Cultural and personal values interact to predict divorce
Introduction
Divorce is a major life event, leaving an impactful impression on the divorcees and their dependents. Divorce is associated with fundamentally important outcomes such as economic status1,2, well-being3,4,5,6, health condition7,8, early mortality9,10,11, and more. The last century has seen a striking global surge in divorce cases12. Given the growing population of individuals experiencing divorce and the scope of consequences they suffer, it is crucial to better understand its antecedents.
Much of the existing literature on divorce have successfully outlined the characteristics of individuals who get divorced (e.g., their age, financial status, parenthood, past divorce, relationship satisfaction, and emotional stability13,14,15,16,17,18,19. While these studies focused on who gets divorced, little is known about why people get divorced. The current research addresses this gap by focusing on basic, trans-situational motivations, namely, values. Values represent abstract goals that motivate perceptions, decisions, and actions. They can therefore shed light on what drives divorce. In addition, whereas previous studies have pointed to attributes that are associated with divorce concurrently, they are less useful in predicting who will get divorced in the future. Values are useful in that respect as well: they develop early in life and are relatively stable throughout adulthood. Therefore, even at a relatively young age, values may prospectively inform about the likelihood of divorce. By examining the role of values in divorce, we aim to offer a more comprehensive understanding of the phenomenon.
The current research expands past findings in several ways. First, we investigate values at both the individual and cultural levels, as well as the interaction between them. Taking this approach allows us to study the effects of motivations (i.e., personal values) and norms (i.e., cultural values), both separately and in interaction. In addition, our research employs comprehensive theories of cultural and personal values, that allow for a systematic examination of the full spectrum of values. The study may therefore shed light on which values are related to divorce and which are not. Finally, past studies often assumed that divorce marks individualism20,21,22, but rarely tested their association empirically (for an exception see23). We empirically test our hypotheses using data from three independent sources, studying a larger and more diverse set of countries than before.
Taking a cost-benefit approach to divorce may reveal the motivation to seek or avoid it. We reason that from the societal perspective, divorce threatens the social order and depletes societal resources. These costs may be apparent in the need to control ex-family revengeful violence or to provide support (e.g., financial, emotional) for single parents and their children. At the same time, divorce allows for societal transformation and is a right claimed by society members. Allowing individuals to pursue their wishes benefits the collective by increasing overall contentment. At the individual level, the cost end includes the risk to one’s public image, difficulty in disappointing family members, financial and emotional challenges, and the threat to children’s well-being. On the benefit end, divorce allows for change and growth. It allows the freedom to fully express oneself, experience a variety of relationships, and pursue own goals.
At both levels of analysis, these up and downsides of divorce can be arranged around the balance between collective and individual interests. Drawing on research showing that both societies and individuals vary in their emphasis on collective functioning versus individual growth, we suggest that perceptions of divorce and the action of divorcing, are shaped by cultural and personal values that express this broad issue.
Values are general and abstract principles regarding what is good and desired24,25,26,27. They exist at both the collective and individual levels28,29. Cultural (collective) values are the widely shared premises about what individuals and societies should aspire for—fundamental goals that members are encouraged to pursue30,31,32,33. Cultural values give rise to practices and norms, which serve as standards for expected behavior. Personal values are abstract and desired trans-situational goals that serve as guiding principles in people’s lives34,35,36. They motivate behavior both directly and through their impact on attitudes and perceptions37,38,39.
Past research has investigated the associations between values and the perceived quality of marital relationships. At the individual level, the majority of studies found that spousal similarity in personal values predicts relationship satisfaction40,41,42,43, cf.44. Only a handful of studies have examined whether the content of values predicts divorce, perhaps because values, which carry a highly positive valence25, seem irrelevant to a suboptimal outcome, such as divorce. In the few studies conducted, marital satisfaction was positively associated with collectivism (measured at both the individual and cultural levels45,46). The current research is the first to systematically examine the full range of both personal and cultural values as determinants of the dissolution of marriage.
To study cultural values of nations we draw on Schwartz’s framework32,47,48, which is the only framework that conceptualizes and measures both cultural and personal values. Schwartz identified three basic conflicts of cultural values: autonomy vs. embeddedness, egalitarianism vs. hierarchy, and harmony vs. mastery. Most relevant to the current investigation is the first conflict, which concerns the relationships between the individual and the group. In cultures high in embeddedness values, members are viewed as pieces in a big mosaic, expected to blend in the collective, identify with it, and adopt its goals. By contrast, in cultures high in affective or intellectual autonomy values, members are viewed as autonomous entities, encouraged to set their own goals and express uniqueness. Divorce predominantly promotes individuals’ goals, letting them a way out of unsatisfactory relationships even at a potential cost to the social order. We therefore hypothesized that the higher the cultural embeddedness (vs. autonomy), the more negative the attitudes towards divorce and the lower the divorce rate.
Schwartz (1992)26 further identified ten basic personal values (see Fig. 1) that could be summarized into two value conflicts: conservation versus openness to change, and self-enhancement versus self-transcendence. The former is most relevant to the current research. Conservation values (security, conformity, and tradition) express the motivation of deference to and devoutness toward abstract authorities and familiar customs (tradition); adherence to social norms and expectations along with the exercise of self-restriction (conformity); and stability, order, and the safety of close others (security). Divorce hinders the attainment of each of these values. First, divorce is religiously disputed if not prohibited altogether. Divorce is also a non-conformist decision that violates a societal norm. Lastly, divorce poses emotional and financial threats to the security of the self and family. We therefore hypothesize that emphasizing tradition, conformity, and security values is negatively related to divorce.
Schwartz’s ten basic personalized values are organized along two main conflicts: Self-transcendence values (in pink) are opposed to self-enhancement values (in purple) and openness-to-change values (in green) are opposed to conservation values (in orange).
Openness-to-change values (self-direction, stimulation, and hedonism) express the motivation of independence of thought and action (self-direction); novelty and excitement in life (stimulation); and self-indulgence and pleasure (hedonism). Divorce may allow the expression of all three values. Some relationships undermine the ability to pursue an individual’s goals and aspirations. Divorce may then allow to exercise independence. Divorce further paves a path to novel and exciting new relationships and experiences. Finally, divorce may offer an opportunity for more pleasurable relationships than one currently has. We therefore hypothesize that emphasizing self-direction, stimulation, and hedonism values is positively related to divorce.
The second value conflict contrasts the motivation to care for other people and for nature (benevolence and universalism) with the motivation to promote oneself and achieve dominance and control over others (achievement and power). We reason that values on each pole of this conflict could motivate individuals either to seek divorce or to avoid it. For example, a person who emphasizes self-transcendence is likely to prioritize the interests of their child above their own interest. That person may choose to divorce to shield their child from a high-stress relationship, or they may choose to remain married to protect the child from the distress of parental separation. For another example, a person who emphasizes self-enhancement values is likely to focus on their own interests; consequently, they might neglect their partner’s needs, precipitate a divorce, or they might cling to their marriage to avoid declining social status. Thus, despite their strong relevance to interpersonal relationships, the importance ascribed to self-transcendence versus self-enhancement values is not likely to predict divorce.
Building on recent theorizing39,49,50, we reason that the impact of personal values on divorce varies across cultures. In autonomous cultures, individuals are encouraged to express their individuality—their tastes, ideas, aspirations, and values. In embeddedness cultures, by contrast, individuals are expected to identify with the collective interest and obey social norms, to the extent that violating a norm might result in a direct punishment51. Hence, members of autonomous cultures are assumed to act on their personal values more than members of embedded cultures. We thus hypothesize that the magnitude of associations between personal values and divorce is bigger in cultures emphasizing autonomy (vs. embeddedness) values.
We tested our hypotheses in three studies, using archival datasets. Study 1 investigated the relationships between cultural values and the national divorce-to-marriage ratio. Study 2 investigated the unique effects of cultural values, personal values, and the interaction between them. Study 3 focused on personal values, identifying the values of people who have never been divorced, are currently divorced, or were divorced.
Methods
This research draws on publicly available data collected in accordance with the ethical standards of the WVS and the ESS which include providing informed consent. All analyses conducted for this research were approved by the researchers’ Institutional Review Board (approval no. 2023_05_017). All sample sizes were determined solely by the availability of the data. We did not exclude any participants unless their data were insufficient to be included in the analyses (i.e., there were missing data in either the independent or the dependent variables).
Study 1
Study 1 examined the relationships between cultural values and national divorce rates. We matched two independent datasets: one containing cultural values scores as defined by Schwartz52, and another containing information on national divorce-to-marriage ratios between the years 1920–2018 (retrieved from Our World in Data, see https://ourworldindata.org/marriages-and-divorces). The 59 countries for which both cultural values and divorce data were available served as the data points in the analysis (see full list in the Supplementary Table 1).
Cultural values
Data on nations’ embeddedness vs. autonomy values were derived from Schwartz52. Schwartz computed national scores of 80 countries using data gathered from 1988–2007. Research on many countries indicates that change in cultural value orientations is very slow even in the presence of major political and institutional change52. The embeddedness vs. autonomy composite was calculated by subtracting the average of intellectual and affective autonomy scores (based on the items: broadminded, creativity, curious, freedom, enjoying life, exciting life, pleasure, varied life, self-indulgent) from the embeddedness score (based on the items: clean, devout, forgiving, honoring parents and elders, moderate, national security, obedient, politeness, protecting my public image, reciprocation of favors, respect for tradition, self-discipline, social order, wisdom). The hierarchy vs. egalitarianism composite was calculated by subtracting the score for egalitarianism (based on the items: equality, helpful, honest, loyal, responsible, social justice) from the score for hierarchy (based on the items: authority, humble, social power, wealth). The mastery vs. harmony composite was calculated by subtracting nations’ harmony score (based on the items: a world of beauty, a world at peace, protecting the environment, unity with nature) from nations’ mastery score (based on the items: ambitious, capable, choosing own goals, daring, independent, influential, social recognition, successful).
Divorce-to-marriage ratio
The initial public data included separate information about crude divorce and marriage rates by country and year. These rates denote the number of divorces/marriages per 1000 people in the country in a specific year. In order to compute divorce-to-marriage ratios, we matched the two lists of rates, while excluding incomplete observations. We then divided divorce by marriage to obtain the ratio of interest. The number of years for which the divorce-to-marriage ratio was available varied across countries: Several countries, such as Yemen, only had a single report of divorce and marriage rates, whereas other countries had reports over many years (the USA, for example, had reports for 99 years). For each country, we created two ratio indexes: The first was an average of all divorce-to-marriage ratios available for the country over the years. This index was used in the main analyses. The second was the most recent divorce-to-marriage ratio available for the country. When we analyzed the data using this index, the predicted effect of embeddedness (vs. autonomy) was slightly stronger than the effect for the average index that we report above.
Study 2
In Study 2 we tested our full model, investigating the unique effects of cultural and personal values and the interaction between them. Individual-level data were obtained from Wave 5 and 6 of the World Value Survey (WVS, www.worldvaluessurvey.org/wvs.jsp53,54. The WVS consists of nationally representative surveys conducted in all of the world’s major cultural zones using a common questionnaire. The questionnaire includes dozens of questions about human beliefs, attitudes, and values. Information about personal values, as defined by Schwartz26, was collected only in these two waves of data collection. Wave 5 took place between 2005 and 2009 whereas Wave 6 wave took place between 2010 and 2014. We included in the analysis the data of all respondents from the 46 nations that could be matched with national-level data on Schwartz’s cultural values (N = 86,436, 52.4% female, 47.6% male, Mage = 41.96, age range 16–99). Supplementary Table 2 presents demographic characteristics by nation.
Embeddedness vs. autonomy cultural values
See cultural values in Study 1.
Personal values
Participants’ personal values were measured with an abbreviated version of the Portrait Values Questionnaire (PVQ55). The PVQ gauges the ten values by describing hypothetical figures, and participants are asked to rate their similarity to each on a scale from 1 (very much like me) to 6 (not at all like me). For example, participants’ similarity rating to the figure described as “It is important to this person to have a good time; to “spoil” oneself” indicates their endorsement of hedonism values. The version used in the current study consisted of 10 items, one per value. For ease of interpretation, we reversed-coded all value items so that higher scores reflect a higher endorsement of the value. In line with the vast recommendation in values research, we centered each value on an individual basis26,55. During the administration of Wave 6, the item measuring benevolence was rephrased, with participants receiving either the old version, the new version, or both. We used the response to either version or the average rating (when both were available) as the indicator for benevolence values.
Marital status
To indicate their current marital status, participants chose among the following categories: married, living together as married, divorced, separated, widowed, or single (117 participants did not report their status). Because we were interested in how values help distinguish between individuals who leave versus stay in the marriage, we only included participants who reported being married or widowed (grouped together, coded 0) and those who reported being divorced or separated (grouped together, coded 1) in the analysis concerning marital status.
Justification of divorce
The participants were asked to indicate how justifiable divorce is on a scale from 1 (never justifiable) to 10 (always justifiable). Participants completing this question were included in the analyses predicting divorce justification regardless of their marital status.
Analytical strategy
We used Hierarchical Linear Modelling (HLM) to account for the dependency between individuals nested within the different nations. We used R packages lme4 (for the hierarchical modeling56) and reghelper (for the simple slopes57). In each hierarchical regression the index of cultural embeddedness vs. autonomy was entered as a cultural-level predictor, one personal value type was entered as an individual-level predictor, and their cross-level interaction was estimated. Overall, we employed ten linear models, with maximum likelihood optimization, to predict the justification of divorce (a continuous outcome), and another ten logistic models to predict marital status (a dichotomous outcome). The data distribution was assumed to meet the tests’ requirements, but this was not formally tested.
Study 3
To test the robustness of our findings, this study investigated again the impact of individual-level values. In addition, considering that divorced and married individuals carry different value profiles (as suggested by this research), we sought to further investigate an intriguing group of individuals: those who divorced and later remarried. Are remarried individuals more similar in their values to others who are currently divorced? Or are they more similar to individuals who have always been married? Understanding the pattern of similarities could help inform whether having the experience of divorce or the current bearing of marital status is more strongly related to people’s values.
The participants were 32,588 individuals (53.0% female, 47.0% male, Mage = 50.69, age range 15–90) taking part in the 9th round of the European Social Survey (ESS, www.europeansocialsurvey.org). The ESS is an academically driven survey, sampling representatively from various European nations. The survey is administered face-to-face, and measures attitudes, beliefs, and behavior patterns. The ninth round of the survey consists of data collected between 2018 and 2019 from 19 countries (Austria, Belgium, Bulgaria, Cyprus, Czechia, Estonia, Finland, France, Germany, Hungary, Ireland, Italy, Netherlands, Norway, Poland, Serbia, Slovenia, Switzerland, and the United Kingdom). Because European countries are relatively homogenous with respect to their autonomous culture, this study focused on investigating the effect of personal values only. Supplementary Table 3 presents demographic characteristics by nation.
Personal values
Personal values were measured as in Study 2 except that each value was assessed with 2–3 items (rather than a single item). As in Study 2, for ease of interpretation, we reversed-coded all value items so that higher scores reflect a higher endorsement of the value. Values indexes were computed by averaging the items reflecting each value type and centering on an individual basis.
Marital status in the past and present
Three questions concerned participants’ marital status indicating, in combination, their marital experience in the past and present. First, the participants were asked whether or not they are or have ever been married. We used this question to select only participants confirming marital experience to the analyses predicting staying vs. leaving marriage. Then, they were asked whether or not they had ever been divorced. We used this question as our primary dependent variable (coding No as 0 and Yes as 1). Finally, the participants indicated their current legal marital status by choosing from among the following options: legally married, in a legally registered civil union, legally separated, legally divorced/civil union dissolved, widowed/civil partner died, or none of these. We grouped the first two categories as Currently Married (coded 0) and the following two categories as Currently Divorced (coded 1). This question helped us distinguish between individuals who were never divorced, those who divorced and remarried, and those who divorced and have not remarried.
Approval of divorce while having young children
The participants were asked to indicate on a scale from 1 (strongly disapprove) to 5 (strongly approve), how much they approve or disapprove if a person gets divorced while they have children aged under 12. Participants who completed this question were included in the analyses predicting divorce approval regardless of their marital status.
Analytical Strategy
We used Hierarchical Linear Modelling (HLM) to account for the dependency between individuals nested within the different nations. We used R package lme456 to conduct ten hierarchical linear regressions predicting divorce approval (a continuous outcome) and another ten hierarchical logistic regressions predicting ever being divorced (a binary outcome). Each of the ten regressions used a different personal value as a predictor.
This research draws on publicly available data collected in accordance with the ethical standards of the WVS and the ESS, which include providing informed consent. All analyses conducted for this research were approved by the Institutional Review Board of the Hebrew University Business School (approval no. 2023_05_017). The studies were not pre-registered.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Results
Study 1
As hypothesized, greater emphasis on cultural embeddedness (over autonomy) predicted a lower divorce-to-marriage ratio in the country (b = -0.20, SE = 0.04, 95% CI [-0.28; -0.11], t = -4.72, p < 0.001), explaining 28% of the variance. Serving as single predictors in separate regressions, mastery (over harmony) and hierarchy (over egalitarianism) explained substantially smaller portions of the variance (mastery: b = -0.09, SE = 0.04, 95% CI [-0.18; -0.00], t = -2.01, p = 0.049, 7% explained variance; hierarchy: b = -0.05, SE = 0.03, 95% CI [-0.11; 0.002], t = -1.91, p = 0.061, 6% explained variance). When the three cultural dimensions were entered into the regression equation simultaneously, embeddedness (over autonomy) was the only predictor that remained significant (b = -0.23, SE = 0.05, 95% CI [-0.37; -0.13], t = -4.50, p < 0.001; see Table 1). In sum, the findings of Study 1 support our hypothesis, indicating that divorce is more prevalent in highly embedded (vs. autonomous) cultures.
Study 2
Table 2 presents the full regression results.
Cultural values
Replicating Study 1, and as hypothesized, the country’s embeddedness (vs. autonomy) values negatively predicted justification of divorce in all regression models, security: ɣ = -2.35, SE = 0.33, 95% CI [−3.00; −1.70], p < 0.001, conformity: ɣ = −2.30, SE = 0.33, 95% CI [−2.94; −1.65], p < 0.001, tradition: ɣ = −2.21, SE = 0.34, 95% CI [−2.87; −1.55], p < 0.001, benevolence: ɣ = −2.04, SE = 0.33, 95% CI [−2.70; −1.39], p < 0.001, universalism: ɣ = −2.22, SE = 0.33, 95% CI [−2.86; −1.57], p < 0.001, self-direction: ɣ = −2.31, SE = 0.32, 95% CI [−2.95; −1.68], p < 0.001, stimulation: ɣ = −2.55, SE = 0.32, 95% CI [−3.18; −1.92], p < 0.001, hedonism: ɣ = −2.25, SE = 0.34, 95% CI [−2.91; −1.59], p < 0.001, achievement: ɣ = −2.26, SE = 0.33, 95% CI [−2.91; −1.62], p < 0.001, power: ɣ = −2.29, SE = 0.34, 95% CI [−2.95; −1.63], p < 0.001. Likewise, the country’s embeddedness (vs. autonomy) values negatively predicted divorce rate, security: ɣ = −1.93, SE = 0.31, 95% CI [−2.54; −1.32], p < 0.001, conformity: ɣ = −1.86, SE = 0.31, 95% CI [−2.46; −1.25], p < 0.001, tradition: ɣ = −1.82, SE = 0.31, 95% CI [−2.43; −1.22], p < 0.001, benevolence: ɣ = −1.73, SE = 0.31, 95% CI [−2.33; −1.14], p < 0.001, universalism: ɣ = −1.81, SE = 0.30, 95% CI [−2.41; −1.20], p < 0.001, self-direction: ɣ = −1.81, SE = 0.30, 95% CI [−2.42; −1.20], p < 0.001, stimulation: ɣ = −1.92, SE = 0.32, 95% CI [−2.53; −1.30], p < 0.001, hedonism: ɣ = −1.82, SE = 0.31, 95% CI [−2.42; −1.21], p < 0.001, achievement: ɣ = −1.84, SE = 0.31, 95% CI [−2.43; −1.24], p < 0.001, power: ɣ = −1.84, SE = 0.32, 95% CI [−2.46; −1.21], p < 0.001.
Personal values
Also as hypothesized, emphasizing conformity and tradition values negatively predicted divorce justification, conformity: ɣ = −0.12, SE = 0.02, 95% CI [−0.16; −0.08], p < 0.001; tradition: ɣ = −0.25, SE = 0.03, 95% CI [−0.30; −0.19], p < 0.001, and the likelihood of being divorced, conformity: ɣ = −0.06, SE = 0.02, 95% CI [−0.10; −0.03], p = 0.001, tradition: ɣ = −0.16, SE = 0.02, 95% CI [−0.20; −0.16], p < 0.001, whereas emphasizing self-direction, stimulation, and hedonism values positively predicted divorce justification, self-direction: ɣ = 0.15, SE = 0.02, 95% CI [0.11; 0.18], p < 0.001; stimulation: ɣ = 0.09, SE = 0.02, 95% CI [0.06; 0.12], p < 0.001 & hedonism: ɣ = 0.17, SE = 0.02, 95% CI [0.13; 0.21], p < 0.001, and the likelihood of being divorced, self-direction: ɣ = 0.08, SE = 0.02, 95% CI [0.04; 0.11], p < 0.001; stimulation: ɣ = 0.11, SE = 0.02, 95% CI [0.08; 0.15], p < 0.001; hedonism: ɣ = 0.08, SE = 0.02, 95% CI [0.05; 0.12], p < 0.001. Our hypothesis was not confirmed for security values.
Cross-level interaction
Finally, the findings supported our interaction hypothesis (see Table 2). In nations high in autonomy values, both divorce justification and actual divorce were associated with all six values, including security, divorce justification: security: ɣ = −0.07, SE = 0.02, 95% CI [−0.12; −0.03], p = 0.004; conformity: ɣ = −0.15, SE = 0.02, 95% CI [−0.22; −0.11], p < 0.001; tradition: ɣ = −0.30, SE = 0.04, 95% CI [−0.37; −0.23], p < 0.001; self-direction: ɣ = 0.16, SE = 0.02, 95% CI [0.11; 0.20], p < 0.001; stimulation: ɣ = 0.14, SE = 0.02, 95% CI [0.10; 0.18], p < 0.001; hedonism: ɣ = 0.23, SE = 0.03, 95% CI [0.18; 0.28], p < 0.001; actual divorce: security: ɣ = −0.08, SE = 0.02, 95% CI [−0.11; −0.04], p < 0.001; conformity: ɣ = −0.10, SE = 0.02, 95% CI [−0.14; −0.06], p < 0.001; tradition: ɣ = −0.19, SE = 0.02, 95% CI [−0.23; −0.15], p <0.001; self-direction: ɣ = 0.08, SE = 0.02, 95% CI [0.05; 0.12], p < 0.001; stimulation: ɣ = 0.16, SE = 0.02, 95% CI [0.12; 0.20], p < 0.001; hedonism: ɣ = 0.14, SE = 0.02, 95% CI [0.10; 0.18], p < .001. In contrast, in nations high in embeddedness values, these associations were either weaker or non-significant, divorce justification: security: ɣ = 0.04, SE = 0.03, 95% CI [−0.01; 0.09], p = 0.145; conformity: ɣ = −0.06, SE = 0.03, 95% CI [−0.11; 0.00], p = 0.061; tradition: ɣ = −0.17, SE = 0.04, 95% CI [−0.25; −0.09], p = 0.000; self-direction: ɣ = 0.13, SE = 0.03, 95% CI [0.08; 0.18], p < 0.001; stimulation: ɣ = 0.02, SE = 0.02, 95% CI [−0.02; 0.07], p = .300; hedonism: ɣ = 0.09, SE = 0.03, 95% CI [0.03; 0.15], p = .003; actual divorce: security: ɣ = 0.05, SE = 0.03, 95% CI [−0.02; 0.11], p = 0.188; conformity: ɣ = −0.01, SE = 0.03, 95% CI [−0.08; 0.05], p = 0.678; tradition: ɣ = −0.12, SE = 0.03, 95% CI [−0.19; −0.05], p < 0.001; self-direction: ɣ = 0.07, SE = 0.03, 95% CI [0.01; 0.13], p = 0.030; stimulation: ɣ = 0.04, SE = 0.03, 95% CI [−0.01; 0.10], p = 0.102; hedonism: ɣ = 0.003, SE = 0.03, 95% CI [−0.05; 0.06], p = 0.905. The interaction effects were significant for all hypothesized values except self-direction, which predicted divorce positively regardless of cultural embeddedness, divorce justification: security: ɣ = 0.16, SE = 0.05, 95% CI [0.06; 0.25], p = 0.002; conformity: ɣ = 0.15, SE = 0.05, 95% CI [0.05; 0.26], p = 0.006; tradition: ɣ = 0.18, SE = 0.07, 95% CI [0.04; 0.33], p = 0.018; self-direction: ɣ = −0.04, SE = 0.05, 95% CI [−0.13; 0.06], p = 0.488; stimulation: ɣ = −0.17, SE = 0.04, 95% CI [−0.25; −0.09], p = 0.000; hedonism: ɣ = −0.20, SE = 0.05, 95% CI [−0.30; −0.09], p = 0.000; actual divorce: security: ɣ = 0.17, SE = 0.05, 95% CI [0.06; 0.28], p = 0.002; conformity: ɣ = 0.12, SE = 0.05, 95% CI [0.02; 0.22], p = 0.022; tradition: ɣ = 0.10, SE = 0.06, 95% CI [−0.01; 0.21], p = 0.067; self-direction: ɣ = −0.03, SE = 0.05, 95% CI [−0.12; 0.07], p = 0.580; stimulation: ɣ = −0.16, SE = 0.05, 95% CI [−0.26; −0.07], p = 0.001; hedonism: ɣ = −0.20, SE = 0.05, 95% CI [−0.29; −0.10], p = 0.000. Self-direction values express the motivation for independence of thought and action. Emphasizing these values may thus lead people to act on their personal values even in cultures that emphasize embeddedness values.
To test the robustness of our findings, we repeated the analyses while controlling for various demographic measures at both the national and individual levels (national level: gross domestic product per capita, median age, average religiosity; individual level: sex, age, education, income, number of children, religiosity, technology usage). The results obtained were highly similar (see detailed analysis in Supplementary Methods 1 and Supplementary Table 4).
We further calculated the proportion of explained variance (R2) as an indicator of overall association strength between personal values and justification of divorce. Correlating this proportion with cultural embeddedness (vs. autonomy) indicated that, as expected, cultural embeddedness negatively predicted the explanatory power of personal values in divorce justification (r = −0.48, 95% CI [−0.68; −0.22], p < 0.001; see Fig. 2).
The dark blue line represents a regression line and the shade represents the 95% confidence interval for the slope. Note that the Ethiopian sample was a substantial outlier and was hence excluded from this analysis.
Study 3
Table 3 presents the results from the regression analyses. As hypothesized, security, conformity, and tradition values negatively predicted justification of divorce, security: ɣ = −0.12, SE = 0.02, 95% CI [−0.16; −0.09], p < 0.001; conformity: ɣ = −0.18, SE = 0.02, 95% [−0.21; −0.14], p < 0.001; tradition: ɣ = −0.20, SE = 0.02, 95% [−0.24; −0.16], p < 0.001, and the likelihood of experiencing divorce in the past or present, security: ɣ = −0.10, SE = 0.04, 95% CI [−0.18; −0.03], p = 0.006, conformity: ɣ = −0.19, SE = 0.02, 95% CI [−0.24; −0.14], p < 0.001; tradition: ɣ = −0.28, SE = 0.04, 95% [−0.35; −0.20], p < 0.001, whereas self-direction, stimulation, and hedonism values predicted stronger approval of divorce, self-direction: ɣ = 0.14, SE = 0.02, 95% CI [0.11; 0.18], p < 0.001; stimulation: ɣ = 0.12, SE = 0.01, 95% CI [0.09; 0.14], p < 0.001; hedonism: ɣ = 0.14, SE = 0.02, 95% CI [0.11; 0.17], p < 0.001, and a higher likelihood of experiencing it, self-direction: ɣ = 0.31, SE = 0.03, 95% CI [0.25; 0.37], p < 0.001; stimulation: ɣ = 0.20, SE = 0.03, 95% CI [0.14; 0.25], p < 0.001; hedonism: ɣ = 0.14, SE = 0.02, 95% CI [0.10; 0.18], p < 0.001. Controlling for various demographic variables (sex, age, education, having children, religiosity, domicile, and political orientation) did not change the results (see Supplementary Methods 1 and Supplementary Table 5).
Finally, we compared the value profiles of re-married individuals (n = 1643) to those of always-married individuals (n = 15,399) and currently divorced individuals (n = 3286). Pairwise comparisons using Tukey’s HSD confirmed that all differences between re-married and currently divorced individuals were non-significant, security: difference = 0.04, 95% CI [−0.02; 0.09], p = 0.302; conformity: difference = 0.01, 95% CI [−0.06; 0.07], p = 0.977; tradition: difference = −0.01, 95% CI [−0.07; 0.05], p = 0.950; benevolence: difference = −0.03, 95% CI [−0.07; 0.02], p = 0.383; universalism: difference = −0.02, 95% CI [−0.06; 0.03], p = 0.609; self-direction: difference = 0.02, 95% CI [−0.04; 0.07], p = 0.736; stimulation: difference = 0.01, 95% CI [−0.06; 0.08], p = 0.944; hedonism: difference = 0.02, 95% CI [−0.05; 0.09], p = 0.783; achievement: difference = −0.05, 95% CI [−0.11; 0.02], p = 0.224; power: difference = 0.02, 95% CI [−0.04; 0.09], p = 0.614), whereas all differences between re-married and always-married individuals were significant (security: difference = −0.08, 95% CI [−0.13; −0.03], p < 0.001; conformity: difference = −0.15, 95% CI [−0.21; −0.10], p < .001; tradition: difference = −0.20, 95% CI [−0.26; −0.15], p < .001; benevolence: difference = 0.06, 95% CI [0.02; 0.10], p = 0.002; universalism: difference = 0.05, 95% CI [0.02; 0.09], p = 0.003; self-direction: difference = 0.20, 95% CI [0.16; 0.25], p < 0.001; stimulation: difference = 0.17, 95% CI [0.11; 0.23], p < 0.001; hedonism: difference = 0.17, 95% CI [0.11; 0.23], p < 0.001; achievement: difference = −0.11, 95% CI [−0.17; −0.05], p < 0.001; power: difference = −0.14, 95% CI [−0.20; −0.09], p < 0.001; see Table 4. Note, however, that differences in groups’ sizes could affect the significance tests. We therefore computed two series of standardized differences (Cohen’s d), comparing the values of re-married individuals to each of the other two groups. For all ten values, the standardized differences between re-married and always-married individuals were consistently different from zero, security: Cohen’s d = −0.10, 95% CI [−0.15; −0.05]; conformity: Cohen’s d = −0.17, 95% CI [−0.22; −0.12]; tradition: Cohen’s d = −0.24, 95% CI [−0.29; −0.19]; benevolence: Cohen’s d = 0.09, 95% CI [0.04; 0.14]; universalism: Cohen’s d = 0.09, 95% CI [0.04; 0.14]; self-direction: Cohen’s d = 0.26, 95% CI [0.21; 0.31]; stimulation: Cohen’s d = 0.18, 95% CI [0.13; 0.23]; hedonism: Cohen’s d = 0.17, 95% CI [0.12; 0.23]; achievement: Cohen’s d = −0.12, 95% CI [−0.17; −0.07]; power: Cohen’s d = −0.16, 95% CI [−0.21; −0.11]; see Table 4, whereas the standardized differences between re-married and currently divorced individuals were consistently close to zero, security: Cohen’s d = −0.04, 95% CI [−0.10; 0.02]; conformity: Cohen’s d = −0.01, 95% CI [−0.06; 0.05]; tradition: Cohen’s d = 0.01, 95% CI [−0.05; 0.07]; benevolence: Cohen’s d = 0.04, 95% CI [−0.02; 0.10]; universalism: Cohen’s d = 0.03, 95% CI [−0.03; 0.09]; self-direction: Cohen’s d = −0.02, 95% CI [−0.08; 0.04]; stimulation: Cohen’s d = −0.01, 95% CI [−0.07; 0.05]; hedonism: Cohen’s d = −0.02, 95% CI [−0.08; 0.04]; achievement: Cohen’s d = 0.05, 95% CI [−0.01; 0.11]; power: Cohen’s d = −0.03, 95% CI [−0.09; 0.03]; see Table 4.
Discussion
We investigated the relationships between values and divorce in three large-scale studies. Taken together, the results lend support to our hypotheses. Investigating cultural values revealed that the more cultures emphasize embeddedness (versus autonomy) values, the less their members legitimize divorce and the less are likely to divorce (Studies 1 & 2). Investigating personal values revealed that tradition and conformity discourage people from supporting divorce and getting divorced, whereas self-direction, stimulation, and hedonism encourage both (Studies 2 & 3). In addition, cultural and personal values interact in predicting divorce, such that personal values predict individuals’ outcomes more powerfully in nations that emphasize autonomy (Study 2). Lastly, we found that re-married individuals were more similar in their values to divorced than married individuals (Study 3). The findings are consistent for attitudes (divorce justification/approval) and behavior (actual divorce).
Previous studies have listed a range of concrete and proximal motives for divorce, such as poor relationship quality and a weak commitment to marriage58, problems related to work and the division of labor59, infidelity, incompatibility, substance use, and growing apart60, money-wasting, jealousy, moodiness, and irritating habits61, see also62. In this research we broaden the investigation to distal, abstract, general, and cross-situational motivations (i.e., values) that conceptually differ from such specific motives. Our investigation of personal values thus deepens our understanding of the motivations underlying divorce and its psychological meaning.
A great deal of evidence supports the importance of self-transcendence vs. self-enhancement values in pro- and anti-social behavior (see review in ref. 28). Recently, van der Wal et al. 44 showed that people higher in self-transcendence values enjoy better relationship quality. Despite their impact on interpersonal relationships, we reasoned and found them unrelated to divorce. Instead, divorce is primarily associated with values related to conservation and openness values. Using Lönnqvist et al. 63 terminology, divorce is “value-expressive” with regard to the latter value conflict, as it enables the expression of one (openness) and not the other (conservation). However, divorce is “value ambivalent” with respect to self-transcendence vs. self-enhancement, as it can express either one or the other.
The findings were inconsistent with regard to security values. We expected security to be negatively related to divorce because it promotes a safe social and financial life. It is possible, however, that divorce is sometimes initiated in order to gain family or financial security. When the marriage relationship is perceived as insecure or involves risk to children, stepping out of the marriage may reduce insecurity. Thus, the motivation for security may lead not only to sustaining the marriage but also to breaking it.
Studying cultural values in addition to personal values allowed us to investigate not only the impact of motivations (i.e., personal values) but also that of norms (i.e., cultural values), and the interplay between the two. Past studies often assumed the relationships between divorce and a cultural emphasis on individualism20,21,22. Across two large and diverse datasets we found that divorce is consistently related to a cultural emphasis on autonomy, which is conceptually and empirically related to individualism64. Relying on a comprehensive theory of cultural values allowed us to investigate whether any other cultural dimension is related to divorce. Our studies also revealed negative associations between divorce and hierarchy (vs. egalitarianism) as well as mastery (vs. harmony) cultural values. These associations turned weak and inconsistent, however, when controlling for embeddedness (vs. autonomy values). It is possible that the zero-order associations with hierarchy and mastery are driven by their own relationships with embeddedness (vs. autonomy) values32.
The current study also contributes to the research on values. First, we investigate both attitudes toward divorce and actual divorce, showing consistent associations across outcome types. This finding therefore speaks to the ongoing discussion about the impact of values, strengthening the claim that values are associated not only with subjective attitudes but also with actual behavior, e.g.,38,39,65. Moreover, this study joins an emerging body of research that simultaneously tests the effects of personal and cultural values and the interplay between them. This examination is important as it points to an interactive effect of the personal and cultural levels and helps identify the boundaries of the values effect rather than their mere existence.
Limitations
Divorce is a dyadic decision made by two interdependent individuals. Our findings indicate that knowing even one spouse’s values is beneficial to predicting whether a marriage will fail or sustain. The impact of one’s values may depend on their partner’s values, however. Future research may focus on dyads to investigate the interactive effect of both spouses’ values. Another limitation of our study is that our measurement of divorce indicates the respondent’s marital status at the time of taking the survey. This status may change in the future (e.g., a respondent might get divorced after their participation). This may suggest that the current study is likely to underestimate the impact of values on marriage dissolution.
This study observes a relationship between openness (vs. conservation) and autonomy (vs. embeddedness) values and divorce. Our theorizing focuses on the tradeoff between collective and individual interests at both the societal and personal levels, suggesting that the collective benefits from the institution of marriage while the individual benefits from the possibility of divorce. While this perception is common, it is not necessarily accurate: High conflict, misfunctioning marriages might be harmful to both person and society, whereas harmonious separation may benefit both. Indeed, values’ impact on attitudes and behavior depends on their subjective perception (see review in ref. 39). It is possible, then, that the observed association results from community-based perception and stigma rather than the objective reality.
We reason that the notion of divorce—the idea that people can step out of a marriage suppose they are willing to—favors individuals’ freedom over commitment. However, divorce does not come without its costs, and for some groups of individuals, it might result in reduced freedom. For example, studies show that women’s income decreased66 and chances of falling under the poverty line increased67 post-divorce, probably leading to impaired personal freedoms in many life domains. Future studies could investigate perceptions of divorce among disadvantaged groups such as women.
Lastly, our research focused on the impact of values on divorce. Future research could investigate the role of values in marriage. Over the last century, the motives for marriage have changed—from pragmatism to love to self-expression as of today68. Values could inform us as to the motivations underlying the decision to marry and account for cross-cultural variation in this behavior.
Conclusion
Divorce is a possible solution to an undesirable situation. This study suggests that whether or not one will opt for this solution depends, in part, on their personal and cultural values. Some individuals (likely, those who prioritize conservation) would rather avoid divorce at all possible costs. Other individuals (likely those who prioritize openness to change) are open to considering divorce when push comes to shove. This way or the other, marrying spouses who share one’s own values elevates the chances that they also share their worldview on the matter of divorce. Know your partner’s values before your wedding celebration.
Responses