bad are the least affected. Making a comparison with adult data, our results are in line with the findings of Williams ( 2003 ). Second, the proposed model specification based on the augmented inverse probability weighted estimator which requires to specify a potential outcome model along with the propensity score model, has never been proposed in this context. Third, most previous research has focused on parental conflict measured in terms of the frequency of disagreements, rather than on a measure of overall marital discord and quality, the millennium cohort study does not, however, provide a direct measure of the disagreement among parents. Also, our measure of parental relationship quality is derived from the mother’s perceptions and therefore misses the father perceptions. Due to these limitations, we are not able to capture the overall level of relationship quality that children experience at home. For this reason, we cannot rule out the possibility that if we had better measures to assess disharmonious families, we may have found positive effects of parental divorce for children living in them. Future research may be improved by using more subtle measurements of relationship quality and focusing on the heterogeneity effects of divorce on very young children.

Another reason that could explain why parental separation is not beneficial for children who experienced poor parental relationship quality is that we are not able to capture its duration. Research shows that there are different trajectories of parental relationship quality over time and that experiencing persistent poor parental relationship quality has more negative effects on children’s well-being than experiencing temporary poor parental relationship quality. Hence, it is reasonable to hypothesize that parental separation can be positive for children whose parents experienced a chronically poor parental relationship quality while the same event can be damaging for children that only experienced temporary poor parental relationship quality prior to separation. An important contribution of future research would be to include the measurement of the duration and trajectories of parental relationship quality in the studies of the heterogeneity effects of parental divorce. In addition to that, future studies should not only analyze the heterogeneity effects of parental divorce but also the heterogeneity effects of other forms of family transitions such as parental temporary separation. Children experience an increase in the number and types of family dissolution experiences during childhood: parents separate and then they may get back together or have a new partner and may separate again (Amato 2010 ).

Overall, the present study makes two main contributions to the literature by employing a suitable model to infer causality by reporting that parental temporary separation has detrimental effects for children, and that parental divorce exerts the most harmful effects among children whose parents enjoyed a very good parental relationship quality prior to separation. Our results can be used also to define policies targeted to those children for whom the detrimental effect of divorce might be stronger.

We are only able to empirically study the heterogeneous effects of parental divorce and not the heterogeneous effects of parental temporary separation due to data limitations. For this reason, we only theoretically discuss the heterogeneous effects of parental temporary separation in this section. In addition to that, the literature does not provide any theoretical explanation about the heterogeneity of the effects of parental temporary separation.

Details on the survey, its origins, objectives, and sampling, as well as the content of the survey waves, are contained in the documentation attached to the data deposited at the UK Data Archive at Essex University.

These percentages are calculated by considering the estimated ATE multiplied by 100 and divided by the estimated PO referred to stable family.

The percentages are computed as explained in the previous footnote.

Amato, P. R. (1993). Children’s adjustment to divorce: Theories, hypotheses, and empirical support. Journal of Marriage and Family, 55, 23–38.

Article   Google Scholar  

Amato, P. R. (2010). Research on divorce: Continuing trends and new developments. Journal of Marriage and Family, 72, 650–666.

Amato, P. R., & Hohmann-Marriott, B. (2007). A comparison of high-and low-distress marriages that end in divorce. Journal of Marriage and Family, 69, 621–638.

Amato, P. R., Loomis, L. S., & Booth, A. (1995). Parental divorce, marital conflict, and offspring well-being during early adulthood. Social Forces, 73, 895–915.

Bang, H., & Robins, J. M. (2005). Doubly robust estimation in missing data and causal inference models. Biometrics, 61, 962–973.

Booth, A., & Amato, P. R. (1991). Divorce and psychological stress. Journal of Health Social Behavior, 32, 396–407.

Booth, A., & Amato, P. R. (2001). Parental predivorce relations and offspring post divorce well-being. Journal of Marriage and Family, 63, 197–212.

Brooks-Gunn, J., Han, W. J., & Waldfogel, J. (2010). First-year maternal employment and child development in the first seven years. Monographs of the Society for Research in Child Development, 75, 7–9.

Google Scholar  

Brown, S. L. (2004). Family structure and child well‐being: The significance of parental cohabitation. Journal of Marriage and Family, 66, 351–367.

Bussell, D. A. (1996). A pilot study of African American children’s cognitive and emotional reactions to parental separation. Journal of Divorce and Remarriage, 24, 1–22.

Cao, W., Tsiatis, A. A., & Davidian, M. (2009). Improving efficiency and robustness of the doubly robust estimator for a population mean with incomplete data. Biometrika, 96, 723–734.

Elliott, C. D., Smith, P., & McCulloch, K. (1997). Technical manual British Ability Scales II . Berks: NFER-NELSON Publ. Co, Windsor.

Fitzgerald, K. A. (2010). Interparental conflict and emotional in security: coparenting and parent - child relationships as mediating family processes . Masters Thesis. Western Washington University.

Fomby, P., & Osborne, C. (2010). The influence of union instability and union quality on children’s aggressive behaviour. Social Science Research, 39, 912–924.

Gähler, M., & Palmtag, E. L. (2015). Parental divorce, psychological well-being and educational attainment: Changed experience, unchanged effect among Swedes born 1892–1991. Social Indicators Research, 123, 601–623.

Glynn, A. N., & Quinn, K. M. (2010). An introduction to the augmented inverse propensity weighted estimator. Political Analysis, 18, 36–56.

Goodman, R. (1997). The strengths and difficulties questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38, 581–586.

Graham, A. M., Fisher, P. A., & Pfeifer, J. H. (2013). What sleeping babies hear: A functional MRI Study of interparental conflict and infants’ emotion processing. Psychological Science, 24, 782–789.

Halpern-Meekin, S., & Turney, K. (2016). Relationship churning and parenting stress among mothers and fathers. Journal of Marriage and Family, 78, 715–729.

Hansen, K., & Joshi, H. (2007). Millennium Cohort Study second survey: A user’s guide to initial findings . London: Centre for Longitudinal Study, Institute of Education.

Hansen, K., et al. (2012). Millennium Cohort Study: A guide to datasets (seventh editions). First, second, third and fourth surveys . London: Centre for Longitudinal Study, Institute of Education.

Hanson, T. L. (1999). Does parental conflict explain why divorce is negatively associated with child welfare? Social Forces, 77, 1283–1316.

Hill, V. (2005). Through the past darkly: A review of the British ability scales second edition. Child and Adolescent Mental Health, 10, 87–98.

Holland, P. W., & Rosenbaum, P. R. (1986). Conditional association and unidimensionality in monotone latent variable models. Annals of Statistics, 14, 1523–1543.

Idstad, M., Torvik, F. A., Borren, I., Rognmo, K., Røysamb, E., & Tambs, K. (2015). Mental distress predicts divorce over 16 years: The HUNT study. BMC Public Health, 15, 1.

Imbens, G. W., & Wooldridge, J. M. (2009). Recent developments in the econometrics of program evaluation. Journal of Economic Literature, 47, 5–86.

Jekielek, S. M. (1998). Parental conflict, marital disruption and children’s emotional well-being. Social Forces, 76, 905–936.

Kalmijn, M. (2015). How childhood circumstances moderate the long-term impact of divorce on father–child relationships. Journal of Marriage and Family, 77, 921–938.

Kalmijn, M., & Monden, C. W. (2006). Are the negative effects of divorce on well-being dependent on marital quality? Journal of Marriage and Family, 68, 1197–1213.

Karraker, A., & Latham, K. (2015). In sickness and in health? Physical illness as a risk factor for marital dissolution in later life. Journal Health and Social Behavior, 56, 420–435.

Ketende, S. (2010). MCS technical report on response . London: Centre for Longitudinal Studies.

Kiernan, K. E., & Huerta, M. C. (2008). Economic deprivation, maternal depression, parenting and children’s cognitive and emotional development in early childhood. Bristish Journal Sociology, 59, 783–806.

Kiernan, K., McLanahan, S., Holmes, J., & Wright, M. (2011). Fragile families in the US and UK. Centre for Research on Child Well-Being. Working Paper WP 11 - 04 - FF , Princeton University, Princeton, NJ.

Kiernan, K. E., & Mensah, F. K. (2009). Poverty, maternal depression, family status and children’s cognitive and behavioural development in early childhood: A longitudinal study. Journal of Social Policy, 38, 569–588.

Kim, H. S. (2011). Consequences of parental divorce for child development. American Sociological Review, 76, 487–511.

Kim, L. S., Sandler, I. N., & Tein, J. Y. (1997). Locus of control as a stress moderator and mediator in children of divorce. Journal of Abnormal Child Psychology, 25, 145–155.

Levinger, G. (1976). A social psychological perspective on marital dissolution. Journal of Social Issues, 32, 21–47.

Lunceford, J. K., & Davidian, M. (2004). Stratification and weighting via the propensity score in estimation of causal treatment effects. Statistics in Medicine, 23, 2937–2960.

Maes, S. D., De Mol, J., & Buysse, A. (2012). Children’s experiences and meaning construction on parental divorce: A focus group study. Childhood, 19, 266–279.

McLanahan, S., Tach, L., & Schneider, D. (2013). The causal effects of father absence. Annual Review of Sociology, 39, 399–427.

Morrison, D. R., & Coiro, M. J. (1999). Parental conflict and marital disruption: Do children benefit when high-conflict marriages are dissolved? Journal of Marriage and Family, 61, 626–637.

Muluk, N. B., Bayoğlu, B., & Anlar, B. (2014). Language development and affecting factors in 3-to 6-year-old children. European Archives of Oto-Rhino-Laryngology, 271, 871–878.

Nepomnyaschy, L., & Teitler, J. (2013). Cyclical cohabitation among unmarried parents in fragile families. Journal of Marriage and Family, 75, 1248–1265.

Neugebauer, R., & van der Laan, M. (2005). Why prefer double robust estimators in causal inference? Journal of Statistical Planning and Inference, 129, 405–426.

OECD. (2017). Social policy division—directorate of employment, labour and social affairs. OECD Family Database. CO2.2: Child poverty, http://www.oecd.org/els/family/database.htm . Accessed 20 April 2020.

Oláh, L. S., & Gähler, M. (2014). Gender equality perceptions, division of paid and unpaid work, and partnership dissolution in Sweden. Social Forces, 93, 571–594.

Petterson, S. M., & Albers, A. B. (2001). Effects of poverty and maternal depression on early child development. Child Development, 72, 1794–1813.

Plewis, I. (2007). Non-response in a birth cohort study: the case of the Millennium Cohort Study. International Journal of Social Research Methodology, 10, 325–334.

Plewis, I., Calderwood, L., Hawkes, D., Hughes, G., & Joshi, H. (2000). Millennium Cohort Study: Technical Report on Sampling . London: Centre for Longitudinal Study, Institute of Education.

Raghunathan, T. E., Lepkowski, J. M., Van Hoewyk, J., & Solenberger, P. (2001). A multivariate technique for multiply imputing missing values using a sequence of regression models. Survey Methodology, 27, 85–96.

Robins, J. M., Hernan, M. A., & Brumback, B. (2000). Marginal structural models and causal inference in epidemiology. Epidemiology, 11, 550–560.

Robins, J. M., Rotnitzky, A., & Zhao, L. P. (1994). Estimation of regression coefficients when some regressors are not always observed. Journal of American Statistical Association, 89, 846–866.

Romano, E., Babchishin, L., Pagani, L. S., & Kohen, D. (2010). School readiness and later achievement: Replication and extension using a nationwide Canadian survey. Developmental Psychology, 46, 995–1007.

Rosenbaum, P. R. (1987). Model-based direct adjustment. Journal of the American Statistical Association, 82, 387–394.

Rosenbaum, P. R. (2020). Modern algorithms for matching in observational studies. Annual Review of Statistics and its Applications, 7, 143–176.

Rosenbaum, P. R., & Rubin, D. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70, 41–55.

Rubin, D. B. (1974). Estimating causal effects of treatments in randomized and nonrandomized studies. Journal of Educational Psychology, 66, 688–701.

Rubin, D. B. (1996). Multiple imputation after 18 + years. Journal of American Statistical Association, 91, 473–489.

Rubin, D. B. (2002). Introduction to multiple imputation . Stat. Anal. Missing Data, 2nd Ed NY, Wiley, pp. 85–93.

Rust, J., Bennun, I., Crowe, M., & Golombok, S. (1990). The GRIMS: A psychometric instrument for the assessment of marital discord. Journal of Family Therapy, 12, 45–57.

Sabates, R., & Dex, S. (2015). The impact of multiple risk factors on young children’s cognitive and behavioural development. Children and Society, 29, 95–108.

Strohschein, L. (2005). Parental divorce and child mental health trajectories. Journal of Marriage and Family, 67, 1286–1300.

Vennum, A., Lindstrom, R., Monk, J. K., & Adams, R. (2014). “It’s complicated” The continuity and correlates of cycling in cohabiting and marital relationships. Journal of Social and Personal Relationship, 31, 410–430.

Waite, L. J., Luo, Y., & Lewin, A. C. (2009). Marital happiness and marital stability: Consequences for psychological well-being. Social Science Research, 38, 201–212.

Wheaton, B. (1990). Life transitions, role histories, and mental health. American Sociological Review, 55, 209–223.

Williams, K. (2003). Has the future of marriage arrived? A contemporary examination of gender, marriage, and psychological well-being. Journal of Health Social Behavior, 44, 470.

Wilson, S. E., & Waddoups, S. L. (2002). Good marriages gone bad: Health mismatches as a cause of later-life marital dissolution. Population Research Policy Review, 21, 505–533.

Ye, M., DeMaris, A., & Longmore, M. A. (2017). Role of marital quality in explaining depressive symptoms after marital termination among older adults. Marriage Family Review, 1, 1–16.

Yu, T., Pettit, G. S., Lansford, J. E., Dodge, K. A., & Bates, J. E. (2010). The interactive effects of marital conflict and divorce on parent-adult children’s relationships. Journal of Marriage and Family, 72, 282–292.

Zhou, N., Cao, H., & Leerkes, E. M. (2017). Interparental conflict and infants’ behavior problems: The mediating role of maternal sensitivity. Journal of Family Psychology, 31, 464–474.

Download references

Acknowledgements

Open access funding provided by Università degli Studi di Milano - Bicocca within the CRUI-CARE Agreement. Garriga thanks the grants of the Spanish Ministry of Economy and Competitiveness (Grants CSO2012-33476 and CSO2015-69439-R). Pennoni thanks the grant “Finite mixture and latent variable models for causal inference and analysis of socio-economic data” (FIRB—Futuro in Ricerca) funded by the Italian Government (RBFR12SHVV). The authors thank Isabella Romeo for her contribution to preliminary data analyses.

Author information

Authors and affiliations.

Department of Political and Social Science, Pompeu Fabra University, Barcelona, Spain

Anna Garriga

Department of Statistics and Quantitative Methods, University of Milano-Bicocca, Milan, Italy

Fulvia Pennoni

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Fulvia Pennoni .

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Supplementary material 1 (PDF 140 kb)

Rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Garriga, A., Pennoni, F. The Causal Effects of Parental Divorce and Parental Temporary Separation on Children’s Cognitive Abilities and Psychological Well-being According to Parental Relationship Quality. Soc Indic Res 161 , 963–987 (2022). https://doi.org/10.1007/s11205-020-02428-2

Download citation

Accepted : 06 July 2020

Published : 27 July 2020

Issue Date : June 2022

DOI : https://doi.org/10.1007/s11205-020-02428-2

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Augmented inverse propensity treatment weighted estimator
• Conduct problems
• Family instability
• Missing values
• Find a journal
• Publish with us
• Track your research

Wealth and Divorce

Corresponding author: [email protected]

• Standard View
• Article contents
• Figures & tables
• Supplementary Data
• Peer Review
• Open the PDF for in another window
• Permissions
• Cite Icon Cite
• Search Site

Alexandra Killewald , Angela Lee , Paula England; Wealth and Divorce. Demography 1 February 2023; 60 (1): 147–171. doi: https://doi.org/10.1215/00703370-10413021

Download citation file:

• Reference Manager

In the United States, wealthier couples have lower divorce risk. Wealth may stabilize marriage through its material value, especially by easing financial stress, or by providing symbolic resources, especially signaling that couples meet normative financial standards for marriage. We first show that the negative association between wealth and divorce holds net of a rich set of controls. All else being equal, having $40,000 in wealth rather than $0 is associated with as big a decline in average predicted divorce risk as having no nonmarital births versus at least one. Second, we show that the negative association between wealth and divorce risk is steepest at low positive wealth levels. Net of covariates, having $40,000 in wealth rather than $0 is associated with as big a decline in average predicted divorce risk as having $400,000 rather than $40,000. Third, we consider evidence for the symbolic perspective, which emphasizes the stabilizing role of owning visible physical assets, and the material perspective, which suggests unsecured debt heightens divorce risk. Consistent with the symbolic perspective, we find that with net worth held constant, ownership of homes and vehicles is negatively associated with divorce risk. However, more research is needed to fully adjudicate between the symbolic and material perspectives.

• Introduction

Wealth inequality in the United States is high and rising ( Pfeffer and Schoeni 2016 ). Variation in marital stability amplifies wealth inequality: wealthier couples are more likely to stay married ( Eads and Tach 2016 ; Eads et al. forthcoming ), and remaining married is associated with higher wealth ( Addo and Lichter 2013 ; Yamokoski and Keister 2006 ; Zagorsky 2005 ). If wealth stabilizes marriages, rising wealth inequality may also further stratify American family life.

Wealth may increase marital stability through its material value, providing couples with economic resources they can use to improve their marital quality, thereby lowering their divorce risk. Alternatively, wealth may stabilize marriages through its symbolic value, allowing couples who have achieved the economic success expected of married couples to receive more social support or perceive greater value in their marriage. Or the association between wealth and divorce may not reflect a causal relationship.

In this article, we describe the association between wealth and divorce risk for Americans born in the late 1950s and early 1960s and explore the underlying mechanisms linking wealth and marital stability. We first consider that the association between wealth and marital stability may be spurious rather than causal. Although we cannot estimate the causal effect of wealth on divorce, we show that the association between wealth and marital stability remains net of a richer set of control variables than considered in prior research. Thus, we show that wealth is a distinct predictor of divorce risk, above and beyond other measures of socioeconomic resources, such as income and education.

Next, we seek to uncover the sociological processes that lead to greater marital stability for wealthier couples. Again, we cannot draw firm causal conclusions, but we evaluate whether empirical patterns are consistent with the predictions of various theoretical perspectives. We consider that wealth's material benefits may largely derive from avoiding acute financial stress and that its symbolic benefits may largely derive from clearing a wealth “marriage bar” that defines a minimum appropriate level of affluence for married couples. Both possibilities suggest the wealth–divorce association will be more negative toward the bottom of the net worth distribution than at the top. Consistent with these predictions, we find the largest differences in divorce risk across modest values of positive net worth, not between the highest wealth levels and more moderate wealth.

Finally, we continue to investigate the possible underlying mechanisms linking wealth and marital stability by seeking to determine whether the negative association between net worth and divorce risk is due to wealth's material or symbolic value. Because each perspective foregrounds particular assets or debts as especially important for marital stability, we examine how portfolio composition is associated with divorce, conditional on total wealth. Unsecured debt—not backed by collateral—may increase financial stress. Home and vehicle ownership, in addition to their material value, are visible symbolic markers of middle-class status. We find that, holding net worth constant, ownership of homes and vehicles is negatively associated with divorce risk, consistent with the symbolic perspective. However, further research is needed to fully evaluate and distinguish between the material and symbolic perspectives.

• Wealth's Material Value in Reducing Divorce Risk

In this section, we describe how wealth's material value may shape couples' risk of divorce, and we denote the associated hypotheses with “M” for “material.” In the next section, we consider the possibility that wealth carries symbolic value that reduces divorce risk, and we denote the associated hypotheses with “S” for “symbolic.”

Material economic resources may improve couples' relationship quality by allowing them to face fewer contentious decisions about spending, outsource household labor, experience less crowded living quarters, and engage in valued leisure activities. By contrast, limited economic resources are associated with increased economic strain, which is associated with poorer marital quality ( Conger et al. 1990 ). Eads and Tach (2016) theorized that assets may stabilize relationships by buffering negative economic shocks and reducing material hardship. Consistent with this idea, consumer debt is positively associated with economic pressure and negatively associated with marital quality, whereas assets tend to have the opposite associations ( Dew 2007 , 2009 , 2011 ). Thus, the material perspective on wealth suggests wealthier couples are less likely to divorce because they can use their financial resources to purchase valued goods and services and achieve greater economic stability in ways that improve their relationship.

Hypothesis M1 : Net worth is negatively associated with divorce risk, all else being equal.

One material benefit of net worth is the reduction of financial stress. The marginal returns to net worth for reducing financial stress are expected to be highest at low net worth levels and to decline as net worth rises. Put differently, avoiding net debt and having some asset safety net—being comfortable versus precarious—should reduce financial stress more than being very wealthy versus moderately wealthy. If the material benefits of wealth reduce divorce risk by reducing financial stress, higher net worth should reduce divorce risk most among those experiencing acute financial stress: those with low net worth.

Hypothesis M2 : Net worth is more negatively associated with divorce risk at low values of net worth than at higher values, all else being equal.

Finally, the material perspective suggests that different types of assets and debts may have different consequences for marital stability. Imagine that two couples have the same total net worth, but the first couple holds $10,000 in unsecured debt, whereas the second couple has no unsecured debt and $10,000 less in home equity than the first couple. In theory, assets and debts are fungible. However, in practice, transaction costs may prevent converting home equity to cash to pay off debt, and the first couple may be strained by monthly bills to repay the loans. Thus, the couple holding unsecured debt is expected to experience greater financial stress and marital instability. Having a small amount of unsecured debt may not be disruptive because it may permit couples to cushion short-term costs, such as by charging car repair costs on a credit card, but more unsecured debt is expected to be associated with greater marital instability.

Hypothesis M3 : The amount of unsecured debt is positively associated with divorce risk, all else being equal, including total net worth.

• Wealth's Symbolic Value in Reducing Divorce Risk

A second perspective predicts that wealth promotes marital stability through the symbolic cultural resources it provides rather than its financial value. Under this perspective, prior scholarship has theorized that asset accumulation may be one way that spouses perform and meet expected marital roles ( Dew 2007 , 2009 ) and that wealth's symbolic benefits may improve couples' interactions and commitment ( Eads and Tach 2016 ). Thus, the symbolic perspective, too, predicts a negative association between net worth and divorce risk.

Hypothesis S1 : Net worth is negatively associated with divorce risk, all else being equal.

Marriage has become a capstone event, to be entered into only after achieving other life course milestones ( Cherlin 2004 ). Unmarried couples describe delaying marriage until they have achieved what they perceive as a sufficient economic standard—clearing an economic bar—which may include owning a car and home, paying off debts, or saving enough to pay for a wedding ( Cherlin 2004 ; Edin and Kefalas 2005 ; Gibson-Davis et al. 2005 ; Smock et al. 2005 ).

Once married, couples whose marriages do not meet these normative standards may experience marital dissatisfaction and higher divorce risk. This logic suggests that wealth's symbolic value for marriage primarily distinguishes between couples who have and have not achieved a minimum level of affluence deemed appropriate for married couples. A couple who once met the marriage bar may divorce because they no longer meet the bar. A couple who never met the marriage bar may divorce because other aspects of their marriage become less rewarding and no longer offset the costs of not meeting the bar.

Hypothesis S2 : Net worth is more negatively associated with divorce risk at low values of net worth than at higher values, all else being equal.

So far, the predictions of the material and symbolic perspectives on wealth are identical: both predict a negative relationship between net worth and divorce risk, and versions of each perspective suggest the steepest association at the bottom of the wealth distribution. Therefore, in what follows, we drop the “M” and “S” designators when referring to Hypotheses 1 and 2.

However, the two perspectives highlight different wealth portfolio components as particularly relevant for marital stability. The material perspective suggests that unsecured debt will increase financial stress and, therefore, marital instability, even conditional on total net worth (Hypothesis M3). By contrast, the symbolic perspective emphasizes the importance of visible assets in demonstrating that the couple has cleared the marriage bar. Visible assets, such as homes and cars, feature in unmarried couples' descriptions of desired economic thresholds to clear before marriage ( Edin and Kefalas 2005 ; Gibson-Davis et al. 2005 ; Smock et al. 2005 ). The same logic may apply to decisions to remain married: ownership of visible assets demonstrates to the couple and others that they meet the marriage bar. Thus, the symbolic perspective predicts greater marital stability for owners of homes and vehicles, even conditional on total net worth. Put differently, wealth in the form of homeownership or vehicle ownership is likely to confer more stabilizing symbolic resources than holding the same amount of wealth in a less visible asset, such as a retirement account.

Homes and vehicles are not the only components of wealth with symbolic value; as mentioned earlier, couples considering marriage also attach meaning to paying down debts and saving for a wedding. However, homes and vehicles are two visible, symbolic, and easily measured aspects of net worth.

Hypothesis S3 : Ownership of homes and vehicles is negatively associated with divorce risk, all else being equal, including total net worth.

We caution that a lack of support for Hypothesis 2, Hypothesis M3, or Hypothesis S3 would not rule out the material or symbolic perspectives. Wealth's material and symbolic benefits for marital stability may operate through processes other than reducing acute financial stress or clearing a normative marriage bar (Hypothesis 2). Even if these processes are at work, our analyses may not capture them: unsecured debt may not heighten financial stress more than other components of net worth (Hypothesis M3), and homes and vehicles may not carry more symbolic value than other assets (Hypothesis S3). Thus, our analyses probing the potential underlying mechanisms linking wealth and divorce risk are exploratory rather than definitive.

• Prior Research and Remaining Questions on the Wealth–Divorce Association

Is Wealth Associated With Lower Divorce Risk?

Using Survey of Income and Program Participation (SIPP) data, Eads and Tach (2016) found that, net of controls, a one-standard-deviation increase in net worth was associated with a 31% decline in the risk of union dissolution. 1 Their model included a parsimonious set of controls: family income; both partners' employment statuses; the household reference person's race/ethnicity, age, and educational attainment; whether there are children in the household; and whether the union is a marriage or cohabitation.

In earlier research, Galligan and Bahr (1978) and Ross and Sawhill (1975) also found a negative association between wealth and marital disruption. More recently, Dew (2009 , 2011 ) found that marital dissolution risk rose with consumer debt and fell with assets, whereas Sanchez and Gager (2000) found that divorce risk was not significantly associated with capital assets, capital debt, or consumer debt. Because Eads and Tach (2016) provided the most comprehensive evaluation of the association between wealth and marital stability, considering both net worth and asset and debt components and seeking to distinguish between the symbolic and material perspectives on wealth, we focus on how our analyses differ from theirs.

Our first contribution is to include a richer set of control variables to provide a more rigorous test that the wealth–divorce association is robust to adjusting for confounding factors (Hypothesis 1). We use National Longitudinal Survey of Youth 1979 cohort (NLSY79) data. Relative to the SIPP, the NLSY79 has the key advantage of including detailed measures of respondents' social origins and family demographic characteristics. Although some controls we add may be endogenous to wealth, both the material and symbolic perspectives predict a direct effect of current net worth on marital stability, above and beyond how prior net worth may have shaped characteristics such as marriage timing or fertility. We reiterate that we cannot estimate the causal effect of wealth on divorce, but our inclusion of additional control variables provides stronger evidence that the association between wealth and divorce is not entirely spurious.

Is the Wealth–Divorce Association Due to Wealth's Material Benefits?

Like us, Eads and Tach (2016) considered financial stress to be one possible manifestation of the material perspective. They found that being in the fourth quartile of unsecured debt holdings was associated with a statistically significantly higher risk of union dissolution than being in the first quartile. Furthermore, they showed that material hardship, measured as reported difficulty meeting essential household expenses, explained approximately 30% of this disparity. The NLSY79 does not include measures of respondent-assessed material hardship, so we cannot replicate that portion of their analysis.

However, we extend Eads and Tach's (2016) analysis of unsecured debt by conditioning on total net worth (Hypothesis M3). This allows us to test whether there is anything distinctive about unsecured debt's relationship to marital stability, as opposed to merely increasing divorce risk as any other decrease in total net worth would. For example, recalling our two hypothetical couples from above, finding that the two couples, who have the same total net worth, have the same divorce risk, despite their different portfolio compositions and differences in unsecured debt, would not provide equally strong evidence for the financial stress perspective as showing that differences in unsecured debt are related to divorce risk with total wealth held constant. 2

Although Eads and Tach (2016) estimated a model in which they specified secured debt, unsecured debt, liquid assets, and illiquid assets with quartiles, they did not formally test whether the associations were consistent with linearity. Thus, our tests of the shape of the association between net worth and divorce (Hypothesis 2) are a new way to evaluate whether empirical patterns are consistent with the expectations of financial stress (and the economic marriage bar). Furthermore, describing the shape of the wealth–divorce association sheds empirical light on the economic stratification of family life, clarifying which wealth positions are associated with particularly distinctive divorce risks.

Is the Wealth–Divorce Association Due to Wealth's Symbolic Benefits?

Eads and Tach (2016) considered that wealth's symbolic benefits may result from clearing a marriage bar. However, as noted earlier, they did not test the implication that the wealth–divorce association is more negative at lower wealth levels, so our analyses testing for this nonlinearity are new (Hypothesis 2).

Like us, Eads and Tach (2016) argued that the symbolic perspective suggests that asset ownership affects union stability. They found that binary indicators for having any secured debt, any liquid assets, and any illiquid assets were each negatively associated with the risk of union dissolution, whereas having any unsecured debt had the opposite association. We argue that the interpretation of this pattern is not straightforward. Those who own a given asset, unless they owe more on the asset than it is worth, hold more value in the asset than nonowners; thus, an association between binary measures of ownership and marital stability does not reveal whether ownership has a larger role in marital stability than increases in asset value conditional on ownership. In our models seeking to evaluate whether asset ownership has symbolic value that reduces divorce risk (Hypothesis S3), we include both indicators of asset ownership and measures of the value of the asset, distinguishing the role of ownership from that of financial value. Furthermore, we control for net worth, isolating the predictive power of portfolio composition net of total wealth. Controlling for net worth allows us to assess whether asset ownership is associated with divorce risk separate from the fact that owners of an asset tend to be wealthier overall.

Our analyses also highlight that not all assets and debts have equal symbolic value. We focus on two examples of visible assets identified as symbolically meaningful in prior research: vehicles and homes ( Edin and Kefalas 2005 ; Gibson-Davis et al. 2005 ; Townsend 2002 ). Prior research has found that homeownership is negatively associated with divorce ( Cooke 2006 ; Ono 1998 ; South 2001 ) but has not evaluated whether this association merely reflects the overall negative association between net worth and divorce risk.

Last, Eads and Tach (2016) found homeownership to be associated with a lower risk of union dissolution only when partners jointly own the home, consistent with the possibility that joint investments facilitate relationship stability by symbolizing or encouraging commitment. In subsequent research, Eads et al. ( forthcoming ) found that couples who held a greater share of their assets and debts jointly had lower divorce risk. Because NLSY79 measures wealth only at the couple level, we cannot replicate this analysis.

• Data and Methods

Data Source: NLSY79

We used data from the 1979–2018 waves of the NLSY79 ( U.S. Bureau of Labor Statistics 2022 ). The NLSY79 first surveyed a nationally representative sample of young adults aged 14–22 in 1979 and, except for some discontinued subsamples, attempted to interview them annually until 1994 and then biennially.

The NLSY79 collected information on net worth in all waves between 1985 and 2000 except 1991, and subsequently in 2004, 2008, 2012, and 2016. We excluded observations before the first collection of wealth information in 1985. We included observations from 1991, 2002, 2006, 2010, and 2014, even though net worth was not collected in these years; as described later, we multiply imputed wealth in those years.

Defining the Risk Set

At each survey wave, we identified whether each respondent was married and therefore at risk of divorce. The NLSY79 asked respondents to report their marital status at each survey wave and the month and year of changes in marital status. Using this information, NLSY79 created variables indicating the start and end dates of respondents' marriages, with the end of a marriage defined by divorce or widowhood. We created revised marriage dissolution dates, treating a marriage as dissolved once either the reported marriage end date occurred or the respondent reported separated as their marital status and was never observed reunited with their spouse at a subsequent interview. Thus, although we refer to “divorce” for brevity throughout, we considered marriages to have ended as soon as permanent separation began, regardless of whether or when divorce occurred. We defined a respondent as currently married if the interview month was after the start date of one of their marriages and before our constructed month of the marriage's dissolution. When necessary, such as when the interview month and the month of the marital transition were the same, we broke ties using the respondent's current marital status and changes in marital status since the last interview. Once a marriage had ended, the respondent exited the risk set until they married again (if ever), at which point they reentered the risk set and analytic sample. 3

Outcome Variable: Divorce

Our outcome variable is whether a respondent in the risk set divorced or permanently separated before the next survey wave in which the respondent was observed. For respondents whose marriages ended, we used wave-specific reports of marital status and interwave changes in marital status to determine whether the marriage ended in widowhood rather than separation or divorce.

Core Predictor Variables: Net Worth, Assets, and Debts

The NLSY79 collected data on whether respondents and their spouses held a variety of assets and debts and, for those they held, the value. Questions about assets and debts differed across waves but included assets such as homes; vehicles; valuable items or collections; farms, businesses, and real estate other than residential homes; financial assets, including bank accounts, investment accounts, and retirement accounts; and other debt, such as to stores, hospitals, and banks, which we refer to as unsecured debt .

The NLSY79 computed couples' total net worth, which is the sum of the values of all assets the respondent and their spouse held, less the values of all debts. We converted all financial variables to constant 2020 dollars using the Consumer Price Index and then top- and bottom-coded at the 95th and 5th percentiles of the unweighted distribution in the analytic sample before multiple imputation.

We specified net worth as a linear spline with knots at the quartiles of the pooled, weighted distribution of married couple-years in the sample: $29,937; $108,813; and $319,011. We included an additional knot at $0 because net debt may be a distinctive state compared with low positive net worth.

In the analyses examining portfolio composition, testing Hypotheses M3 and S3, we also included three indicators for whether the couple had any unsecured debt, owned a vehicle, or owned a home, as well as three linear terms for home value, vehicle value, and unsecured debt value. In a second model, we included measures of home and vehicle equity (values less debts) rather than their values.

Control Variables

We controlled for marital duration with linear and quadratic terms for the number of years between the interview year and the year the respondent's marriage began.

We used the NLSY79 household screener to categorize respondents as Hispanic, non-Hispanic Black, or non-Black and non-Hispanic. We measured nativity with an indicator for whether the respondent was born in the United States.

We controlled for the respondent's parents' highest grade completed using four categories: less than 12th grade, 12th grade, one to three years of college, or four or more years of college. We also included an indicator for whether the respondent lived with two biological parents at age 14.

We measured each spouse's current educational attainment using the same categories as for parents' education, and we included an indicator variable for whether the respondent was currently enrolled in school. For each spouse, we measured whether they were employed full-time (at least 1,500 hours) in the prior calendar year by multiplying reported weeks and hours per week worked. We controlled for family income in the prior calendar year with a linear spline with knots at the weighted quartiles ($58,894; $88,247; and $127,706).

We controlled for respondents' age at marriage using three categories: younger than 21, 21–24, and older than 24. We measured prior nonmarital cohabitation with an indicator set to 1 if the respondent ever previously reported (1) a partner on the household roster while not married, (2) having cohabited before marriage with the most recent spouse (asked in 1990–2000), or (3) nonmarital cohabitation between survey waves (asked beginning in 2002). We measured nonmarital fertility with an indicator set to 1 if the respondent's reported birthdates of any of their children born to date fell outside the respondent's marriage spells as defined in our construction of the risk set. Using three categories, we controlled for whether the current marriage was the respondent's first, second, or third or higher. We controlled for the presence of children in the household with counts of the respondent's biological children and stepchildren in the home, each divided into those under 5 and those aged 5–17 and each top-coded at four. 4

We controlled for region with four categories: Northeast, North Central, South, or West. We included an indicator for whether the respondent lived in an urban versus rural area.

Analytic Plan

We estimated discrete-time hazard models with a logit link, modeling the log odds of a married respondent divorcing before the next survey wave in which they were observed and adjusting for the length of exposure—the difference between the wave in which the predictors were measured and the wave in which the divorce outcome was measured. Our baseline model describes the association between net worth and the hazard of divorce, controlling only for marital duration. This model describes how divorce risk is stratified by wealth position.

Next, we estimated the full model, which includes the control variables described previously. We tested Hypothesis 1 by evaluating whether the slopes on the net worth spline pieces were jointly statistically significantly different from zero. We tested Hypothesis 2 by evaluating whether the slopes on all the spline pieces were identical and therefore that the association was linear in log odds. We used a .05 significance level with two-tailed tests throughout. 5

To describe the magnitude of the association between net worth and divorce risk, we computed predictive margins, which give the average predicted probability of divorce in the next year for respondents if they were all assigned a particular value of net worth but otherwise had their own covariates. The predictive margins do not refer to the average predicted probability of divorce for those observed to have a particular net worth; they refer to the average predicted probability of divorce for the entire sample if their wealth were set to the given value. We annualized estimates by assigning each observation an exposure period of one year when generating predicted probabilities.

To test Hypotheses M3 and S3, we added to the full model indicators for homeownership, vehicle ownership, and holding any unsecured debt, plus linear measures of home value, vehicle value, and the value of unsecured debt. We then repeated this model but replaced the measures of vehicle and home values with measures of their equity. 6

To contextualize the magnitudes of the associations in these models, we again used predictive margins, this time examining how the predicted probability of divorce changed with different portfolio compositions while holding net worth constant.

We weighted all analyses using the NLSY79 year-specific weights and clustered standard errors at the 1979 household level.

Sample Restrictions and Missing Data

We censored individuals if they attrited before the next survey wave and censored all other respondents at the final wave in 2018. We censored marriage spells that ended in widowhood before the next survey wave. For 4% of the remaining observations, we could not determine whether they were in the risk set because their marriage start dates created by the NLSY79 were incomplete or inconsistent. We excluded these cases from the analytic sample, along with observations from currently unmarried respondents. 7 In the remaining sample of married couples, we excluded 2% of the observations because the exposure period—the number of years between the survey wave in which the predictors were measured and the survey wave in which the respondent was next observed—exceeded two years.

Only 22 same-sex couples met the criteria for inclusion in our sample. We restricted our analysis to different-sex couples because the correlates of union dissolution may differ for different-sex and same-sex couples (e.g., Weisshaar 2014 ).

Our analytic sample includes 88,660 couple-year observations from 8,351 respondents and 10,286 marriages. The sample includes 4,161 divorces.

We used multiple imputation with 20 imputations for item-level missing data. Because the NLSY79 did not collect wealth information in all waves, the highest missing data rates are for net worth (36%) and the home, vehicle, and unsecured debt components (23% to 30%). In waves in which the NLSY79 collected net worth data, the missing rate is 17% for net worth and no more than 9% for the home, vehicle, and unsecured debt components. Our imputation models include respondents' most recent report of the net worth and home, vehicle, and unsecured debt variables provided that it was within the last four years and while the respondent was in the same marriage. Including this most recent report allowed the couple's prior wealth to inform their imputed wealth in the years that the NLSY79 did not collect wealth data. The online supplement shows missing data rates and the results of models that treat missing data using listwise deletion and that exclude observations from survey waves in which wealth information was not collected.

As shown in Table 1 , which describes our analytic sample at the couple-wave level, 4% of the married respondents observed at a given wave divorce before the next survey wave in which they participated. Mean net worth after top- and bottom-coding is $226,597, and the median is $108,813. Net worth is less than zero for 7% of the sample. Further, 76% of the sample are homeowners, vehicle ownership is near universal (97%), and 57% hold unsecured debt. Among owners, median values are $191,938 for homes, $21,081 for vehicles, and $5,580 for unsecured debt.

The Robust Nonlinear Association Between Wealth and Divorce

Table 2 shows the logit coefficients for the net worth terms from the baseline and full models. 8 As shown in the bottom row of Table 2 , for both models, the net worth spline terms are jointly statistically significantly associated with divorce (i.e., we can reject the null hypothesis that the slopes on the wealth spline terms are all zero). In both models, net worth is negatively and statistically significantly associated with divorce for low positive values of net worth ($0–$29,937). For the middle 50% of the wealth distribution ($29,937–$319,011), net worth remains negatively associated with divorce, but the associations are statistically significant only in the baseline model. In both models, the association between wealth and divorce is not statistically significant for either the very bottom (below $0) or the top quartile (above $319,011) of the wealth distribution. Overall, the results support Hypothesis 1 that net worth is negatively associated with divorce risk, although perhaps not across the entire distribution. Given the rich set of controls in our full model, our results provide more rigorous support for the claim that wealth is a distinct predictor of divorce, although we still cannot draw firm causal conclusions. 9 , 10

In both models, the negative association between wealth and the log odds of divorce is most pronounced for positive values of net worth below the 25th percentile of the distribution (between $0 and $29,937) and then attenuates as net worth rises. As shown in the second-to-bottom row of Table 2 , for both models, we can reject the null hypothesis that the association between wealth and the log odds of divorce is linear (i.e., that the wealth spline coefficients are all equal). On the log odds scale, these results support Hypothesis 2, that the association between net worth and divorce is more negative at lower (positive) values of net worth. The association between net worth and divorce risk remains statistically significant and nonlinear when sibling fixed effects are added to the full model, although the pattern of coefficients across the net worth distribution is somewhat different (see the online supplement).

In both models in Table 2 , increases in net worth among net debtors (i.e., declines in net debt) are positively and not statistically significantly associated with divorce risk. Only 7% of the sample are net debtors ( Table 1 ), and the associations are imprecisely estimated. Therefore, we cannot draw firm conclusions about how, if at all, divorce risk changes across net debt values. However, divorce risk may not increase at higher net debt levels if access to credit allows couples to meet unexpected costs and smooth consumption.

Next, we calculated predictive margins, which allow us to assess whether the wealth–divorce association is nonlinear in predicted probabilities as well as log odds. To visualize stratification in divorce risk by net worth, Figure 1 shows the predictive margins and 95% confidence band for the baseline model across values of net worth between –$20,000 and $400,000 in increments of $20,000. 11 For context, the overall annualized hazard of divorce, generated from a hazard model with no covariates, is 2.6% (see the online supplement). 12 The average predicted probability of divorce in the next year is 5.1% when net worth is set to $0 but falls by approximately 25%, to 3.8%, when net worth is just $20,000 higher. Divorce risk continues to fall across moderate levels of net worth but is approximately constant at 1.8% when wealth is at least $300,000. These patterns highlight the substantial differences in marital stability between couples with low net worth compared with at least moderate wealth and smaller differences across the top of the wealth distribution.

Figure 2 spotlights these differences, showing the differences in predictive margins between $0 net worth—when the average predicted probability of divorce is highest—and $20,000 increments of net worth from $20,000 to $400,000 for both the baseline model and the full model. These differences are similar to average marginal effects (“effect” does not imply causality here) of net worth on divorce, except that the “margin” is a change in net worth from $0 to a specified alternative value rather than an instantaneous rate of change. As shown in Figure 1 , the predictive margin for $20,000 in net debt has a very wide confidence interval. We therefore do not show it in Figure 2 .

In the full model, compared with having $0 net worth, the average predicted probability of divorce in the next year is 0.4 percentage points lower when net worth is $20,000, 0.7 percentage points lower when it is $40,000, and 1.4 percentage points lower when it is $400,000. These disparities are 56% to 66% smaller than in the baseline model. Although control variables substantially reduce the wealth–divorce association, net worth remains meaningfully negatively associated with divorce risk net of controls, supporting Hypothesis 1. 13

The nonlinearity in the wealth–divorce association also remains net of controls, supporting Hypothesis 2. All else constant, the average decline in the predicted probability of divorce in the next year resulting from increasing net worth from $0 to $40,000 (0.7 percentage points) is the same as the average decline resulting from increasing net worth from $40,000 to $400,000.

To contextualize the magnitude of wealth's association with divorce, Table 3 shows the average marginal effects (for categorical variables) and average change in predictive margins across illustrative values (for quantitative variables) for all other predictors in the full model, analogous to the results for net worth in Figure 2 . The average decline in the predicted probability of divorce in the next year associated with having $400,000 in net worth rather than $0 (1.4 percentage points, from Figure 2 ) is similar to that of getting married after age 24 rather than before age 21 (1.5 percentage points) or of being in a first rather than a third marriage (1.4 percentage points). The average decrease in the predicted probability of divorce associated with having $40,000 in net worth rather than $0 (0.7 percentage points, from Figure 2 ) is similar to that of being in a first rather than second marriage (0.5 percentage points) or of having no nonmarital births rather than at least one (0.7 percentage points).

Although our focus is on net worth, family income also shows a negative association with divorce concentrated toward the bottom of the distribution. Having a family income of $60,000 rather than $20,000 is associated with an average decline of 1.1 percentage points in the predicted probability of divorce in the next year, but an income of $100,000 rather than $60,000 is associated with an average additional decline of only 0.4 percentage points.

Portfolio Composition and Assessing the Symbolic and Material Perspectives

Next, we assessed whether asset and debt types foregrounded by the material and symbolic perspectives are associated with divorce risk, holding constant total net worth. Table 4 shows the logit coefficients from these analyses. As shown in the first column, consistent with the symbolic perspective (Hypothesis S3), homeownership and vehicle ownership are negatively and statistically significantly associated with divorce holding net worth and the other asset/debt measures constant. The value of the home or vehicle is not significantly associated with divorce risk. 14

The results do not support Hypothesis M3: neither ownership of unsecured debt nor its amount is statistically significantly associated with divorce risk, nor are the terms jointly statistically significant.

We repeated the previous model but replaced home and vehicle values with equity (value less debts). If homes and vehicles reduce divorce risk purely through their symbolic value, we would not expect equity to predict divorce, conditional on total wealth. By contrast, according to the material perspective, equity may reduce financial stress because it can be borrowed against, or higher equity may indicate less burdensome debt payments. The results show that vehicle equity, in addition to home and vehicle ownership, is negatively and statistically significantly associated with divorce risk. This finding raises the possibility that vehicles have material as well as symbolic value in decreasing divorce risk.

To contextualize the magnitudes of the associations in Table 4 , we generated predicted probabilities of divorce in the next year for an illustrative scenario: $50,000 in net worth and varied portfolio composition. Both owners and nonowners of homes and unsecured debt are common at this level of net worth, making a range of portfolio compositions realistic. We compared the average predicted probability of divorce under three scenarios: when a couple does not own their home, when they own a home worth $75,000, and when they own a home worth $150,000. In each case, we set couples' net worth to $50,000 and otherwise used couples' observed covariates. This analysis allows us to compare the change in divorce risk associated with two equal-sized changes in home value, only one of which includes a change in ownership. We conducted analogous analyses for vehicle value (at $0; $10,000; and $20,000) and unsecured debt (at $0; $5,000; and $10,000). Because homes, vehicles, and unsecured debt have different ranges of plausible values, we considered a different range of values for each. Thus, the coefficients should be compared only within asset/debt types, not between them.

Panel A of Table 5 shows the results based on the first column of Table 4 , using home and vehicle values. The decline in the average predicted probability of divorce associated with owning a home worth $75,000 rather than not owning (0.7 percentage points) is statistically significantly different from the decline associated with owning a home worth $150,000 rather than $75,000 (less than 0.1 percentage points). The same pattern holds for vehicles: the decline in the average predicted probability of divorce associated with owning a vehicle worth $10,000 compared with not owning a vehicle (0.7 percentage points) is statistically significantly different from the decline associated with owning a vehicle worth $20,000 rather than $10,000 (less than 0.1 percentage points). By contrast, the increase in the average predicted probability of divorce associated with having $5,000 in unsecured debt versus $0 is not statistically significantly different from the increase associated with having $10,000 versus $5,000, nor is either of the individual effects statistically significant.

Panel B of Table 5 shows the results based on the second column of Table 4 , using home and vehicle equity. Because home equity values are often much lower than home values, here we estimated the change in the predicted probability of divorce associated with owning a home with $20,000 in equity versus not owning a home and with owning a home with $40,000 in equity versus $20,000 in equity. The results for homes and unsecured debt are similar to those in panel A. For vehicles, having $20,000 in equity rather than $10,000 is associated with a statistically significant decline in the predicted probability of divorce, just as vehicle equity was a significant predictor of divorce in the logit model in Table 4 . However, ownership remains more important: the decline in the average predicted probability of divorce associated with holding $10,000 in vehicle equity compared with not owning a vehicle (0.8 percentage points) is statistically significantly larger than the decline associated with holding $20,000 rather than $10,000 in vehicle equity (0.1 percentage points).

These results confirm that for homes and vehicles, asset ownership is associated with larger declines in divorce risk than equivalent changes in asset value or equity conditional on ownership, consistent with Hypothesis S3. The results in Tables 4 and 5 do not provide evidence for Hypothesis M3 that unsecured debt increases divorce risk, holding net worth constant. Of course, this does not mean unsecured debt is unimportant; it merely means that we cannot rule out the possibility that its association with divorce risk is well captured by its contribution to overall net worth. 15

Divorce risk varies substantially by wealth, especially across modest positive values of net worth. When we adjusted only for marital duration, the annualized average predicted probability of divorce fell from 5.1% at $0 net worth to 3.8% at $20,000 and to 1.8% at $300,000 or above. Descriptively, these patterns illuminate how marital stability is patterned by social class.

We sought to distinguish among three alternative explanations for the association between wealth and marital stability: (1) material benefits of wealth that increase marital stability, including by reducing financial stress; (2) symbolic benefits of wealth that stabilize marriage through the signal it provides of status and success, including that the couple exceeds a minimum economic bar considered appropriate for married couples; and (3) confounding factors that lead to a spurious association between wealth and divorce risk.

Drawing on the material and symbolic perspectives on wealth's benefits for marital stability, we articulated four empirical predictions. First, both perspectives suggest a negative association between wealth and divorce risk, net of controls (Hypothesis 1). After controlling for a richer set of variables than Eads and Tach (2016) , we confirmed their finding that, all else being equal, higher net worth is associated with greater marital stability. Net of controls, we found that having $40,000 in net worth rather than $0 is associated with, on average, a decrease of 0.7 percentage points in the annualized risk of divorce—comparable in magnitude to the decreases associated with being in a first rather than second marriage or having no prior nonmarital births versus at least one. Thus, as a correlate of divorce, net worth is on par with other correlates that have received far more attention.

Second, versions of the material and symbolic perspectives suggest that wealth should have diminishing marginal returns for marital stability. At the bottom of the wealth distribution, small increases in net worth may be valuable for reducing financial stress (a material benefit) or clearing a threshold for the minimum expected level of affluence for married couples (a symbolic benefit) (Hypothesis 2). Consistent with this prediction, we found that all else being equal, the average annualized predicted probability of divorce fell about as much when net worth was $40,000 rather than $0 as when it was $400,000 rather than $40,000.

Of course, we could not entirely rule out selection. Future research is needed to estimate the causal effect of wealth on divorce. This sort of examination might include, for example, exploiting exogenous wealth changes due to lottery winnings (e.g., for wealth's effect on labor supply, see Cesarini et al. 2017 ) or housing booms (e.g., for housing wealth's effect on offspring college outcomes, see Lovenheim and Reynolds 2013 ). Further, our analyses provide only a partial portrait of how wealth may affect marital stability because we controlled for characteristics that may be endogenous to prior wealth, such as income and fertility. We encourage future research that takes a broad array of approaches to collectively provide greater evidence on the effect of wealth on divorce.

Our final set of analyses sought to distinguish between the material and symbolic perspectives on wealth's benefits for marital stability. Here, we relied on the perspectives' different predictions about which asset or debt types are likely to be associated with divorce, above and beyond total net worth, and whether ownership or the dollar value is expected to be more consequential. We improved on prior research by isolating the predictive power of each asset or debt type net of overall wealth and not conflating ownership and value of the focal asset or debt. Consistent with the symbolic perspective, we found ownership of both homes and vehicles to be negatively and statistically significantly associated with divorce risk, holding constant total net worth and all other controls (Hypothesis S3). Using an illustrative case, we further showed that the same increase in asset value or equity is associated with greater declines in average divorce risk when it is paired with moving to ownership rather than an increase in value or equity conditional on ownership. This finding suggests that ownership of homes and vehicles is associated with greater marital stability in ways explained by neither their consequences for total net worth nor their implications for home or vehicle value or equity. Conditional on total net worth, we found that neither ownership of unsecured debt nor its amount is significantly associated with divorce risk, contrary to our prediction based on the material perspective (Hypothesis M3).

However, we urge caution in interpreting these results. The material and symbolic perspectives are broad, so versions of each may be consistent with various empirical patterns. Empirical estimation is challenging given the collinearity of different asset and debt types and limitations in which portfolio components can be disaggregated in the NLSY79. All else being equal, unsecured debt may be associated with divorce risk in the population, although the association is not statistically significant in our sample. Even if no such population-level association exists, this does not rule out other manifestations of the material perspective. Our analyses are exploratory rather than a definitive adjudication among spurious associations and the material and symbolic perspectives on wealth's association with marital stability.

Matters are even more complicated when we consider other possible mechanisms linking wealth and divorce risk. Wealth may affect divorce risk by shaping the expected costs and benefits of divorce as spouses divide assets ( Dew 2009 ). Prior research has considered how divorce laws, including unilateral divorce and the division of marital property, shape in-marriage behavior and divorce risk ( Clark 1999 ; Stevenson 2007 ; Voena 2015 ; Zang 2020 ). This line of research could be expanded to analyze how these laws moderate the association between wealth and divorce. We encourage future research that develops additional empirical predictions to distinguish among theoretical perspectives and future qualitative research that explores how couples understand the role of financial circumstances in their marital satisfaction and marital stability.

In an era of high and rising wealth inequality ( Pfeffer and Schoeni 2016 ), we found that wealthier couples experience lower divorce risk, net of a host of controls. We further found that reductions in divorce risk are concentrated over modest positive values of net worth and are much less pronounced above the median of the wealth distribution. Last, ownership of visible, symbolic assets—homes and vehicles—is negatively associated with divorce risk, even conditional on total net worth. Our findings support calls to consider wealth as a distinct indicator of economic stratification, and we encourage future research to recognize wealth as a substantial correlate of divorce.

• Acknowledgments

We are grateful for the helpful comments we received from three anonymous Demography reviewers. We also benefited from helpful comments at conference and seminar presentations. Killewald is grateful for support received from the Harvard Radcliffe Institute Fellowship Program. The NLSY79 survey is sponsored and directed by the U.S. Bureau of Labor Statistics and managed by the Center for Human Resource Research at The Ohio State University. Interviews are conducted by the National Opinion Research Center at the University of Chicago.

Eads and Tach ( 2016 ) examined relationship dissolution for both married and cohabiting couples and did not find strong evidence that assets and debts were differently associated with union stability for these groups. In an analysis of different-gender married couples, Eads et al. ( forthcoming ) found that a one-standard-deviation increase in net worth was associated with a 38% decline in the risk of union dissolution, net of controls.

Like Eads and Tach ( 2016 ), we considered that wealth may especially buffer financial stress following a negative income shock. However, neither they nor we found that wealth (for them, liquid and illiquid assets; for us, net worth) statistically significantly moderated the association between income loss and divorce risk.

We excluded 251 marriages because they did not overlap any survey waves, so the couple was never in the risk set.

We treated household rosters as complete and ignored the few biological children or stepchildren of unknown age and household members of unknown relationship to the respondent.

With the publicly available NLSY79 data, we could not adjust for the NLSY79’s multistage, stratified sampling (National Longitudinal Surveys n.d. ). Therefore, our analyses understate estimates’ uncertainty.

We tested for nonlinearity in these models, using linear splines for each asset value or equity. We could not reject the joint null hypothesis of linearity, so we used the linear terms in our analyses.

We considered respondents missing data on the start date of their first marriage to be currently unmarried if they had consistently reported being never married through the current wave.

The online supplement shows coefficients for control variables.

Results from the full model stratified by race/ethnicity are in the online supplement.

Premarital wealth is not a statistically significant predictor of divorce when added to either the baseline or the full model. For this analysis, we excluded marriages begun prior to 1985, since premarital wealth is not available for these marriages.

Tabular versions of Figure 1 and Figure 2 are in the online supplement.

This figure differs from the 4% of the sample who divorce by the next survey wave ( Table 1 ) because the interwave period is not always one year.

Results from models that add various groups of control variables sequentially and from models that include additional or alternative control variables are in the online supplement.

The associations of homeownership and vehicle ownership with divorce risk are not statistically significant in a complete-case analysis (see the online supplement).

For example, when net worth is removed from the full model and wealth is specified only with asset- and debt-specific measures, greater unsecured debt is positively and statistically significantly associated with divorce risk (see the online supplement).

Supplementary data

Data & figures.

Predictive margins of the probability of divorce in the next year, by net worth, baseline model. N  = 88,660. The 95% confidence band is indicated by the shaded area. Standard errors were clustered at the 1979 household level. The model is weighted and adjusts for exposure time. All financial variables are adjusted for inflation to 2020 values. See the text for details of the model specification.

Differences in predictive margins of divorce between $0 and alternative net worth values. N  = 88,660. The 95% confidence bands are indicated by the shaded areas. Standard errors were clustered at the 1979 household level. The model is weighted and adjusts for exposure time. All financial variables are adjusted for inflation to 2020 values. See the text for details of each model specification.

Descriptive statistics

Notes: For imputed variables, values are averaged across imputations. Results are weighted. All financial variables are adjusted for inflation to 2020 values.

Discrete-time hazard models of the association between net worth and divorce

Notes: N  = 88,660. Standard errors, clustered at the 1979 household level, are shown in parentheses. All models are weighted and adjust for exposure time. All financial variables are adjusted for inflation to 2020 values. See the text for details of each model specification. See the online appendix for coefficients on control variables.

* p  < .05; *** p  < .001

Variation in predictive margins of divorce in the next year, control variables

Notes: N  = 88,660. Standard errors, clustered at the 1979 household level, are shown in parentheses. All models are weighted and adjust for exposure time. All financial variables are adjusted for inflation to 2020 values. See the text for details of the model specification.

* p  < .05; ** p  < .01; *** p  < .001

Discrete-time hazard models of the association between net worth and divorce, specific assets

Notes: N  = 88,660. Standard errors, clustered at the 1979 household level, are shown in parentheses. All models are weighted and adjust for exposure time. All financial variables are adjusted for inflation to 2020 values. See the text for details of each model specification.

* p  < .05; ** p  < .01

Variation by portfolio composition in predictive margins of divorce in the next year, holding net worth at $50,000

Notes: N  = 88,660. Standard errors, clustered at the 1979 household level, are shown in parentheses. All models are weighted and adjust for exposure time. All financial variables are adjusted for inflation to 2020 values. Results in panels A and B are based on the Values and Equity models in Table 4 , respectively. See the text for details of each model specification. NS = not significant at the .05 significance level.

• Previous Issue
• Previous Article
• Next Article

Advertisement

Supplements

Citing articles via, email alerts, related articles, related topics, related book chapters, affiliations.

• About Demography
• Editorial Board
• For Authors
• Rights and Permissions Inquiry
• Online ISSN 1533-7790
• Print ISSN 0070-3370
• Copyright © 2024
• Duke University Press
• 905 W. Main St. Ste. 18-B
• Durham, NC 27701
• (888) 651-0122
• International
• +1 (919) 688-5134
• Information For
• Advertisers
• Book Authors
• Booksellers/Media
• Journal Authors/Editors
• Journal Subscribers
• Prospective Journals
• Licensing and Subsidiary Rights
• View Open Positions
• email Join our Mailing List
• catalog Current Catalog
• Accessibility
• Get Adobe Reader

This Feature Is Available To Subscribers Only

Sign In or Create an Account

Study: Conflict between divorced parents can lead to mental health problems in children

Karey O'Hara, assistant research professor of psychology. Photo by Robert Ewing.

Conflict between divorced or separated parents increases the risk of children developing physical and mental health problems.

A new study from the Arizona State University Research and Education Advancing Children’s Health (REACH) Institute has found that children experience fear of being abandoned when their divorced or separated parents engage in conflict. Worrying about being abandoned predicted future mental health problems in children. The work was published in Child Development on Jan. 12. 

“Conflict is a salient stressor for kids, and the link between exposure to interparental conflict and mental health problems in children is well established across all family types — married, cohabitating, separated and divorced,” said Karey O’Hara, a research assistant professor of psychology at ASU and first author on the paper.

“Conflict between divorced or separated parents predicted children experiencing fear that they would be abandoned by one or both parents. This feeling was associated with future mental health problems, especially for those who had strong relationships with their fathers.”

Based on studies including children from families with married or cohabitating parents, the researchers knew that children view interparental conflict as a threat, often wondering if their parents will get divorced. 

To understand how children with divorced or separated parents interpreted interparental conflict, the researchers surveyed families participating in the  New Beginnings Program , asking 559 children (aged 9–18 years) about their exposure to conflict. The questions included topics like whether their parents fought in front of them, spoke poorly of the other parent or asked children to carry messages. Children exposed to interparental conflict were more likely to report worrying about being abandoned by one or both of their parents.

“When parents who are married or cohabitating engage in conflict, the child might worry about their parents separating,” O’Hara said. “But children whose parents are divorced or separated have already seen the dissolution of their family. The idea that they might be abandoned might be unlikely, but it is not illogical from their perspective.” 

The fear of abandonment was persistent: Exposure to parental conflict predicted fear of abandonment three months later. And, worrying about abandonment predicted mental health problems, as reported by the children themselves and their teachers, 10 months later.

Because quality parent-child relationships are known to  buffer children  against stress, the researchers expected children who had strong relationships with a parent to experience less fear of abandonment and mental health problems. But the team did not find a general buffering effect of parenting. 

“A strong father-child relationship came at a cost when interparental conflict was high,” O’Hara said. “Having a high quality parenting relationship is protective, but it is possible that quality parenting alone is not enough in the context of high levels of interparental conflict between divorced parents.” 

The goal of ASU’s  REACH institute  is to bring research promoting children’s well-being from the lab into practice, and the research team is currently working on designing an intervention to help children cope with parental conflict after divorce. 

C. Aubrey Rhodes, Sharlene Wolchik, Irwin Sandler and Jenn Yun-Tien, all of ASU’s REACH Institute, also contributed to the work. This study was supported by the National Institute on Drug Abuse, National Institute of Mental Health and National Institute of Child Health and Human Development.

More Science and technology

Mapping a new field

A child reaches for a cookie on a hot tray, despite their parent’s warning, and experiences the sensation of a burn. That “burn…

ASU professors explore benefits of sharing 'hidden' identities in the classroom

Every person has characteristics, beliefs, values and affiliations that they identify with. These identities can define and…

2 ASU juniors awarded prestigious Goldwater Scholarship for excellence in STEM research

The Lorraine W. Frank Office of National Scholarships Advisement has announced that Arizona State University juniors Leslie…

Intermarriage and the risk of divorce in the Netherlands: the effects of differences in religion and in nationality, 1974-94

Affiliation.

• 1 Tilburg University, Netherlands. [email protected]
• PMID: 15764135
• DOI: 10.1080/0032472052000332719

A textbook hypothesis about divorce is that heterogamous marriages are more likely to end in divorce than homogamous marriages. We analyse vital statistics on the population of the Netherlands, which provide a unique and powerful opportunity to test this hypothesis. All marriages formed between 1974 and 1984 (nearly 1 million marriages) are traced in the divorce records and multivariate logistic regression models are used to analyse the effects on divorce of heterogamy in religion and national origin. Our analyses confirm the hypothesis for marriages that cross the Protestant-Catholic or the Jewish-Gentile boundary. Heterogamy effects are weaker for marriages involving Protestants or unaffiliated persons. Marriages between Dutch and other nationalities have a higher risk of divorce, the more so the greater the cultural differences between the two groups. Overall, the evidence supports the view that, in the Netherlands, new group boundaries are more difficult to cross than old group boundaries.

• Divorce / statistics & numerical data*
• Middle Aged
• Netherlands / epidemiology

Divorce and the status of women

Journal of marriage and the family • vol/iss. 41 • published in 1979 • pages: 375-385 •   cite, by pearson, jr., willie , hendrix, lewellyn.

Female status will be positively associated with divorce rate (378, 381)

Tested for region. Prediction holds everywhere except North America. Control variables do not significantly effect the relationship between status and divorce.

Related Hypotheses

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• HHS Author Manuscripts

Divorce and Death: A Case Study for Health Psychology

Marital separation and divorce are associated with increased risk for early death, and the magnitude of this association rivals that of many well-established public health factors. In the case of divorce, however, the mechanisms explaining precisely why and how some people are at risk for early death remain unclear. This paper reviews what is known about the association between divorce and risk for all-cause mortality, then discusses four emerging themes in this area of research: the biological intermediaries linking divorce to pathophysiology and disease onset, moving beyond the statistical mean, focusing research on the diathesis-stress model, and studying how opportunity foreclosures may place people on a trajectory toward poor distal health outcomes. These ideas are grounded in a set of public lay commentaries about the association between divorce and death; in this way, the paper seeks to integrate current research ideas with how the general public thinks about divorce and its correlates. Although this paper focuses on divorce, many of the emerging themes are applicable to the study of psychosocial stress and health more generally. Therefore, the study of divorce and death provides a good case study for health psychology and considers new questions that can be pursued in a variety of research areas.

The study of marital separation and divorce provides an ideal vantage point for understanding the association between psychological stress and health. 1 First, divorce is a major interpersonal stressor that has the potential to increase risk for a range of negative health outcomes, including increased risk for early death ( Amato, 2010 ; Sbarra, Law, & Portley, 2011 ). Thus, in and of itself, divorce is a significant public health risk, and studying the factors that contribute to good or poor adjustment is an important and timely endeavor. Second, the prevalence, demographics, and range of psychosocial responses to divorce make it a life event that can shed light on stress, coping, and health more generally. Nearly 40% of first marriages end in divorce ( Bramlett & Mosher, 2002 ; Kreider & Ellis, 2011 ) and two million adults are newly impacted by the end of marriage each year. For almost everyone involved, divorce often involves moving, dividing financial assets and, when kids are involved, negotiating visitation and custody arrangements. Despite these logistical challenges, most people cope well and report either minimal disruptions to their psychological wellbeing ( Mancini, Bonanno, & Clark, 2011 ) or a pattern of common grief defined by a period of acute distress that dissipates steadily over time ( Hetherington & Kelly, 2002 ). However, some people—perhaps between 10–20% of all divorcees ( Mancini et al., 2011 )— suffer quite substantially when their marriage ends; in these situations, divorce becomes a chronic stressor that is associated with lasting decrements in psychological wellbeing and physical health (e.g., Lucas, 2005 ). Identifying exactly who these people are and what processes unfold (prior to and after their separation) to place them at greatest risk are the key questions for the next generation of research on divorce and health.

This paper reviews what is known about the association between divorce and risk for all-cause mortality. In doing so, our primary goal is to outline four emerging themes that can advance the field. Although our discussion is specific to divorce, we believe the topics discussed here are germane to understanding the association between negative life events, psychosocial stress and distal health outcomes in general. For these reasons, the study of divorce and health is a useful case study for health psychology.

The emerging themes we discuss in this paper are based on a strong foundation of research, but one additional way to illustrate the topics of interest is to consider how lay people discuss the association between divorce and death. Lay conceptions of the self, intention, causality, and meaning provide an excellent counterbalance to psychological science’s pursuit of universal principles of human behavior ( Molden & Dweck, 2006 ). Related to divorce and death, a recent Huffington Post blog entry on the topic presents a unique opportunity to look carefully at how people outside of the research enterprise view this topic. In late 2011, Robert Hughes posted a column on his Huffington Post blog entitled, Will Divorce increase your Chances of Early Death? This post generated nearly 700 comments from The Huffington Post readership. To illustrate our emerging themes and, importantly, to underscore how people outside the research community think about these topics, we have chosen a small group of the comments to serve as jumping-off points for discussing the relevant research. As shown in Figure 1 , the comments remind us that nomothetic approaches to human behavior often fail to capture the richness of idiographic data (cf. Beck, 1953 ; Nesselroade & Ram, 2004 ). It is our view that a complete scientific analysis of divorce and death must integrate what people think about why this association exists with data generated from empirical (and largely nomothetic) studies. We discuss these comments in detail as we address each emerging theme.

Major themes of the paper and comments from The Huffington Post readership that illustrate these themes. A major challenge for nomothetic research is to fully capture and represent the richness of idiographic accounts of divorce.

Divorce and Death: What is known?

Integrating research in psychology, sociology, and epidemiology, Sbarra and colleagues ( Sbarra et al., 2011 ) recently published a meta-analysis on the association between divorce and death that synthesized data from 32 prospective studies (involving more than 6.5 million people, 160,000 deaths, and over 755,000 divorces in 11 different countries). The overall hazard ration (HR) linking divorce and risk for early death was 1.23, which indicates that relative to married adults divorced adults evidenced, on average and across all the studies in the review, a 23% greater risk of being dead each successive follow-up period. Although both men and women are at increased risk for early death following divorce, the meta-analysis revealed that divorced men evidenced greater risk for early death (HR = 1.31) than divorced women (HR = 1.13). More recently, Shor, Roelfs, Bugyi, and Schwartz (2012) have taken a broader look at the potential association between divorce and death by studying every published report on this topic, including cross-sectional studies from very large census samples. Their final sample included 600 million people from more than 24 countries. Consistent with the results of the more narrowly focused meta-analysis, Shor et al. (2012) observed a significant average RH of 1.30, as well as significant differences between men (HR = 1.37) and women (HR = 1.22).

The fact that divorce is associated with increased risk for early death raises a very interesting set of questions. Does marital separation play a causal role in hastening one’s time to death among people who are not otherwise at risk? Alternatively, is the association between divorce and death epiphenomenal and due to third variable confounds that increase risk for both divorce and death individually? Sbarra et al. (2011) addresses these types of questions by considering four potential pathways linking marital separation and subsequent health outcomes.

First, as mentioned, any serious consideration of this topic must contend with the fact that divorce is non-random and heritable ( Jocklin, McGue, & Lykken, 1996 ), and the factors selecting people out of marriage also predict poor health outcomes (e.g., Fu & Goldman, 2000 ). Although it appears that at least some of the association between divorce and death is due to third variable confounds, the best evidence for a potential causal association between relationship transitions and health comes from a study of co-twin control designs, which allow investigators to examine risk associated with the experience of a life event in one twin (proband) relative to his/her identical twin. Because monozygotic twins share 100% of their genotypes, discordant life events that result in increased risk for a negative outcome can be presumed to be causally associated with the outcome. Osler and colleagues ( Osler, McGue, Lund, & Christensen, 2008 ) recently used a co-twin control design to investigate rates of health outcomes between twins who were discordant for widowhood or divorce. The results indicated that depression and rates of smoking may be consequences of ending a marriage, but differences in many other health outcomes (e.g., self-rated health, alcohol use, BMI) may be due to underlying genetic explanations, not the stress of a relationship transition.

Using a similar co-twin design in a sample of over 1,900 pairs of twins discordant for spousal bereavement, Lichtenstein, Gatz, and Berg (1998) found evidence for a causal effect of bereavement on mortality. Of course, divorce is an entirely different life event, but the results of this work suggest that it is possible for the experience of social loss to causally increase risk for early mortality. A co-twin control study focused on divorce and mortality has yet to be conducted, and research of this nature will be a major advance in the study of divorce and health outcomes.

Assuming that at least a small portion of the association between divorce and death is causal or spurred by the stress of divorce itself, Sbarra et al. (2011) outlined three additional pathways of action. First, divorce has a negative impact on financial and social resources; these changes, in turn, can make the acute stress of marital separation much more chronic ( Lorenz, Wickrama, Conger, & Elder, 2006 ). Second, divorce is associated with a variety of poor health behaviors, including, for example, substantially elevated risk for severe insomnia and problems of sleep maintenance ( Doi, Minowa, Okawa, & Uchiyama, 2000 ; Hajak, 2001 ), cannabis initiation among previous abstainers ( Agrawal & Lynskey, 2009 ), increased risk for non-abstinent recovery among adults diagnosed with prior alcohol dependence ( Dawson, Grant, Stinson, & Chou, 2006 ), and increases in alcohol consumption immediately prior to a marital separation ( Mastekaasa, 1997 ). Finally, Sbarra et al. (2011) reviewed evidence that the psychosocial stress of divorce is associated with biological responses (e.g., blood pressure reactivity; Sbarra, Law, Lee, & Mason, 2009 ) that, if sustained over time, have direct relevance for negative health outcomes.

Together, these four pathways (selection, social and financial resource change, health behaviors, and psychosocial stress) are ripe for future research. Indeed, one of the central points raised by the Sbarra et al. (2011) review of potential mechanisms is that the available data on this topic are severely impoverished . Consider the following example: If most people cope well with a divorce in terms of their psychological adjustment, how can this life event be associated with such a large increase in the risk for poor health? These observations do not comport well, but the field has few studies that are in a position, for example, to determine whether remaining in a poor quality marriage is as potentially deleterious as a stressful separation experience. In the remaining sections, we address four emerging themes that have the potential to shed new light on this and many other questions of interest.

We Die from Illness, Not Divorce (Theme 1)

As the comments associated with Theme 1 ( Figure 1 ) suggest, responses to marital separation can range from relief to complete exhaustion. For some people, the prospect of a difficult divorce seems much easier than staying in a torturous marriage. With respect to health outcomes, few people would argue that divorce itself hastens disease; rather, if divorce exerts a causal effect on increased risk for early death, this life experience sets in motion a series of processes that contribute to the development and progression of disease pathophysiology. What are these processes?

This question is not unique to the divorce literature. One of the primary challenges facing health psychology is to identify biologically plausible pathways from psychosocial variables to disease outcomes ( Miller, Chen, & Cole, 2009 ). Miller et al. (2009) have advanced a conceptual approach to studying how biobehavioral responses to life events are associated with and may regulate (and be regulated by) molecular and cellular responses that are disease-relevant. This approach highlights the role of biological intermediaries, or, as Miller et al. (2009) state, “Once a robust linkage between a psychosocial factor and a clinical health outcome is identified, the next step is to determine what biological processes convey those effects into the physical environment of disease pathogenesis” (p. 504).

Related to divorce, Miller et al.’s (2009) first criterion is satisfied. Beyond the meta-analytic findings we described above ( Sbarra et al., 2011 ), other studies have demonstrated that divorce can result in a “proliferation of stress” that mediates the association between the end of marriage and poor self-rated health up to a decade later ( Lorenz et al., 2006 ).We do not yet know, however, the biological intermediaries that translate the psychosocial stress of divorce into poor physical health. We have some hints as to how these effects might operate, but lack clear-cut, disease-relevant pathways. For example, Sbarra et al. (2009) demonstrated that high levels of divorce-related emotional intrusion are associated with elevated blood pressure, and that men who find thinking about their separation especially difficult evidence substantial blood pressure increases when asked to do so. Similarly, Kiecolt-Glaser and colleagues ( Kiecolt-Glaser et al., 1987 ) demonstrated that relative to married adults, divorced adults had significantly higher antibody titers to the Epstein Barr Virus and lower percentage of Natural Killer cell activity, both of which indicate compromised immune functioning. Among the divorced group, a shorter separation period and continued attachment to a former spouse were associated with poorer physiological outcomes, suggesting that psychological variables specific to divorce adjustment (e.g. continued attachment to a former spouse) are associated with impaired immune functioning.

Conceptually, the notion of allostatic load ( McEwen, 1998 , 2000 ) provides a compelling model for understanding the cumulative wear-and-tear on the body associated with divorce. Allostasis describes a physiological or behavioral adaptation to environmental changes in order to restore normality, or homeostasis ( McEwen, 2000 ). Excessive cardiovascular, neuroendocrine, and immune activations in response to stress can promote vascular remodeling, initiate atherosclerotic plaque growth, and alter gene expression in a manner that contributes to disease pathogenesis ( Cole, 2009 ; Cole et al., 2007 ; Libby, 2002 ; Miller, et al., 2009 ; Miller, Cohen, & Ritchey, 2002 ).

When seeking to understand this cumulative wear-and-tear in the case of divorce, we must contend with two important processes: equifinality and multifinality ( Cicchetti & Rogosch, 1996 ). Equifinality refers to the idea that multiple starting points can converge on a single outcome; conversely, multifinality refers to the idea that a single starting point can yield multiple distinct end points. For example, non-acceptance of divorce (often thought of as a longing for reunion with one’s ex-partner), loneliness, and emotional inhibition may all increase risk for depression in the aftermath of a separation. Depression, in turn, can promote cardiovascular and inflammatory responding that potentiates the development of atherosclerosis and cardiovascular disease ( Miller & Blackwell, 2006 ). Alternatively, through the same biological mechanism, inflammation can increase risk for a range of different disease outcomes.

To untangle this complex web of associations, Miller et al. (2009) suggest researchers “reverse engineer” adverse health outcomes into their disease precursors, then identify the specific psychological or behavioral variables that are associated with the biological intermediaries shaping these molecular and cellular processes. For research on divorce and health to advance, we must heed this advice and identify (a) the specific processes that make divorce a chronic stressor (see Lorenz et al., 2006 ), and (b) why, exactly, divorce increases risk for a major depressive episode (see Kendler, Hettema, Butera, Gardner, & Prescott, 2003 ) or other forms of persistent psychological distress. Although it is likely that chronic stress and mood disturbances represent the major pathways that drive divorce “under the skin,” the divorce-specific tributaries that feed into these pathways remain to be identified.

What Does the “Statistical Mean” Mean ? (Theme 2)

Any consideration of the meta-analytically derived association between divorce and risk for early mortality must contend with the fact that the overall effect size represents the statistical mean across all the studies in question. The statistical mean is highly susceptible to the influence of outliers and, as a stand-alone indicator, masks substantial variability. We see from the comments associated with Theme 2 ( Figure 1 ) that many people report improved functioning following their divorce. If a large percentage of people are relieved—and even happy— to be free of the weight of a bad marriage, how can divorce also increase risk for poor health?

The answer to this question rests in moving beyond the statistical mean when studying divorce and health. As an illustration of the central problem in this literature, consider two studies that rely on data from the German Socio-Economic Panel Study ( Lucas, 2005 ; Mancini et al., 2011 ). Lucas (2005) used multilevel modeling to examine mean trajectories of life satisfaction prior to and following divorce. The primary findings from this report were that life satisfaction steadily decreases prior to divorce and that people do not return to their initial set-point after the divorce ( Lucas, 2005 ). Using data from the same sample, Mancini et al. (2011) applied a series of latent growth mixture models to identify potential sub-samples, or classes, of how people respond to divorce. This study revealed that nearly 72% of adults reported high levels of life satisfaction prior to and following their divorce; 9% reported low levels of satisfaction that increased substantially following the divorce; and 19% reported a moderate decline in satisfaction over the study period ( Mancini et al., 2011 ). The difference between the main findings in these studies, which use almost entirely overlapping data, is substantial. The illustration serves as an important reminder that, in many instances, the statistical mean may be essentially meaningless for drawing veridical conclusions about the psychosocial and health sequelae of divorce.

With respect to health outcomes, this problem is magnified because the available data are extremely impoverished. The primary explanation for the so-called “impoverished data situation” is that representative epidemiological studies rarely measure psychosocial variables with any degree of complexity (see Sbarra & Mason, in press). Consequently, prospective studies that can speak to the long-term health correlates of divorce do not typically assess key moderators such as whether a person initiated the separation or was left by their partner, or whether distal health outcomes are a function of participants’ longing for their ex-partner.

Given these concerns, we suggest moving forward to examine two distinct avenues. First, epidemiological evidence indicates that specific aspects of adults’ “marital biographies” may be associated with differential outcomes. Amato and Hohmann-Marriott (2007) , for example, found that adults in high-conflict marriages reported an increase in life happiness following divorce, whereas adults in low-conflict marriages reported a decrease . An obvious implication of this study is that responses to divorce are highly dependent on marital quality prior to the separation.

With respect to physical health outcomes, Sbarra and Nietert (2009) found that divorce was associated with significant risk for early mortality among people who separated from their partner and never remarried; in contrast, divorced adults who subsequently remarried showed no elevated risk (for related studies, see Brockmann & Klein, 2004 ; Hughes & Waite, 2009 ; Zhang & Hayward, 2006 ). To the extent that remarrying mitigates risk in some cases, what processes explain this effect? The study of remarriage provides a natural interrupted time series for the study of divorce and health. A detailed study of the extent to which changes in loneliness, social capital, financial resources, and chronic stress explain the potential benefits of remarriage is needed. These variables can be studied as mediators or moderators of recovery from divorce, and doing so may help reveal important variability in average effects across time.

A second avenue of study for improving the impoverished data situation is to collect more detailed psychological measures about the experience of divorce, then link these measures to health-relevant or disease endpoints. High-quality, processes-focused research in this area is almost absent. This lack of empirical research on the psychology of divorce is surprising given longstanding observations that divorce can induce shame, longing, loneliness, humiliation, rumination, identity disruptions, and prolonged anger or grief ( Weiss, 1975 ). Presumably, it is these emotional experiences that give rise to, or at least co-vary with, more severe forms of psychopathology ( Afifi, Cox, & Enns, 2006 ), and comorbid psychological disorders also represent important (potential) moderators. An excellent example of this type of work is a study by Kendler and colleagues ( Kendler, et al., 2003 ), which demonstrated that stressful loss events characterized by a high degree of humiliation (e.g. your partner leaving you in a very public way) are associated with a 22% increase in risk for a depressive episode within the month of event occurrence. These findings suggest that the psychological dimensions of a loss event—including humiliation, entrapment, and danger (see Brown, Harris, & Hepworth, 1995 ), not the events in question—are the most salient predictors of important mental health outcomes. To overcome the problem of impoverished data, work of this nature is greatly needed in the study of divorce and health; deepening our understanding of the psychosocial moderators of the association between divorce and death will surely help the field move beyond the statistical mean.

Back to the Future: Revisiting the Diathesis-Stress Model (Theme 3)

The quote associated with Theme 3 reminds us that it is essential to understand divorce within the broader context of a person’s complete set of life experiences. When we do so, it becomes hard to view the stress of divorce as a unique, stand-alone main effect; rather, the essence of Theme 3 is that we look to a person’s past experiences to understand how they might respond to their marital separation. In many ways, Theme 3 is an extension of our argument for studying moderators and can be located squarely within the diathesis-stress model.

Figure 2 provides a multilevel model for conceptualizing how distal vulnerabilities may interact with more proximal social, cognitive/affective, and health behavior responses to predict health following a stressful divorce. An obvious criticism of the model is that we suggest an “everything is associated with everything else” process, which, by definition, lacks theoretical and conceptual specificity. We recognize this concern, but note that the model captures much of the complexity inherent in understanding why some people may be at unique risk for divorce, as well as poor outcomes following divorce. Although it is well known that many different variables may serve as underlying vulnerabilities within the diathesis-stress model ( Monroe & Simons, 1991 ), we limit our discussion to two processes that we believe are ripe for investigation related to divorce, both of which have an origin in early childhood experience: life history strategies and deoxyribonucleic acid (DNA) methylation.

A multilevel framework for depicting how distal vulnerabilities may interact with more proximal social, cognitive/affective, and health behavior responses to predict health following divorce.

Life history strategies

Life History Theory (LHT; Charnov, 1993 ; Stearns, 1977 ) is an evolutionary framework for understanding how people allocate resources toward their survival— or, more specifically, the survival of their genes in the population. How people allocate resources toward or away from both reproductive efforts and parenting investments constitutes a fundamental trade-off within LHT; individual differences in life history strategies appear to be a function of both heritable individual differences and early rearing environments (see Ellis, 2004 ; Figueredo et al., 2006 ).

Most humans devote their resources to continued survival (vs. reproductive efforts) and to deep parental investments (vs. mating efforts), but within this broad range of functioning, people vary quite substantially. These behavioral and psychological strategies are referred to as differential K strategies (the “K” terminology is derived from basic mathematic models in ecology and evolutionary biology; Rushton, 2004 ). People at the low end of the differential K-strategy continuum are more likely to evidence risk-taking, low parental investment, and disregard for social conventions; in contrast, people at the high end of the continuum evidence a psychological and behavioral repertoire that appears much more stable and suited for both longevity and parental investment (see Figuredo et al., 2006).

We believe that studying differential K-strategies in the context of divorce holds considerable promise. If people who engage in low-K strategies (as a function of their genetic propensity and developmental history) are more likely to be promiscuous in general and, after their marriage ends, to engage in more health-damaging behaviors, and/or to have difficulty maintaining high-quality social relationships, we suspect it is these people who will also evidence the worst health outcomes. Thus, LHT provides an integrative framework for thinking about who will (and will not) engage in a constellation of risk behaviors and be faced with—or perhaps create— chronic stress after divorce. Ultimately, in terms of studying variability in distal health outcomes (see Figure 2 ), LHT-derived behavioral and psychological profiles may prove especially tractable.

Epigenetic programming

Beyond contributing to behavioral and psychological profiles that may convey health risk, early rearing experiences can program biological responses in a manner that places people at risk for poor health later in life. Early social behaviors can cause molecular changes at the level of the genome that promote or constrain health-relevant biological responses. As a discipline, epigenetics investigates how experience can change gene expression without altering the underlying nucleotide structure, and startling evidence reveals that these changes can alter observable phenotypes across multiple generations (e.g., Singh, Murphy, & O'Reilly, 2002 ).

The most well-developed animal model for the lasting effect of early care is the rat dam-pup dyad. Maternal rats (dams) show natural variability in two caregiving behaviors, licking and grooming (LG) and arched-back nursing (ABN), and these behaviors lead to changes in pups’ hypothalamic-pituitary-adrenal (HPA) responses to stress ( Meaney, 2001 ). Pups of high-LG-ABN damns show less fearful behavioral responding and less HPA responding than pups of low-LG-ABN mothers ( Liu et al., 1997 ). These differences emerge over the first week of life, are a function of changes in the glucocorticoid receptor gene promoter region in the hippocampi, and are maintained into adulthood ( Liu et al., 1997 ). Specifically, these maternal behaviors result in changes in DNA methylation and chromatin structures, which are chemical changes to the genome that result in differences in gene expression ( Weaver et al., 2004 ). In effect, these changes represent a genomic imprinting by maternal behavior that has long-term implications for stress responding as a function of hippocampal glucocorticoid receptor (GR) gene expression ( Zhang et al., 2006 ).

How can epigenetics inform the study of adult relationships and health? Miller, Chen, and Parker (2011) have recently advanced a model of biological embedding that explains how early adverse experiences can alter cellular signals and sculpt biological processes in a manner that results in increased risk for poor health in later life. One view of these biological changes is a “defensive programming hypothesis”— the idea that early stress sensitizes biological systems to be highly responsive in order to promote survival. This sensitization, or programming, has adaptive value in the short-term but can be physiologically destructive over the long-term through exaggerated neuroendocrine and inflammatory responses ( Miller et al., 2009 ; Zhang, et al., 2006 ).

This idea maps nicely onto the LHT strategies discussed above and can more easily be understood from a behavioral ecology standpoint. Predictive adaptive responses (PARs) are biologically programmed reactions to expected environmental stressors ( Gluckman & Hanson, 2004 ). They serve the purpose of creating a biological set point for reactivity to stressors and ready the organism for the rest of its life based on its current environment. The programming occurs during the critical period of child development, allowing these responses to be functionally embedded in the body’s physiology. These responses take diverse forms, involving the cardiovascular, neuroendocrine, and immune systems. The traditional ancestral environment contained dangerous situations relevant to these PARs. If separated from the social group, cuts and scratches from predators and the stresses of living alone would favor an ancestral human with a primed proinflammatory response ( Cole, et al., 2007 ). In conditions of constant fight-or-flight activation, it would behoove our ancestors to develop an efficient system for the rapid delivery of glucocorticoids in service of mobilizing bodily resources for action. What in contemporary times may be considered a hyper-adaptation would, in the ancestral environment, be considered the most efficient strategy for survival. If living beyond a few decades was rare, the wear-and-tear of allostatic load associated with PARs was never an issue. These vestigial adaptations may still function in today’s society under specific childhood circumstances, but now result in deterioration of the body’s various systems over time.

These ideas have direct applications to studying susceptibility for disease following divorce. We can make a strong moderation hypothesis that the effects of divorce on health outcomes should depend on the extent of DNA methylation in specific gene regions. This gene by environment approach is emerging with increased frequency in the biological psychiatry literature (e.g., Ressler et al., 2011 ) and can be easily extended to the study of divorce adjustment.

Opportunity Foreclosures and Branching Trees (Theme 4)

Sbarra et al. (2011) suggested that changes in social capital and resources might explain the association between divorce and increased risk for early mortality, but did not articulate a framework for understanding how such processes might unfold. As the quotes associated with Theme 4 suggests, it is obvious and sensible that large movements across the gradient of social resources can impair health (e.g., Adler et al., 1994 ). We do not yet know specifically how these types of changes combine with other life experiences to canalize a person on a pathway toward good or poor outcomes. From our vantage point, understanding associations between life stress and health can be greatly enhanced by invoking models from the field of developmental psychopathology, which defines pathological outcomes in terms of developmental deviations.

Sroufe (1997) conceptualizes development and pathology using the metaphor of a branching tree: Development begins in the trunk of the tree with many potential endpoints; these endpoints quickly become limited or foreclosed as people experience difficult life events. For example, not completing high school is a deviation from normative development that, on average, forecloses many lifetime opportunities for the accumulation of wealth. Similarly, initiating tobacco use at a young age moves a person from the trunk of good health to a branch with an increased likelihood of cardiovascular disease or cancer; it is not impossible to return to good health, but the likelihood of doing so is diminished and, potentially, foreclosed over time.

The branching tree metaphor provides an excellent framework for understanding how divorce may be associated with distal health outcomes given a person’s unique life experiences. In terms of data analysis, classification and regression tree (CART) analysis ( Breiman, Friedman, Olshen, & Stone, 1984 ) can be used to represent the branching limbs toward good or poor health. CART analysis has its roots in machine learning and is a nonparametric modeling approach that uses recursive partitioning procedures to create a series of binary outcomes. A detailed discussion of CART techniques is beyond the scope of this review, but we note that these procedures can use both continuous and binary data for classification in node membership.

To illustrate how this might work in the case of divorce, Figure 3 depicts a fictional CART beginning with someone’s propensity to divorce (see Rosenbaum & Rubin, 1983 ). Propensity score matching is a tool for adjusting “exposure” effects for measured confounds in a non-randomized design ( Thoemmes & Kim, 2011 ). As mentioned above, some people are much more likely than other people to become divorced. Thus, every divorced adult has a propensity to divorce—e.g., people who report considerable marital conflict are more likely to divorce than people with low marital conflict. We can hypothesize that people who report marital conflict (or more depression, hostility, neuroticism, and/or a history of parental divorce) have a greater propensity to divorce than those people who are low on this dimension. Propensity score matching is akin to statistical control, but it allows for a complete quantification of a person’s probability for having been exposed to a specific life event. A propensity score, in turn, can serve as an important moderator of divorce adjustment (i.e. who is most vulnerable to the stress of divorce) and thus represents the first node in the classification tree.

A fictional example of how opportunity foreclosures can be studied using classification and regression tree (CART) analyses following divorce. The key conceptual ideas are that movement from the divorce toward distal health outcomes (pictured at the top of the figure) can be understood in terms of cumulative risk exposures, each of which foreclose opportunities for positive outcomes and increase the probability for poor outcomes in the future. People at low risk (i.e., who have low propensity to be divorced) can realize negative health outcomes, just as people at high risk can realize positive health outcomes. CRP = C-reactive protein; BP = Blood pressure; BMI = Body mass index.

As shown in the figure, the tree ends with three negative health outcomes: a cardiovascular event, elevated blood pressure, and elevated C-reactive protein (CRP), an index of systemic inflammation. We have created a fictional tree in which the midpoint nodes are fairly different in each branch of the tree—for some, low social support is highly correlated with smoking initiation, whereas for others high levels of loneliness are linked with sleep disturbances. To ultimately link the stress of divorce with distal health outcomes will require multi-wave prospective data. Until this data is available, cross-sectional CART models should be explored. The notion of branching trees and opportunity foreclosures are appealing theoretical ideas for the study of psychosocial stress and health, and it will be highly productive if health psychology tackles ideas that have roots in developmental psychopathology (see Sroufe, 1997 ).

Conclusions

This paper discusses four emerging themes in the study of divorce and health. Our goal was to present a case study for health psychology by illustrating how the emerging themes can shed increased light on exactly why and how some people are at risk for early death following a marital separation. To contextualize the ideas in everyday language, we linked each of our themes to comments on a Huffington Post blog about divorce and death. These selected quotes illustrate many important ideas that have turned the study of divorce into a useful exploration of emerging trends in the field of psychosocial stress and health more generally.

Acknowledgments

Preparation of this paper was supported by grants from the National Science Foundation (BCS#0919525) and the National Institute of Aging (AG#036895) to the first author.

1 For simplicity, we refer to separated and divorced adults as divorced throughout this proposal. When distinctions between marital separation and legal divorce are meaningful, we use more precise terminology.

• Adler NE, Boyce T, Chesney MA, Cohen S, Folkman S, Kahn RL, Syme SL. Socioeconomic status and health: The challenge of the gradient. American Psychologist. 1994; 49 :15–24. [ PubMed ] [ Google Scholar ]
• Afifi TO, Cox BJ, Enns MW. Mental health profiles among married, never-married, and separated/divorced mothers in a nationally representative sample. Social psychiatry and psychiatric epidemiology. 2006; 41 :122–129. [ PubMed ] [ Google Scholar ]
• Agrawal A, Lynskey MT. Correlates of later-onset cannabis use in the National Epidemiological Survey on Alcohol and Related Conditions (NESARC) Drug and Alcohol Dependence. 2009; 105 :71–75. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Amato PR. Research on divorce: Continuing trends and new developments. Journal of Marriage & Family. 2010; 72 :650–666. [ Google Scholar ]
• Amato PR, Hohmann-Marriott B. A Comparison of High- and Low-Distress Marriages That End in Divorce. Journal of Marriage and Family. 2007; 69 :621–638. [ Google Scholar ]
• Beck SJ. The science of personality: nomothetic or idiographic? Psychological Review. 1953; 60 :353–359. [ PubMed ] [ Google Scholar ]
• Bramlett MD, Mosher WD. Cohabitation, Marriage, Divorce, and Remarriage in the United States. Hyattsville, MD: National Center for Health Statistics; 2002. [ PubMed ] [ Google Scholar ]
• Breiman L, Friedman JH, Olshen RA, Stone CJ. Classification and regression trees. Belmont, CA: Wadsworth; 1984. [ Google Scholar ]
• Brockmann H, Klein T. Love and death in Germany: The marital biography and its effect on mortality. Journal of Marriage and Family. 2004; 66 :567–581. [ Google Scholar ]
• Brown GW, Harris TO, Hepworth C. Loss, humiliation and entrapment among women developing depression: a patient and non-patient comparison. Psychological Medicine. 1995; 25 :7–22. [ PubMed ] [ Google Scholar ]
• Charnov EL. Life history invariants: some explorations of symmetry in evolutionary ecology. New York: Oxford University Press; 1993. [ Google Scholar ]
• Cicchetti D, Rogosch FA. Equifinality and multifinality in developmental psychopathology. Development and Psychopathology. 1996; 8 :597–600. [ Google Scholar ]
• Cole SW. Social regulation of human gene expression. Current Directions in Psychological Science. 2009; 18 :132–137. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Cole SW, Hawkley LC, Arevalo JM, Sung CY, Rose RM, Cacioppo JT. Social regulation of gene expression in human leukocytes. Genome Biol. 2007; 8 :R189. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Dawson DA, Grant BF, Stinson FS, Chou PS. Maturing out of alcohol dependence: the impact of transitional life events. J Stud Alcohol. 2006; 67 :195–203. [ PubMed ] [ Google Scholar ]
• Doi Y, Minowa M, Okawa M, Uchiyama M. Prevalence of sleep disturbance and hypnotic medication use in relation to sociodemographic factors in the general Japanese adult population. J Epidemiol. 2000; 10 :79–86. [ PubMed ] [ Google Scholar ]
• Ellis BJ. Timing of pubertal maturation in girls: an integrated life history approach. Psychological bulletin. 2004; 130 :920. [ PubMed ] [ Google Scholar ]
• Figueredo AJ, V?squez G, Brumbach BH, Schneider SMR, Sefcek JA, Tal IR, Jacobs WJ. Consilience and life history theory: From genes to brain to reproductive strategy. Developmental Review. 2006; 26 :243–275. [ Google Scholar ]
• Fu H, Goldman N. The association between health-related behaviours and the risk of divorce in the USA. Journal of Biosocial Science. 2000; 32 :63–88. [ PubMed ] [ Google Scholar ]
• Gluckman PD, Hanson MA. Living with the past: evolution, development, and patterns of disease. Science. 2004; 305 :1733–1736. [ PubMed ] [ Google Scholar ]
• Hajak G. Epidemiology of severe insomnia and its consequences in Germany. European Archives of Psychiatry and Clinical Neuroscience. 2001; 251 :49–56. [ PubMed ] [ Google Scholar ]
• Hetherington EM, Kelly J. For better or for worse: Divorce reconsidered. New York: Norton & Company; 2002. [ Google Scholar ]
• Hughes ME, Waite LJ. Marital biography and health at mid-life. Journal of Health and Social Behavior. 2009; 50 :344–358. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Jocklin V, McGue M, Lykken DT. Personality and divorce: A genetic analysis. Journal of Personality and Social Psychology. 1996; 71 :288–299. [ PubMed ] [ Google Scholar ]
• Kendler KS, Hettema JM, Butera F, Gardner CO, Prescott CA. Life Event Dimensions of Loss, Humiliation, Entrapment, and Danger in the Prediction of Onsets of Major Depression and Generalized Anxiety. Arch Gen Psychiatry. 2003; 60 :789–796. [ PubMed ] [ Google Scholar ]
• Kiecolt-Glaser JK, Fisher LD, Ogrocki P, Stout JC, Speicher CE, Glaser R. Marital quality, marital disruption, and immune function. Psychosomatic Medicine. 1987; 49 :13–34. [ PubMed ] [ Google Scholar ]
• Kreider RM, Ellis R. Number, timing, and duration of marriages and divorces: 2009. US Department of Commerce, Economics and Statistics Administration, US Census Bureau; 2011. [ Google Scholar ]
• Libby P. Inflammation in atherosclerosis. Nature. 2002; 420 :868–874. [ PubMed ] [ Google Scholar ]
• Lichtenstein P, Gatz M, Berg S. A twin study of mortality after spousal bereavement. Psychological Medicine. 1998; 28 :635–643. [ PubMed ] [ Google Scholar ]
• Liu D, Diorio J, Tannenbaum B, Caldji C, Francis D, Freedman A, Meaney MJ. Maternal care, hippocampal glucocorticoid receptors, and hypothalamic-pituitary-adrenal responses to stress. Science. 1997; 277 :1659–1662. [ PubMed ] [ Google Scholar ]
• Lorenz FO, Wickrama KA, Conger RD, Elder GH., Jr The short-term and decade-long effects of divorce on women's midlife health. J Health Soc Behav. 2006; 47 :111–125. [ PubMed ] [ Google Scholar ]
• Lucas RE. Time does not heal all wounds: A longitudinal study of reaction and adaptation to divorce. Psychological Science. 2005; 16 :945–950. [ PubMed ] [ Google Scholar ]
• Mancini AD, Bonanno GA, Clark AE. Stepping Off the Hedonic Treadmill. Journal of Individual Differences. 2011; 32 :144–152. [ Google Scholar ]
• Mason AE, Sbarra D. Romantic separation, loss, and health: A review of moderators. In: Newman M, Roberts N, editors. Handbook of Health and Social Relationships. Washington, DC: American Psychological Association; In Press. [ Google Scholar ]
• Mastekaasa A. Marital dissolution as a stressor: Some evidence on psychological, physical, and behavioral changes in the pre-separation period. Journal of Divorce & Remarriage. 1997; 26 :155–183. [ Google Scholar ]
• McEwen BS. Stress, adaptation, and disease - Allostasis and allostatic load. Neuroimmunomodulation. 1998; 840 :33–44. [ PubMed ] [ Google Scholar ]
• McEwen BS. Allostasis and allostatic load: Implications for Neuropsychopharmacology. Neuropsychopharmacology. 2000; 22 :108–124. [ PubMed ] [ Google Scholar ]
• Meaney MJ. Maternal care, gene expression, and the transmission of individual differences in stress reactivity across generations. Neuroscience. 2001; 24 [ PubMed ] [ Google Scholar ]
• Miller G, Chen E, Cole S. Health psychology: Developing biologically plausible models linking the social world and physical health. Annual Review of Psychology. 2009; 60 :501–524. [ PubMed ] [ Google Scholar ]
• Miller GE, Blackwell E. Turning Up the Heat: Inflammation as a Mechanism Linking Chronic Stress, Depression, and Heart Disease. Current Directions in Psychological Science. 2006; 15 (6):269–272. [ Google Scholar ]
• Miller GE, Chen E, Fok AK, Walker H, Lim A, Nicholls EF, Kobor MS. Low early-life social class leaves a biological residue manifested by decreased glucocorticoid and increased proinflammatory signaling. PNAS Proceedings of the National Academy of Sciences of the United States of America. 2009; 106 :14716–14721. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Miller GE, Chen E, Parker KJ. Psychological Stress in Childhood and Susceptibility to the Chronic Diseases of Aging: Moving Toward a Model of Behavioral and Biological Mechanisms. Psychological bulletin. 2011; 137 :959–997. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Miller GE, Cohen S, Ritchey AK. Chronic psychological stress and the regulation of pre-inflammatory cytokinesL A glucocorticoid-resistance model. Health Psychology. 2002; 21 :531–541. [ PubMed ] [ Google Scholar ]
• Molden DC, Dweck CS. Finding "Meaning" in Psychology: A Lay Theories Approach to Self-Regulation, Social Perception, and Social Development. American Psychologist. 2006; 61 :192–203. [ PubMed ] [ Google Scholar ]
• Monroe SM, Simons AD. Diathesis-stress theories in the context of life stress research: Implications for the depressive disorders. Psychological bulletin. 1991; 110 :406. [ PubMed ] [ Google Scholar ]
• Nesselroade JR, Ram N. Studying intraindividual variability: What we have learned that will help us understand lives in context. Research in Human Development. 2004; 1 :9–29. [ Google Scholar ]
• Osler MMDD, McGue MMP, Lund RMDP, Christensen KMDD. Marital Status and Twins' Health and Behavior: An Analysis of Middle-Aged Danish Twins. [Review] Psychosomatic Medicine. 2008; 70 :482–487. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Ressler KJ, Mercer KB, Bradley B, Jovanovic T, Mahan A, Kerley K, Myers AJ. Post-traumatic stress disorder is associated with PACAP and the PAC1 receptor. Nature. 2011; 470 :492–497. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983; 70 :41–55. [ Google Scholar ]
• Rushton JP. Placing intelligence into an evolutionary framework or how g fits into the r – K matrix of life-history traits including longevity. Intelligence. 2004; 32 (4):321–328. [ Google Scholar ]
• Sbarra DA, Law RW, Lee LA, Mason AE. Marital dissolution and blood pressure reactivity: Evidence for the specificity of emotional intrusion-hyperarousal and task-rated emotional difficulty. Psychosomatic Medicine. 2009; 71 :532–540. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Sbarra DA, Law RW, Portley RM. Divorce and death: A meta-analysis and research agenda for clinical, social, and health psychology. Perspectives on Psychological Science. 2011; 6 :454–474. [ PubMed ] [ Google Scholar ]
• Sbarra DA, Nietert PJ. Divorce and death: Forty years of the Charleston Heart Study. Psychological Science. 2009; 20 :107–113. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Shor E, Roelfs DJ, Bugyi P, Schwartz JE. Meta-analysis of marital dissolution and mortality: Reevaluating the intersection of gender and age. Social Science & Medicine. 2012 [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Singh S, Murphy B, O'Reilly R. Epigenetic contributors to the discordance of monozygotic twins. Clinical genetics. 2002; 62 :97–103. [ PubMed ] [ Google Scholar ]
• Sroufe LA. Psychopathology as an outcome of development. Development and Psychopathology. 1997; 9 :251–268. [ PubMed ] [ Google Scholar ]
• Stearns SC. The evolution of life history traits: a critique of the theory and a review of the data. Annual Review of Ecology and Systematics. 1977; 8 :145–171. [ Google Scholar ]
• Thoemmes FJ, Kim ES. A systematic review of propensity score methods in the social sciences. Multivariate behavioral research. 2011; 46 :90–118. [ PubMed ] [ Google Scholar ]
• Weaver ICG, Cervoni N, Champagne FA, D'Alessio AC, Sharma S, Seckl JR, Meaney MJ. Epigenetic programming by maternal behavior. Nature neuroscience. 2004; 7 :847–854. [ PubMed ] [ Google Scholar ]
• Weiss RS. Marital separation. New York: Basic Books; 1975. [ Google Scholar ]
• Zhang TY, Bagot R, Parent C, Nesbitt C, Bredy TW, Caldji C, Meaney MJ. Maternal programming of defensive responses through sustained effects on gene expression. Biological Psychology. 2006; 73 (1):72–89. [ PubMed ] [ Google Scholar ]
• Zhang Z, Hayward MD. Gender, the marital life course, and cardiovascular disease in late midlife. Journal of Marriage and Family. 2006; 68 :639–657. [ Google Scholar ]