Abstract
Do timing attitudes—subjective evaluations of particular ages as good ages to marry—influence entrance into marriage? To address this question, we formulated an intergenerational model of how parents’ and children’s timing attitudes influence children’s marriage behavior. We theorized that both parents’ and children’s timing attitudes influence expectations of when children will marry. In turn, both parents’ and children’s marital expectations would influence children’s actual entrance into marriage. We tested the model using intergenerational panel data from Nepal collected in 2008–2014. Timing attitudes of young people and their parents did influence expectations, as well as entrance into marriage. Young people’s own attitudes were more influential than their parents’ attitudes in determining children’s expectations, but not behavior. Further, while the influence of parents was relatively even, mothers appear slightly more influential than fathers.
Keywords: Marriage, Age at marriage, Ideal age, Attitudes, Intergenerational, Nepal, Asia
Introduction
Ideational factors are increasingly recognized as an important influence on family behavior (Johnson-Hanks et al. 2011; Bachrach 2014; Thornton 2005; Lesthaeghe 2010). These ideational factors, which include schemas, values, attitudes, beliefs, scripts, and norms, provide ways for people to understand the world and live in it (Swidler 2001; Johnson-Hanks et al. 2011). One type are timing attitudes—subjective evaluations of particular ages as good or preferred ages, or times, to engage in a particular behavior. A few studies have assessed the influence of timing attitudes on family behaviors other than marriage (Billari and Liefbroer 2007; Tosi 2017; Van Bavel and Nitsche 2013). Further, a growing body of research has examined the influence of many ideational factors other than timing attitudes on marriage (Clarkberg et al. 1995; Jennings et al. 2012; Goldscheider et al. 2009). However, only a handful of studies touch on the connections between marital timing attitudes and marriage simultaneously. One contribution of this paper is in directly examining the influence of timing attitudes on the entrance to marriage using panel data from Nepal.
Another key contribution is our formulation of an intergenerational approach that examines the influence of individuals’ own attitudes, as well as those of both parents. Intergenerational influences on marriage are important even in the USA, where the individualized couple is often idealized (Thornton et al. 2008), but they may be even more important in Nepal and many other contexts where the family plays a more powerful role. In the past, parents and other relatives chose children’s spouses in much of Asia and determined when young people married. Young people in recent cohorts are more involved in making decisions about when and whom they marry, but parents are often still closely involved (Ghimire et al. 2006; Thornton and Lin 1994; Allendorf and Pandian 2016). Thus, we address not only whether timing attitudes are influential, but also whose attitudes are most influential. We do not, however, treat mothers and fathers as a unit, but as individuals with their own attitudes and influence.
Our focus on marriage is motivated by the centrality of marriage to individuals’ lives. Consequences of marital timing can be wide-ranging and diverse in Nepal and other non-Western contexts in which many marry in early adolescence. Delaying marriage by just a few years—from, say, age 15 to 18—can provide adolescents freedom to complete secondary schooling, work for pay, and expand social networks (Yount et al. 2016; Mensch et al. 2005; Mahmud and Amin 2006). For example, Field and Ambrus (2008) find that each year of marital delay translates into 0.22 additional years of girls’ schooling in Bangladesh. Since marriage is often closely followed by childbearing in these contexts, delaying marriage by just a few years also has important health ramifications. Women who give birth at very young ages are at higher risk of mortality, pregnancy complications, and other adverse health outcomes for both themselves and their children (Santhya 2011; Efevbera et al. 2017; Chari et al. 2017). Delaying marriage can also come with disadvantages, including poorer quality spouses, higher dowry costs, and damaged reputations (Hahn et al. 2018; Andrist et al. 2014). In Western contexts, where nearly all marriages are comparatively late, variation in marital timing may also have important consequences for health (Waite and Gallagher 2000; Waite and Lehrer 2003; Kim and McKenry 2002) and accumulation of resources (Killewald and Gough 2013; Yamokoski and Keister 2006; Addo 2014). It is important to note though that it is notoriously difficult to untangle causal effects of marital timing from selection bias (Ludwig and Bruderl 2018; Field and Ambrus 2008).
It is also important to note that the centrality of marriage to union formation varies across contexts. While marriage remains important in Europe and North America, its central roles in organizing society have declined substantially in recent decades—what Cherlin (2004) called “the deinstitutionalization of marriage.” One important aspect of this deinstitutionalization is the growing practice of cohabitation, which emerged as an alternative to marriage. In turn, research on Western contexts often addresses marriage and cohabitation together. In most other parts of the world, including Nepal, marriage remains central to human organization and well-being, although with reductions in importance in some places. In Nepal, cohabitation is not a widespread alternative and marriage remains nearly universal (Yabiku 2005; Ministry of Health [Nepal] et al. 2017; Caltabiano and Castiglioni 2008). Thus, our study addresses only marriage.
Background
As noted above, there is increased attention to ideational factors in studies of family behavior in recent years (Bachrach 2014; Van de Kaa 1987; Lesthaeghe 2010; Johnson-Hanks et al. 2011). Much empirical work in this area focuses on gender ideology. A collection of American studies examine whether attitudes and beliefs about women being homemakers and men being breadwinners, as well as gender equality, influence entrance into marriage (Barber and Axinn 1998; Koball 2004; Sassler and Schoen 1999; Sassler and Goldscheider 2004). Other studies assess the influence of individuals’ valuation of career, family, and leisure on marriage (Barber 2004; Clarkberg et al. 1995). For example, Clarkberg et al. (1995) assess the influence of attitudes about the importance of work, leisure time, and money on timing of union formation and the choice between marriage and cohabitation in the USA. Others address family-related schemas. Sri Lankan women who expect to choose their own spouse, and actually do so, marry earlier (Malhotra and Tsui 1996), while Nepali men’s marital timing is influenced by their parents’ attitudes about family size and parental support in old age (Jennings et al. 2012). Religious affiliation and religiosity also influence marriage (Lehrer 2004; Thornton et al. 1992).
Only a handful of studies address potential links between timing attitudes and entrance into marriage. Two such studies point to the influence of marital timing attitudes at collective levels. While exploring differences in the influence of neighbors’ versus neighborhood context on marriage in Nepal, Yabiku (2006) found that neighbors’ average ideal age at marriage influenced young people’s marital timing. And in Indonesia, ethnicity-based norms about appropriate ages at marriage are related to actual age at marriage (Buttenheim and Nobles 2009).
Other studies address timing attitudes at the individual level, but are focused on other, related research questions. While investigating the influence of parental resources on children’s marital timing in the USA, Axinn and Thornton (1992) found that mothers’ older ideal ages at marriage delayed their children’s marital timing. Similarly, while addressing why a significant minority of Americans continue to marry at a young age, Uecker and Stokes (2008) found that one important predictor is having parents who also married at a young age. They speculated that the transmission of preferences for an early age at marriage links these behaviors.
Previously, we used the same data from Nepal used here to assess if timing attitudes explain why women marry earlier than men and if there are gender differences in the influence of timing attitudes (Allendorf et al. 2017). We found that timing attitudes explain a third of the gender gap in marital timing and the influence of timing attitudes was greater for women. This analysis complements our previous article by taking a step back and focusing on whether young people’s and their parents’ timing attitudes do influence marriage. Unlike the earlier article on gender differences, we provide a comprehensive treatment of links between timing attitudes and marriage by theorizing an intergenerational model of how timing attitudes influence children’s entrance into marriage and empirically testing multiple pathways from the model. Our approach integrates the role of expectations and uses a more comprehensive set of timing attitudes measures. We also examine whose timing attitudes are more influential—young people’s own timing attitudes, those of their mothers, or those of their fathers.
The small number of previous studies on the influence of timing attitudes on marriage is surprising. Life course research has a strong tradition of exploring the timing of key life transitions, conceptualized as age norms and deadlines (Elder 1994; Settersten and Mayer 1997; Liefbroer and Billari 2010; Elder et al. 2015). Age norms and deadlines include both collective and individual perceptions of desirable ages for life events, including marriage, and are comparable to the marital timing attitudes we examine here. However, studies in this tradition focus largely on identifying and describing age norms (Settersten 1997, 1998; Settersten and Hägestad 1996). De Valk and Liefbroer (2007) go a step farther by examining the intergenerational transmission of timing preferences for marriage and other life course transitions in the Netherlands. They find that parents’ marriage timing preferences are strongly associated with their children’s timing preferences.
Only a handful of these studies take the next step of examining the influence of age norms on behavior and, to our knowledge, none focus on marriage. Most notably, Tosi (2017) and Billari and Liefbroer (2007) find young people’s age norms about home-leaving influenced their timing of leaving their parental homes in Italy and the Netherlands. They further found that the age norms of youth’s parents also influenced when the youth left home. These studies are especially relevant because marriage (or cohabitation) was one of the reasons, among others, that young people left the parental home. Ideal ages at motherhood also account for a substantial amount of variation in childbearing timing across Europe (Van Bavel and Nitsche 2013).
Conceptualization, Theory, and Hypotheses
We focus on marital timing attitudes, which are subjective evaluations of particular ages, or times, as desirable or appropriate ages to marry. Our theoretical model of how these timing attitudes influence marital behavior is shown in Fig. 1.
Fig. 1.
Conceptual model of how timing attitudes influence marriage behavior
We begin our description with the intergenerational link between children’s and parents’ timing attitudes. We conceptualize children’s timing attitudes as being shaped by parents’ timing attitudes through the process of socialization (Arrow 1, Fig. 1). Through direct instruction and modeling of behavior, parents teach children how marriage works, including who, when, and how to marry. There is a long tradition of studies documenting that children’s attitudes resemble parents’ attitudes (Bandura 1986; Chodorow 1978; Axinn and Thornton 1993; Axinn et al. 1994; Dhar et al. 2019). To our knowledge, only De Valk and Liefbroer (2007) previously documented transmission of marital timing attitudes specifically.
Hypothesis 1
Children are more likely to value older ages at marriage if their parents value older ages at marriage.
Next, we turn to expectations, which are a key component linking attitudes to behavior (Fishbein and Ajzen 2010; Billari and Liefbroer 2007). We expect that children’s timing attitudes shape their expectations of when they will marry (Arrow 2, Fig. 1). For example, if a 20-year-old woman believes it is acceptable for women to marry once they are 18 years old, she may expect to marry soon. To our knowledge, no previous studies assess whether timing attitudes shape such expectations.
Hypothesis 2
Children who value older ages at marriage are less likely to expect to marry soon.
We expect that parents’ timing attitudes also shape children’s expectations of when they will marry, net of children’s own timing attitudes (Arrow 3, Fig. 1). For example, a young woman may expect to marry at age 16 because her parents believe that is the appropriate age for women to marry, even though she herself thinks women are only ready to marry at age 18. Children may expect to marry in accordance with parents’ attitudes, rather than their own attitudes, to please parents or avoid embarrassing them in the community and among other kin. Further, given the history and continued existence of arranged marriage in this context, it is well accepted that marriage can happen at the behest of parents. Among those married in 2008–2014, parents alone chose 23% of spouses and parents were involved in the choice for an additional 31% of marriages (Allendorf 2017). Six decades earlier, parents alone chose spouses for more than 90% of those married in 1946–1955 in Chitwan Valley (Axinn et al. 2008).
Hypothesis 3
Children are less likely to expect to marry soon when their parents value older ages at marriage, net of children’s own timing attitudes.
Our overarching hypothesis is that timing attitudes influence entrance into marriage. We conceptualize the overarching link between children’s timing attitudes and marriage as operating through two pathways. First, we expect children’s expectations, which are initially shaped by timing attitudes (Hypothesis 2; Arrow 2, Fig. 1), to influence when children actually do marry (Arrow 4, Fig. 1). For example, upon reaching—or getting close to reaching—age 18, her valued age, a young woman will be on the lookout for a husband and ready to accept a proposal. In turn, she will likely marry only after—or near—attaining 18 years of age. A handful of previous studies document the influence of expectations on entrance into marriage (Liefbroer et al. 1994; Guzzo 2009; Waller and McLanahan 2005).
Hypothesis 4
Children who expect to marry soon will marry more quickly.
Children’s timing attitudes may not always be mediated by expectations though. Thus, another pathway directly connects children’s timing attitudes to marriage (Arrow 5, Fig. 1). For example, a young woman, who believes it is acceptable for women in general to marry after reaching age 18, aspires to complete an advanced degree and does not expect to marry herself until she has completed her education at age 23. However, at age 19 she receives a proposal and deems it such an excellent match she cannot turn it down. Conversely, had it been too early to marry according to her timing attitudes, say she was only 16 years of age, she would have likely rejected the proposal. To our knowledge, no previous study examines this link for marriage, although Tosi (2017) and Billari and Liefbroer (2007) previously found that timing attitudes influence parental home-leaving, which did include leaving to marry. The latter also theorize that expectations mediate this link, but identify the total effect of timing attitudes on home-leaving, which may include other pathways.
Hypothesis 5
Children who value older ages at marriage will marry later, net of expectations of when they will marry.
Next, we describe how parents’ timing attitudes, expectations, and behavior may influence when children marry, net of children’s attitudes and expectations (Arrows 6, 7, and 8, Fig. 1). We begin with the idea that just as children’s timing attitudes shape their expectations of when they will marry, parents’ timing attitudes shape their expectations for the timing of their children’s marriages. This influence of parental timing attitudes on expectations for when their children marry is shown by Arrow 6 of Fig. 1. To our knowledge, no previous studies assess whether parental timing attitudes shape parental expectations.
Hypothesis 6
Parents who value older ages at marriage are less likely to expect their children to marry soon.
As the custom of arranged marriage implies, children’s marriages involve the behavior of both children and parents. As portrayed in Arrows 7 and 8 in Fig. 1, parents can exercise social control over children and persuade or force them to behave in ways that are consistent with parent’s attitudes and expectations, rather than the child’s own attitudes and expectations. Parents can directly control when children marry through allocation of material resources, such as the provision of a home or education, and the use or threat of severe disapproval, violence, and abandonment (Grover 2011; Mody 2008; Chowdhry 2007). Thus, we emphasize that our conceptualization of intergenerational goes beyond ideational transmissions from parents to children. We expect that part of the intergenerational effects on marriage stems from parents’ attitudes and expectations guiding parents’ own within generation behavior that shapes when their children marry. There is some previous support for the influence of parents’ timing attitudes on children’s marital timing. Axinn and Thornton (1992) found that mothers’ timing attitudes influence children’s marital timing. Tosi (2017) and Billari and Liefbroer (2007) showed that children’s perceptions of their parents’ timing attitudes influence home-leaving.
Hypothesis 7
Parents who expect their children to marry soon will have children who marry more quickly.
Hypothesis 8
Children marry later when their parents value older ages at marriage, net of parents’ expectations and children’s own timing attitudes and expectations.
The description of our theoretical model implies that timing attitudes may be best conceptualized as a range. Rather than viewing only a single, specific age as a good time for marriage, it is likely that people instead view a range of ages as a desirable period to marry (Liefbroer and Billari 2010). That is, people may have minimum and maximum acceptable ages at marriage, with all ages in-between being acceptable. For example, rather than viewing 25 as the best age for marriage for men or women, an individual may view all ages from 20 to 30 as equally good ages to marry. Ages close to this desirable range may also be seen as better than ages far from it. For example, if the equally acceptable age range is 20 to 30, marrying at 18, which is not desirable, may still be seen as better than marrying at age 16. If timing attitudes are conceptualized as a range, then we can only understand an individuals’ propensity to marry by comparing their age to the range of valued ages. Individuals are likely to marry when it is the “right time” for a person of their gender and unlikely to marry when it is “too early” or “too late.”
Hypotheses restated
When children are younger than their own, or their parents’, acceptable marriage age range for a person of their gender, they are unlikely to expect to marry soon and enter slowly into marriage. As children enter the acceptable age range, they become more likely to expect to marry soon and entrance into marriage increases.
None of the individual elements in this conceptual model are entirely new. As described above, each of these links has been theorized and, in many cases, tested empirically in at least one previous study. Our central contribution is in bringing all of these pieces together and testing them empirically at the same time. There are also aspects of our approach, which are new. Along with our previous study on gender differences (Allendorf et al. 2017), this is the first study to operationalize timing attitudes as a range of acceptable ages, rather than a single ideal or right age. It is also the only study to examine the influence of both mothers’ and fathers’ timing attitudes simultaneously. Even beyond timing attitudes, to our knowledge, there is only one other study that assesses the influence of both mothers’ and fathers’ attitudes—about anything—simultaneously on children (Jennings et al. 2012). Other intergenerational studies include only mothers’ attitudes (Axinn and Thornton 1992; Moen et al. 1997; Malhotra and Tsui 1999), use children’s perceptions of parents’ attitudes (Billari and Liefbroer 2007; Tosi 2017), or randomly select one parent (Dhar et al. 2019).
Methods
Data
We use panel data from the Chitwan Valley Family Study (CVFS), which comprises a representative sample of the population living in Western Chitwan Valley. The population of Chitwan Valley, which is located in south central Nepal, is highly similar to the broader population of Nepal (Barber et al. 1997). Data collection began in 2008 with a baseline survey of a sample of individuals aged 15–59 residing in Chitwan. Current residents aged 12–14 in 2008 were also part of the data collection, but were administered the baseline interview on a rolling basis in 2008–2012 after reaching 15 years of age. The final sample comprises 5840 residents aged 15–59 at the time of the baseline interview and had a response rate of 97.1%. The baseline survey was followed by ongoing collection of data on marriages through 2014 by interviews conducted every 3 months with a member of the household. Marriages were still ascertained if individuals moved outside Chitwan Valley and even outside Nepal.
Our analytical sample comprises the 1562 never married men and women aged 15–24 at baseline whose mothers and fathers were also interviewed in the baseline survey. (These 1562 youth include 972 individuals aged 15–24 in 2008 that were administered the baseline survey in 2008 and 590 individuals aged 12–14 in 2008 who were administered the baseline on a rolling basis after their fifteenth birthday.) Parents of these young people were interviewed even if they did not live with their child or lived outside Chitwan Valley, but within Nepal. There were 208 never married individuals aged 15–24 at baseline for whom one or more parents were not interviewed. Thus, the 1562 individuals in the analytical sample comprise 88% of all youth at risk of first marriage. To highlight our intergenerational approach, we refer to these 1562 individuals as children, as well as sons and daughters.
Measurement
Timing attitudes
Attitudes about marital timing were collected with six questions asked at the baseline interview before any children married. Children and parents were asked what they believe to be, in general, the ideal age at marriage for men and women separately, as well as the lowest and highest acceptable ages at marriage. English translations of the Nepali questions are: (1) “What do you feel is the ideal age for a [woman/man] to get married these days?; (2) What would be the youngest age you would consider acceptable for a [woman/man] to get married these days?; and (3) What would be the oldest age you would consider acceptable for a [woman/man] to get married these days?
One set of measures comprises respondents’ direct responses to the ideal age for the relevant gender. For daughters and their parents, this comprises the ideal age at marriage for a woman, while for sons and their parents it is the ideal marriage age for a man. We use the ideal age for women for daughters’ and their parents and the ideal age for men for sons and their parents because we expect timing attitudes are gender specific. In other words, it is only attitudes about desirable ages for people of the same gender to marry that are influential. The mean ideal age at marriage is 23–24 years of age across family members. Specifically, children’s mean ideal age at marriage for a person of their gender is 23.7 (Table 1). Mothers’ mean ideal age and fathers’ mean ideal age at marriage for a person of their child’s gender are 23.9 and 23.2, respectively (Table 1).
Table 1.
Distributions of marital timing attitudes and expectations among children and their parents (n = 1562)
| Children | Mothers | Fathers | |
|---|---|---|---|
| Mean (SD) | Mean (SD) | Mean (SD) | |
| Lowest acceptable age at marriage | 22.1 (2.7) | 22.3 (3.1) | 21.6 (3.0) |
| Ideal age at marriage | 23.7 (3.0) | 23.9 (3.7) | 23.2 (3.5) |
| Highest acceptable age at marriage | 27.1 (3.6) | 27.6 (3.9) | 27.6 (4.0) |
| % | % | % | |
| Baseline placement in range | |||
| > 5 years too early | 39.8 | 41.9 | 36.8 |
| 3–5 years too early | 40.8 | 35.4 | 36.6 |
| 1–2 years too early | 10.6 | 11.8 | 12.7 |
| In marriage age range | 8.7 | 10.6 | 13.5 |
| Too late | 0.1 | 0.3 | 0.4 |
| Time-varying placement in rangea | |||
| > 5 years too early | 23.7 | 28.8 | 22.7 |
| 3–5 years too early | 31.7 | 27.7 | 28.6 |
| 1–2 years too early | 20.1 | 17.9 | 19.2 |
| In marriage age range | 23.0 | 23.5 | 26.8 |
| Too late | 1.4 | 2.1 | 2.8 |
| Expects marriage soon | |||
| In 8 + years | 63.9 | 56.1 | 32.1 |
| In 6–7 years | 14.8 | 13.4 | 35.5 |
| In 4–5 years | 14.8 | 17.7 | 17.2 |
| Within 3 years | 6.5 | 12.8 | 15.1 |
an person-months = 82,953
In addition, we used the lowest and highest acceptable ages at marriage for women or men in general as a basis for secondary measures. Our conceptualization of how timing attitudes influence individuals’ entrance into marriage is about the interplay of the range of acceptable ages for marriage with the age of the children at risk of marriage. As noted above, when it is “too early” for marriage—children’s age is less than the lowest acceptable marriage age for a person of their gender—we expect that they are unlikely to marry. Once it is the “right time”—as children reach the lowest acceptable marriage age—we expect that they are likely to marry. We operationalized this approach by comparing children’s age to the range of acceptable marriage ages for a person of their gender.
We categorize children into the following five categories: (1) more than 5 years younger than the lowest acceptable age at marriage for a person of their gender (> 5 years “too early”), (2) 3–5 years younger than the lowest acceptable age at marriage (3–5 years “too early”), (3) 1–2 years younger than the lowest acceptable age at marriage (1–2 years “too early”), (4) within the acceptable marriage range (the “right time”), and (5) older than the highest acceptable age at marriage for a person of their gender (“too late”). We use the terms “placement in the marriage age range” and “timing attitudes” to refer to these secondary measures.
We used multiple versions of children’s placement in the acceptable marriage age range. One set used baseline age, while another set used time-varying age. Further, for both baseline and time-varying versions, we created three different measures, one using each family member’s timing attitudes. In other words, we create measures that compare the child’s [baseline/time-varying] age to the: (1) child’s own lowest and highest acceptable marriage ages for a person of their gender, (2) mother’s lowest and highest acceptable marriage ages for a person of her child’s gender, and (3) father’s lowest and highest acceptable marriage ages for a person of his child’s gender.
These multiple versions of placement in the marriage age range can be illustrated with the example of a hypothetical young man who was 18 years old at the time of the baseline interview. This hypothetical son said 25 is the lowest acceptable marriage age for a man, while his mother said 20 is the lowest acceptable marriage age for a man. So, at baseline, it is more than 5 years “too early” for the son to marry according to his own timing attitudes, but only 1–2 years “too early” according to his mother’s attitudes. Three years later when he reached 21 years of age, it was 3–5 years “too early” according to his own attitudes and the “right time” according to his mother.
Children are concentrated in the “too early” categories at the time of the baseline, but many of them aged into the “right time” during the study period. At baseline, two-fifths of the sample were more than five years younger than the youngest acceptable age, while only just over a tenth were within the acceptable marriage range (Table 1). Over the course of the data collection, however, about a quarter of children’s person-months occurred at the “right time” for marriage (Table 1). This skewed distribution is due in part to the staggered administration of the baseline survey; 39% of the analytical sample was age 15 at the time of the baseline interview and 17% age 16 (Table 2). The proportion age 15 and, to a lesser extent, age 16 is elevated because it includes children ages 15 and 16 in 2008, as well as those ages 12–14 in 2008 who were not interviewed until after their fifteenth birthday.
Table 2.
Descriptive statistics for control variables and marriage (n = 1562)
| % | |
|---|---|
| Married (during study period) | 38.7 |
| Child’s gender | |
| Son | 51.2 |
| Daughter | 48.9 |
| Caste/ethnicity | |
| Chhetri-Bahun | 46.4 |
| Dalit | 10.8 |
| Hill Janajati | 20.9 |
| Terai Janajati | 21.9 |
| Child’s education | |
| 0–6 years | 17.4 |
| 7–9 years | 53.5 |
| 10–11 years | 19.4 |
| 12 + years | 9.7 |
| Child’s nonfamily work experience | |
| None | 61.9 |
| Unsalaried work | 31.4 |
| Salaried work | 6.7 |
| Distance from urban area | mn: 8.4 |
| SD: 4.0 | |
| Child’s age at baseline | |
| 15 | 40.3 |
| 16 | 17.4 |
| 17 | 9.7 |
| 18 | 7.8 |
| 19 | 6.7 |
| 20 | 6.3 |
| 21 | 4.6 |
| 22–24 | 7.1 |
| Child’s time-varying agea | |
| 15 | 6.7 |
| 16 | 11.2 |
| 17 | 13.2 |
| 18 | 13.9 |
| 19 | 12.2 |
| 20 | 10.6 |
| 21 | 9.2 |
| 22 | 7.2 |
| 23 | 5.5 |
| 24 | 4.1 |
| 25 | 2.9 |
| 26–27 | 2.8 |
| 28–30 | 0.6 |
| Mother’s education | |
| None | 62.0 |
| 1–9 years | 30.0 |
| 10 + years | 7.9 |
| Father’s education | |
| None | 22.6 |
| 1–9 years | 52.4 |
| 10 + years | 25.0 |
an person-months = 82,953
Marriage Expectations
Children’s marriage expectations were measured with a question on how soon the child expected to marry at the time of the baseline interview. Response categories include: (1) after 8 years or more; (2) in 6–7 years; (3) in 4–5 years; or (4) within 3 years. The original response categories included separate “within 1 year” and “in 2 to 3 years” categories. Only seven children (0.5%) said they expected to marry within 1 year though. So, we collapsed “within 1 year” and “2–3 years” categories into a “within 3 years” category. Children could also volunteer that they never expected to marry. However, less than 1% of the sample did so. So, we grouped those who never expected to marry with those who expected marriage after eight or more years. Parents were also asked the same question in reference to these specific children. So, we used comparable measures for mothers’ and fathers’ expectations for their child’s marriage. In cases where children were interviewed a year or more after their parents because the child was not yet age-eligible at the time of the parents’ 2008 baseline interview, parents’ expectations of how soon they expected their child to marry were adjusted to match the timing anchor of the child’s interview.
Marriage Behavior
Marriage behavior is operationalized as the rate of children entering marriage. The dependent variable is the monthly hazard of entering a first marriage with the period of risk starting the month after the baseline survey, which was in 2008 for children aged 15–24 in 2008 and just after their fifteenth birthday for those aged 12–14 in 2008. Each person-month is coded 0 for every month a child remained unmarried and 1 for months in which a marriage occurred. All children remained at risk until they married or were right-censored when data collection ended in 2014. During data collection, 39% of children married (Table 2).
Controls
The associations we examine could be inflated by correlations with other factors that are associated with both timing attitudes and marriage expectations or behavior. So, we included controls for gender, age, caste/ethnicity, education, nonfamily work experience, distance from the urban center, mother’s education, and father’s education. All of these controls are fixed at the baseline, except age. Age is fixed at the time of the baseline interview in models of timing attitudes and expectations, but is time-varying in models of marriage behavior. Since age is strongly related to both marriage and timing attitudes, we use multiple dummies for single years of age, rather than a single, continuous measure of age (although, as shown in Table 2, oldest ages were grouped together due to small sample sizes). Using single-year dummies provides greater control for age because it does not assume a linear relationship between each successive year of age and marriage expectations or behavior.
Analytical Strategy
We evaluate our conceptual model and hypotheses in three main steps. First, we examine if parents’ timing attitudes are associated with children’s attitudes (Hypothesis 1). Children’s ideal age at marriage is the dependent variable and, since it is continuous, we use ordinary linear regression. Second, we examine whether children’s and parents’ timing attitudes influence children’s expectations (Hypotheses 2 and 3), as well as whether parents’ timing attitudes influence parents’ expectations (Hypothesis 6). We use ordered logistic regression models for this step because children’s and parent’s expectations are measured with ordinal variables on how soon marriage is expected. Lastly, we turn to behavior, testing whether the attitudes and expectations of children and their parents influence children’s entrance into marriage (Hypotheses 4, 5, 7, and 8). We take a discrete time hazard approach using logistic regression models in which the person-month is the unit of analysis (Petersen 1993). Odds ratios from these models are interpreted as speeding up or slowing down entry into marriage. All models are also adjusted for clustering within households and neighborhoods and rely on robust standard errors.
Due to the small sample size of person-months in which it was “too late” to marry, we do not discuss results for this category. At baseline, less than 1% of the children are older than the oldest acceptable age at marriage and less than 5% of children’s person-months fell into the “too late” category during the study period (Table 1). These small sample sizes preclude obtaining robust estimates for the “too late” category. These children and all their person-months remained in the analysis, but we only discuss results for the “too early” and “right time” categories.
To assess the relative influence of children, mothers, and fathers on children’s expectations and marriage behavior, we present a series of nested models. Three models show the effect of each family member’s attitudes without controlling for the other members’ attitudes: one model with only children’s attitudes, one with only mothers’ attitudes, and one with only fathers’ attitudes. A fourth, full model includes all three family members’ timing attitudes together, showing the effect of each family member net of the other two members. We used adjusted Wald type tests to test whether the effects of family members’ attitudes and expectations differ.
Another benefit of using nested models is that they can reveal multicollinearity problems. We expect parents’ and children’s attitudes and expectations to be correlated. If such correlations were large, it could lead to multicollinearity problems in models with multiple measures of attitudes or expectations. However, the nested models suggest that multicollinearity is not affecting results; standard errors are not inflated in models with measures of both children’s and parents’ attitudes or expectations.
Our analysis of entrance into marriage has a clear temporal ordering. We use timing attitudes and expectations of parents and children measured at the baseline to predict children’s subsequent entrance into marriage. Reverse causation could still affect results though if a substantial number of children planned to marry at baseline and those plans influenced their timing attitudes. However, unlike Western contexts, engagements among Nepalis are short, usually a matter of a few weeks (Allendorf 2013; Ahearn 2001). Given the short duration of engagements and preponderance of 15 and 16 year olds in our sample at baseline, it is unlikely that there were a substantial number of engagements. The question on when children expected to marry supports this expectation. As noted above, only seven children (0.5%) said they expected to marry within 1 year and 95 (6.1%) expected to marry in two to 3 years. Of these 102 children who expected to marry within 3 years, 59 actually married within the anticipated 3 years and another 24 married during the study period for a total of 83 marriages. So, even by this conservative measure, there were only a small number of individuals who could have been engaged. To assess if it affected results, we ran the marriage models with the sample excluding the 102 children that expected to marry within 3 years. The results were virtually identical, with odds ratios differing by one-hundredth of a point at most.
Analyses with attitudes and expectations as dependent variables do not have clear temporal ordering because all ideational items were measured simultaneously at baseline. Parental attitudes generally occur prior to children’s attitudes and, while there may be some influence of children’s attitudes on parental attitudes, the bulk of the intergenerational influences concerning attitudes primarily operate from parents to children (Thornton et al. 2008). Further, previous research indicates that attitudes influence expectations, rather than expectations influencing attitudes (Fishbein and Ajzen 2010). It is useful, however, to keep the potential of reciprocal causation between attitudes of parents and children in mind as we interpret results, as well as the possibility of reverse causation in which expectations influence attitudes.
Results
Do Parents’ Timing Attitudes Affect Children’s Timing Attitudes?
First, we tested hypothesis 1 by examining if children’s marital timing attitudes resemble their parents’ attitudes (Arrow 1, Fig. 1). Coefficients from linear regression models of children’s ideal age at marriage are shown in Table 3. When mothers and fathers value older ages at marriage, their children are significantly more likely to value older ages, but the effect size is modest. Each additional year of ideal age given by a mother and father is associated with an increase of 0.05 and 0.07 years, respectively, in children’s ideal age (Models 1 and 2). These positive coefficients for mothers’ and fathers’ ideal ages also remain the same or are slightly reduced when the two variables are included in the same equation (Model 3). While the point estimate for fathers’ ideal age (0.07) is slightly larger than that of mothers’ (0.05), this difference in the coefficients is not statistically significant.
Table 3.
Coefficients and robust standard errors from ordinary least squares regressions of children’s ideal age at marriage (attitudes) (n = 1562)
| Model 1 | Model 2 | Model 3 | |
|---|---|---|---|
| β (SE) | β (SE) | β (SE) | |
| Ideal age at marriage | |||
| Mother | 0.05* (.02) | 0.04* (.02) | |
| Father | 0.07** (.02) | 0.07** (.02) | |
| Child’s gender (daughter) | − 2.99*** (.15) | − 2.91*** (.17) | − 2.73*** (.18) |
| Caste/ethnicity | |||
| Chhetri-Bahun (ref) | 0.0 | 0.0 | 0.0 |
| Dalit | − 0.48 (.25) | − 0.49 (.25) | − 0.46 (.25) |
| Hill Janajati | − 0.55** (.18) | − 0.52** (.18) | − 0.52** (.18) |
| Terai Janajati | − 0.34 (.21) | − 0.38 (.21) | − 0.35 (.21) |
| Child’s education | |||
| 0–6 years (ref) | 0.0 | 0.0 | 0.0 |
| 7–9 years | 0.32 (.21) | 0.33 (.21) | 0.31 (.21) |
| 10–11 years | 0.61* (.26) | 0.61* (.26) | 0.59* (.26) |
| 12 + years | 1.27*** (.32) | 1.27*** (.32) | 1.22*** (.32) |
| Child’s nonfamily work experience | |||
| None (ref) | 0.0 | 0.00.0 | 0.0 |
| Unsalaried work | − 0.38** (.14) | − 0.39** (.14) | − 0.37** (.14) |
| Salaried work | − 0.47 (.29) | − 0.49 (.29) | − 0.49 (.29) |
| Distance from urban area | − 0.07*** (.02) | − 0.07*** (.02) | − 0.07*** (.02) |
| Child’s age at baseline | |||
| 15 | − 0.35 (.28) | − 0.34 (.28) | − 0.33 (.28) |
| 16 | − 0.36 (.28) | − 0.38 (.28) | − 0.36 (.28) |
| 17 | − 0.24 (.28) | − 0.25 (.28) | − 0.25 (.28) |
| 18 (ref) | 0.0 | 0.0 | 0.0 |
| 19 | 0.00 (.30) | − 0.01 (.29) | − 0.00 (.30) |
| 20 | − 0.25 (.32) | − 0.29 (.32) | − 0.25 (.32) |
| 21 | 0.19 (.36) | 0.18 (.36) | 0.18 (.36) |
| 22–24 | 0.05 (.32) | 0.05 (.32) | 0.06 (.32) |
| Mother’s education | |||
| None (ref) | 0.0 | 0.00.0 | 0.0 |
| 1–9 years | 0.34* (.17) | 0.34* (.17) | 0.32 (.17) |
| 10 + years | 0.93** (.30) | 0.92** (.29) | 0.88** (.29) |
| Father’s education | |||
| None (ref) | 0.0 | 0.0 | 0.0 |
| 1–9 years | − 0.09 (.18) | − 0.10 (.18) | − 0.14 (.18) |
| 10 + years | − 0.06 (.24) | − 0.08 (.24) | − 0.10 (.24) |
| Constant | 24.77*** (.81) | 24.25*** (.75) | 23.15*** (.99) |
| R2 | 0.359 | 0.361 | 0.363 |
* p < .05; ** p < .01; *** p < .001; one-tailed tests for ideal age at marriage and two-tailed tests for controls
Do Children’s and Parents’ Attitudes Affect Children’s Expectations?
Next, we tested hypothesis 2 that children’s timing attitudes influence children’s own expectations (Arrow 2, Fig. 1) and hypothesis 3 that parents’ timing attitudes directly influence children’s expectations, net of children’s own attitudes (Arrow 3, Fig. 1). The ordered logistic regression models of children’s expectation of how soon they will marry are listed in Table 4. Models 1–4 measure timing attitudes as ideal age at marriage, while Models 5–8 measure timing attitudes by comparing children’s baseline age to the range of acceptable ages.
Table 4.
Odds ratios from ordered logistic regressions of children’s expectation of marrying soon (child’s expectations) (n = 1562)
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | Model 7 | Model 8 | |
|---|---|---|---|---|---|---|---|---|
| OR | OR | OR | OR | OR | OR | OR | OR | |
| Ideal age at marriage | ||||||||
| Child | 0.80*** | 0.81*** | ||||||
| Mother | 0.93** | 0.94* | ||||||
| Father | 0.97 | 0.98 | ||||||
| Child: baseline placement in range | ||||||||
| > 5 years too early | 0.49*** | 0.52*** | ||||||
| 3–5 years too early (ref) | 1.00 | 1.00 | ||||||
| 1–2 years too early | 1.17 | 1.13 | ||||||
| In marriage age range | 1.30 | 1.35 | ||||||
| Too late | a | a | ||||||
| Mother: baseline placement in range | ||||||||
| > 5 years too early | 0.67* | 0.70* | ||||||
| 3–5 years too early (ref) | 1.00 | 1.00 | ||||||
| 1–2 years too early | 1.67** | 1.66** | ||||||
| In marriage age range | 0.85 | 0.86 | ||||||
| Too late | 0.06 | 0.87 | ||||||
| Father: baseline placement in range | ||||||||
| > 5 years too early | 0.75 | 0.96 | ||||||
| 3–5 years too early (ref) | 1.00 | 1.00 | ||||||
| 1–2 years too early | 1.25 | 1.19 | ||||||
| In marriage age range | 1.74** | 1.80** | ||||||
| Too late | 0.43 | 0.30 | ||||||
| Child’s gender (daughter) | 2.22*** | 2.99*** | 3.49*** | 1.60*** | 2.97*** | 3.53*** | 3.49*** | 2.21*** |
| Caste/ethnicity | ||||||||
| Chhetri-Bahun (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| Dalit | 0.82 | 0.87 | 0.91 | 0.79 | 0.89 | 0.96 | 0.93 | 0.91 |
| Hill Janajati | 0.95 | 1.04 | 1.03 | 0.95 | 1.02 | 1.07 | 1.02 | 1.05 |
| Terai Janajati | 0.77 | 0.79 | 0.84 | 0.74 | 0.82 | 0.85 | 0.82 | 0.83 |
| Child’s education | ||||||||
| 0–6 years (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| 7–9 years | 0.74 | 0.71 | 0.69 | 0.77 | 0.68 | 0.69 | 0.72 | 0.73 |
| 10–11 years | 1.06 | 0.92 | 0.89 | 1.10 | 0.89 | 0.90 | 0.91 | 1.00 |
| 12 + years | 0.87 | 0.70 | 0.84 | 0.94 | 0.63 | 0.63 | 0.68 | 0.72 |
| Child’s nonfamily work experience | ||||||||
| None (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| Unsalaried work | 1.45** | 1.51** | 1.54** | 1.42* | 1.54** | 1.54** | 1.52** | 1.49** |
| Salaried work | 1.78** | 1.96** | 1.96** | 1.81** | 1.66* | 1.77** | 1.84** | 1.53* |
| Distance from urban area | 1.01 | 1.02 | 1.02 | 1.01 | 1.02 | 1.02 | 1.02 | 1.01 |
| Child’s age at baseline | ||||||||
| 15 | 0.12*** | 0.14*** | 0.14*** | 0.12*** | 0.16*** | 0.18*** | 0.18*** | 0.24*** |
| 16 | 0.19*** | 0.21*** | 0.22*** | 0.18*** | 0.23*** | 0.27*** | 0.25*** | 0.32*** |
| 17 | 0.44** | 0.48** | 0.49* | 0.44** | 0.50** | 0.58 | 0.54* | 0.67** |
| 18 (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| 19 | 2.07** | 2.00** | 2.06** | 2.03** | 2.05** | 2.11** | 2.11** | 2.19** |
| 20 | 3.28*** | 3.15*** | 3.34*** | 3.14*** | 2.62*** | 3.49*** | 2.65*** | 2.30*** |
| 21 | 4.28*** | 3.87*** | 3.96*** | 4.17*** | 2.93*** | 4.34*** | 3.14*** | 2.68*** |
| 22–24 | 12.21*** | 10.87*** | 11.12*** | 11.93*** | 8.69*** | 12.93*** | 8.43*** | 7.60*** |
| Mother’s education | ||||||||
| None (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| 1–9 years | 1.21 | 1.16 | 1.15 | 1.22 | 1.19 | 1.20 | 1.15 | 1.23 |
| 10 + years | 0.56* | 0.45** | 0.45** | 0.59 | 0.50* | 0.47* | 0.44** | 0.53* |
| Father’s education | ||||||||
| None (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| 1–9 years | 1.01 | 1.06 | 1.02 | 1.07 | 0.99 | 0.99 | 1.03 | 1.03 |
| 10 + years | 0.78 | 0.85 | 0.83 | 0.82 | 0.79 | 0.81 | 0.84 | 0.82 |
| Log likelihood | − 1197.0 | − 1223.3 | − 1227.4 | − 1192.7 | − 1210.2 | − 1215.2 | − 1221.9 | − 1195.7 |
* p < .05; ** p < .01; *** p < .001; one-tailed tests for timing attitudes and two-tailed tests for controls
aOdds ratio is extremely large due to scant sample size (n = 2)
Children’s and mothers’ ideal ages did affect children’s expectations. As expected, when children and their mothers valued older ideal ages, children were significantly less likely to expect to marry soon. Specifically, the odds ratios for ideal age at marriage are 0.80 for children (Model 1) and 0.93 for mothers (Model 2) (Table 4). By contrast, the odds ratio for father’s ideal age of 0.97 is close to one and not statistically significant (Model 3, Table 4). These effects are also independent; in the full model with all three family members’ ideal ages, the odds ratios hold nearly steady (Model 4, Table 4).
The point estimates suggest children’s own ideal age has the greatest influence on their expectations, while mother’s ideal age is less influential and fathers’ ideal age has a slight influence or none at all. Adjusted Wald tests show that the odds ratios for the child’s ideal age differ significantly from the odds ratios for both mother’s and father’s ideal ages. However, the effects of mothers’ and fathers’ ideal ages on children’s expectations do not differ significantly.
When measuring timing attitudes by comparing children’s age to the acceptable marriage age ranges for a person of their gender, there is further evidence that children and parents are both influential, but in different ways. When children’s age is compared to their own timing attitudes, a clear “too early” effect emerges. When it is more than 5 years “too early,” children have about one-half the odds of expecting to marry soon compared to when it is 3–5 years “too early” (Models 5 and 8). The point estimates of 1.30 (Model 5) and 1.35 (Model 8) suggest that when it is the “right time” children have greater odds of expecting to marry soon compared to when it is 3–5 years “too early,” but these effects are not statistically significant. The point estimate for children just below the acceptable range, at 1–2 years “too early,” suggests they are slightly more likely to expect to marry soon with point estimates of 1.17 (Model 5) and 1.13 (Model 8), but these effects are also not statistically significant.
Mothers’ timing attitudes show a “too early” effect as well (Table 4). When it was more than 5 years “too early” children had significantly lower odds of expecting to marry soon compared to when it was 3–5 years “too early,” (OR 0.67 in Model 6 and OR 0.70 in Model 8). Further, when children nearly reached their mothers acceptable age range and it was only 1–2 years “too early,” they had significantly greater odds of expecting to marry soon (OR 1.67 in Model 7 and OR 1.66 in Model 10). However, children whose age was actually within their mother’s acceptable range were not more likely to expect to marry soon. The odds ratio of 0.85 is not statistically significant, nor is it in the hypothesized direction.
Fathers’ timing attitudes, on the other hand, clearly show a “right time” effect on children’s expectations (Table 4). When it was the “right time” according to fathers, children had 1.74 greater odds of expecting to marry soon, compared to when it was 3–5 years “too early” (Model 7). Further, this effect is independent of mothers’ and children’s attitudes; this odds ratio rises slightly to 1.80 in the full model (Model 8). Fathers’ attitudes do not show an independent much “too early” effect however; the odds ratio for more than 5 years “too early” is 0.96 in the full model (Model 8).
Comparing the results across family members’ indicates that there are differences in the nature and strength of their influences. Children’s attitudes show a strong much “too early” effect, mothers’ attitudes show a moderate “too early” effect and a sizeable almost “right time” effect, while fathers’ attitudes show a strong “right time” effect. Adjusted Wald tests confirm that the effects of children’s, mothers’, and fathers’ attitudes on children’s expectations do differ significantly from each other.
Do Parents’ Attitudes Affect Parental Expectations?
Next, we tested hypothesis 6, examining the effect of parental attitudes on parents’ own expectations about the timing of their children’s marriages (Arrow 6, Fig. 1; Table 5). As expected, when parents’ value older ages at marriage they are significantly less likely to expect their child to marry soon. Odds ratios measuring the effect of mothers’ and fathers’ ideal ages on their own expectations are 0.92 and 0.87, respectively (Models 1 and 3). When comparing children’s baseline age to parents’ acceptable marriage ages, we also find expected results. Mothers and fathers are significantly less likely to expect their child to marry soon when it is 5 years too early for their child to marry (OR 0.65 in Model 2 & OR 0.72 in Model 4). Similarly, mothers and fathers are more likely to expect their child to marry soon when it is 1–2 years too early or the right time compared to when it is 3–5 years too early. For mothers, however, these effects are small and not statistically significant (OR 1.22 and 1.13 in Model 2). For fathers, the odds ratios are larger and statistically significant (OR 1.43 and 1.66 in Model 4).
Table 5.
Odds ratios from ordered logistic regressions of parents’ expectations of their child marrying soon (parents’ expectations) (n = 1562)
| Mothers’ expectations | Fathers’ expectations | |||
|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | |
| OR | OR | OR | OR | |
| Ideal age at marriage | ||||
| Mother | 0.92** | |||
| Father | 0.87*** | |||
| Mother: baseline placement in range | ||||
| > 5 years too early | 0.65** | |||
| 3–5 years too early (ref) | 1.00 | |||
| 1–2 years too early | 1.22 | |||
| In marriage age range | 1.13 | |||
| Too late | a | |||
| Father: baseline placement in range | ||||
| > 5 years too early | 0.72* | |||
| 3–5 years too early (ref) | 1.00 | |||
| 1–2 years too early | 1.43* | |||
| In marriage age range | 1.66* | |||
| Too late | 3.71 | |||
| Child’s gender (daughter) | 3.01*** | 3.38*** | 2.43*** | 3.18*** |
| Caste/ethnicity | ||||
| Chhetri-Bahun (ref) | 1.00 | 1.00 | 1.00 | 1.00 |
| Dalit | 1.05 | 1.11 | 1.18 | 1.24 |
| Hill Janajati | 0.81 | 0.82 | 1.42* | 1.43* |
| Terai Janajati | 0.57** | 0.63* | 0.94 | 0.90 |
| Child’s education | ||||
| 0–6 years (ref) | 1.00 | 1.00 | 1.00 | 1.00 |
| 7–9 years | 0.69 | 0.68* | 0.96 | 0.99 |
| 10–11 years | 0.52** | 0.54** | 0.85 | 0.87 |
| 12 + years | 0.56 | 0.57 | 0.81 | 0.82 |
| Child’s nonfamily work experience | ||||
| None (ref) | 1.00 | 1.00 | 1.00 | 1.00 |
| Unsalaried work | 1.07 | 1.07 | 1.12 | 1.14 |
| Salaried work | 1.21 | 1.18 | 1.09 | 1.08 |
| Distance from urban area | 1.05** | 1.04** | 1.00 | 1.00 |
| Child’s age at baseline | ||||
| 15 | 0.06*** | 0.07*** | 0.05*** | 0.06*** |
| 16 | 0.12*** | 0.15*** | 0.13*** | 0.16*** |
| 17 | 0.45*** | 0.49** | 0.45*** | 0.51** |
| 18 (ref) | 1.00 | 1.00 | 1.00 | 1.00 |
| 19 | 1.71 | 1.67 | 1.34 | 1.27 |
| 20 | 3.11*** | 2.79*** | 3.98*** | 3.08** |
| 21 | 4.62*** | 3.95*** | 5.41*** | 4.16*** |
| 22–24 | 8.96*** | 6.98*** | 15.29*** | 9.98*** |
| Mother’s education | ||||
| None (ref) | 1.00 | 1.00 | ||
| 1–9 years | 0.61** | 0.61** | ||
| 10 + years | 0.38*** | 0.37*** | ||
| Father’s education | ||||
| None (ref) | 1.00 | 1.00 | ||
| 1–9 years | 0.67* | 0.64* | ||
| 10 + years | 0.37*** | 0.35*** | ||
| Log likelihood | − 1325.75 | − 1339.43 | − 1328.38 | − 1354.26 |
* p < .05; ** p < .01; *** p < .001; one-tailed tests for timing attitudes and two-tailed tests for controls
aOdds ratio is extremely large due to scant sample size (n = 4)
Do Children and Parents’ Timing Attitudes Influence Children’s Behavior?
Next, we examined the influences of marital timing attitudes of both children and parents on the children’s actual entrance into marriage. We evaluated these influences of marital timing attitudes on the children’s behavior with and without controlling for expectations of children’s marital timing.
We first assessed the influence of timing attitudes on behavior using the ideal age at marriage measure. Point estimates for ideal age suggest that valuing older ages may result in slight delays in marriage, but may well have no effect at all. The odds ratios are a uniform 0.98 or 0.99 for children’s, mothers’, and fathers’ ideal ages, and none are statistically significant (results not shown).
Second, we examined the influence of timing attitudes on behavior using the time-varying placement of children in the marriage age range (Table 6). The influence of timing attitudes on behavior becomes apparent with this measure. All family members’ timing attitudes show a much “too early” effect. When it is more than 5 years “too early,” children progressed into marriage substantially more slowly than when it was 3–5 years “too early” (Children’s OR 0.57, Model 1; Mother’s OR 0.72, Model 2; Father’s OR 0.55, Model 3). Further, while these effects are attenuated when other family members’ attitudes are taken into account, these much “too early” effects persist independently (Children’s OR 0.61; Mother’s OR 0.89; Father’s OR 0.63, Model 4). However, these odds ratios lack statistical significance for mothers in the full model (Model 4) and for fathers when they are in the model alone (Model 3).
Table 6.
Odds ratios from logistic regressions of children's entrance into marriage (behavior) (n person-months = 82,953)
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | Model 7 | Model 8 | |
|---|---|---|---|---|---|---|---|---|
| OR | OR | OR | OR | OR | OR | OR | OR | |
| Child: time-varying placement in range | ||||||||
| > 5 years too early | 0.57** | 0.61** | 0.58** | 0.63* | ||||
| 3–5 years too early (ref) | 1.00 | 1.00 | 1.00 | 1.00 | ||||
| 1–2 years too early | 1.11 | 1.04 | 1.09 | 1.03 | ||||
| In marriage age range | 1.05 | 0.96 | 1.00 | 0.92 | ||||
| Too late | 0.81 | 0.92 | 0.83 | 0.93 | ||||
| Mother: time-varying placement in range | ||||||||
| > 5 years too early | 0.72* | 0.89 | 0.71* | 0.88 | ||||
| 3–5 years too early (ref) | 1.00 | 1.00 | 1.00 | 1.00 | ||||
| 1–2 years too early | 1.21 | 1.19 | 1.18 | 1.15 | ||||
| In marriage age range | 1.40* | 1.39* | 1.35* | 1.36* | ||||
| Too late | 0.88 | 0.97 | 0.82 | 0.94 | ||||
| Father: time-varying placement in range | ||||||||
| > 5 years too early | 0.55 | 0.63* | 0.54** | 0.61* | ||||
| 3–5 years too early (ref) | 1.00 | 1.00 | 1.00 | 1.00 | ||||
| 1–2 years too early | 1.04 | 0.96 | 1.04 | 0.95 | ||||
| In marriage age range | 1.35* | 1.26 | 1.32* | 1.21 | ||||
| Too late | 1.09 | 1.15 | 1.03 | 0.96 | ||||
| Child: expects marriage soon | ||||||||
| In 8 + years (ref) | 1.00 | 1.00 | ||||||
| In 6–7 years | 1.37** | 1.31* | ||||||
| In 4–5 years | 1.50*** | 1.39* | ||||||
| Within 3 years | 2.02*** | 1.85*** | ||||||
| Mother: expects marriage soon | ||||||||
| In 8 + years (ref) | 1.00 | 1.00 | ||||||
| In 6–7 years | 1.00 | 0.98 | ||||||
| In 4–5 years | 1.29* | 1.25* | ||||||
| Within 3 years | 1.64** | 1.45** | ||||||
| Father: expects marriage soon | ||||||||
| In 8 + years (ref) | 1.00 | 1.00 | ||||||
| In 6–7 years | 1.06 | 1.03 | ||||||
| In 4–5 years | 1.08 | 0.89 | ||||||
| Within 3 years | 1.30* | 0.94 | ||||||
| Child’s gender (daughter) | 2.52*** | 2.41*** | 2.33*** | 2.06*** | 2.28*** | 2.22*** | 2.23*** | 2.38*** |
| Caste/ethnicity | ||||||||
| Chhetri-Bahun (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| Dalit | 0.95 | 0.95 | 0.96 | 0.95 | 0.97 | 0.95 | 0.96 | 0.97 |
| Hill Janajati | 0.88 | 0.89 | 0.88 | 0.87 | 0.89 | 0.88 | 0.87 | 0.88 |
| Terai Janajati | 0.63** | 0.64** | 0.64** | 0.63*** | 0.65** | 0.65** | 0.63** | 0.66** |
| Child’s education | ||||||||
| 0–6 years (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| 7–9 years | 0.77 | 0.76 | 0.80 | 0.77 | 0.78 | 0.77 | 0.79 | 0.80 |
| 10–11 years | 0.85 | 0.85 | 0.86 | 0.86 | 0.81 | 0.84 | 0.85 | 0.83 |
| 12 + years | 0.79 | 0.79 | 0.80 | 0.78 | 0.72 | 0.75 | 0.78 | 0.71 |
| Child’s nonfamily work experience | ||||||||
| None (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| Unsalaried work | 1.35** | 1.34** | 1.35** | 1.32* | 1.29** | 1.33** | 1.35** | 1.27* |
| Salaried work | 1.21 | 1.15 | 1.20 | 1.12 | 1.03 | 1.09 | 1.16 | 0.94 |
| Distance from urban area | 1.03* | 1.03* | 1.03* | 1.03 | 1.03 | 1.03* | 1.03* | 1.02 |
| Child’s time-varying age | ||||||||
| 15 | 0.42* | 0.43* | 0.43** | 0.42* | 0.42* | 0.42* | 0.43** | 0.49* |
| 16 | 0.78 | 0.80 | 0.80 | 0.50* | 0.78 | 0.44* | 0.80 | 0.89 |
| 17 | 0.88 | 0.92 | 0.88 | 0.90 | 0.88 | 0.82 | 0.88 | 0.95 |
| 18 (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| 19 | 1.40 | 1.40 | 1.41 | 1.38 | 1.37 | 1.37 | 1.41 | 1.35 |
| 20 | 1.18 | 1.09 | 1.05 | 0.92 | 1.12 | 1.04 | 1.05 | 0.87 |
| 21 | 1.94** | 1.79** | 1.72** | 1.50 | 1.80 | 1.67* | 1.68** | 1.37 |
| 22 | 2.40*** | 2.14*** | 2.05*** | 1.74* | 2.17** | 1.93** | 1.96*** | 1.53 |
| 23 | 2.11** | 1.89** | 1.83** | 1.54 | 1.86*** | 1.66* | 1.72* | 1.32 |
| 24 | 2.44*** | 2.18*** | 2.10*** | 1.78* | 2.07** | 1.87** | 1.95** | 1.46 |
| 25 | 3.10*** | 2.69*** | 2.48*** | 2.06* | 2.63*** | 2.28** | 2.28** | 1.69 |
| 26–27 | 3.93*** | 3.70*** | 3.16*** | 2.81** | 3.21*** | 3.15*** | 2.89*** | 2.35* |
| 28–30 | 8.62*** | 8.06*** | 7.02*** | 6.19*** | 6.95*** | 6.67*** | 6.42*** | 5.04*** |
| Mother’s education | ||||||||
| None (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| 1–9 years | 0.80* | 0.81 | 0.80* | 0.82 | 0.81 | 0.85 | 0.82 | 0.84 |
| 10 + years | 0.40*** | 0.40*** | 0.40*** | 0.41*** | 0.44** | 0.42*** | 0.41*** | 0.46** |
| Father’s education | ||||||||
| None (ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| 1–9 years | 0.99 | 0.99 | 1.01 | 1.01 | 0.99 | 1.02 | 1.02 | 1.03 |
| 10 + years | 0.90 | 0.91 | 0.92 | 0.93 | 0.90 | 0.96 | 0.93 | 0.95 |
| Log likelihood | − 3366.0 | − 3364.1 | − 3363.4 | − 3354.2 | − 3353.5 | − 3355.3 | − 3361.3 | − 3336.8 |
* p < .05; ** p < .01; *** p < .001; one-tailed tests for placement in marriage range and two-tailed tests for controls
Parents’ timing attitudes also show a “right time” effect on behavior, but children’s attitudes do not (Table 6). Children whose age is within their mother’s or father’s acceptable marriage range progressed into marriage significantly faster than when it was 3–5 years “too early.” Specifically, the odds ratio for being within their mother’s acceptable age range is a statistically significant 1.40 (Model 2), while it is 1.35 for father’s range (Model 3). These “right time” effects of parents’ attitudes are attenuated in the full model, but largely persist (Model 4). By contrast, when compared to when it was only a little “too early” for marriage, children were not more likely to enter into marriage when they were within their own acceptable marriage age range. The odds ratio of 1.11 (Model 1) is suggestive of a slight effect, but it is not statistically significant and falls to 1.04 when parents’ attitudes are added to the model (Model 4).
While the point estimates suggest that the nature of the effects differ among children’s own, mothers’, and fathers’ timing attitudes, these differences among family members do not rise to the level of statistical significance. According to adjusted Wald tests, the effects of children’s placement in the acceptable marriage age range do not differ significantly when children’s own, mothers’, or fathers’ lowest and highest acceptable marriage ages are used as the anchors.
It appears that only a small part of the effect of timing attitudes is mediated by expectations (Arrows 2, 3, 4, 6, and 7, Fig. 1). When expectations are added to the models, odds ratios for children’s, mothers’, and fathers’ timing attitudes generally move closer to one by a few hundredths (Models 5–8, Table 6). For example, when expectations are added to the model, the odds ratio for children being within their father’s range of acceptable marriage ages falls from 1.35 (Model 3) to 1.32 (Model 7). These changes are modest, however, and not all of the odds ratios move closer to one.
Do Children and Parents’ Expectations Influence Children’s Behavior?
Finally, we specifically test hypothesis 4 and examine whether children’s expectations of how soon they will marry influence marriage behavior (Arrow 4, Fig. 1). We also test hypothesis 7 that parents’ expectations influence children’s entrance into marriage (Arrow 7, Fig. 1). Children and mothers’ expectations have strong influences on children’s marriage behavior, but father’s expectations do not (Table 6). Children who expect to marry within 3 years marry at twice the rate of those who expect to marry after 8 years (OR 2.02, Model 5). Children who expect to marry within 6–7 and 4–5 years also have significantly higher marriage rates with odds ratios of 1.37 and 1.50, respectively (Model 5). Mother’s expectations also have a large, independent effect on children’s behavior. When mothers expect them to marry within 3 years, children marry at substantially higher rates compared to those whose mothers expect marriage only after 8 years or more (OR 1.64, Model 6). In the full model, odds ratios for mothers’ and children’s expectations are reduced to 1.85 and 1.45, respectively, but remain sizeable and significant (Model 8). By contrast, when father’s expectations are in a model by themselves, the odds ratio of 1.30 suggests a sizeable and significant effect on children’s behavior (Model 7), but falls to 0.94 in the full model and loses statistical significance (Model 8).
The differences among family members seen in the point estimates of the odds ratios are further supported by adjusted Wald tests. Both mothers and children’s expectations have significantly larger effects than father’s expectations. The effect of children’s expectations does not differ significantly from that of mothers’ expectations however.
Discussion and Conclusion
We found support for our conceptual model and many of its hypotheses. Marital timing attitudes had moderate-sized, statistically significant, influences on children’s marriage expectations and behavior. When children and parents positively view older ages at marriage, children expected to marry at older ages. When it was much “too early” to marry, children expected to marry later and actually did marry at lower rates. Further, when it was the “right time” to marry according to their parents, children were more likely to marry.
We also found strong support for our intergenerational approach. Both generations’ timing attitudes have important influences on children’s expectations and behavior, with children having only a slight advantage over parents. Children’s own timing attitudes had greater influence on children’s expectations than their parents’ timing attitudes. However, children’s timing attitudes did not have a greater influence than parents’ attitudes on the actual entrance into marriage. These findings are consistent with the only previous study that examined both mothers and fathers’ attitudes; Jennings and colleagues’ (2012) found that mothers’ and fathers’ attitudes about fertility and old age support independently influence sons’ marital timing.
Mothers and fathers were relatively even in influencing their children, but mothers had an edge over fathers. Two pathways—parents’ timing attitudes shaping children’s attitudes and expectations—did not differ significantly between mothers and fathers, indicating that mothers and fathers may well be equally influential in instilling timing attitudes in their children and influencing children’s formation of their own marital expectations. However, a third pathway of social control showed a strong gender differential; mothers’ expectations had a large influence on children’s marriage behavior, while fathers’ expectations had none. It may be primarily mothers who are responsible for ensuring that their children marry at the right time. Further, while marriage is important to the identity and well-being of both men and women, it has a special salience for women. While mothers may have less decision-making power than fathers overall, the salience of marriage may motivate mothers to direct the power they do have toward ensuring well-timed marriages for their children.
Overall, this empirical support for our intergenerational model makes several important contributions. For life course research on age norms, we extend the influence of timing attitudes from adolescents’ home-leaving (Billari and Liefbroer 2007; Tosi 2017) and childbearing (Van Bavel and Nitsche 2013) to marriage. For research on marriage, we add attitudes about marital timing to the list of ideational factors that shape the entrance into marriage, alongside gender ideology (e.g., Clarkberg et al. 1995), religion (e.g., Lehrer 2004), and attitudes about other family behaviors (e.g., Jennings et al. 2012). We also provide empirical illustration of the value of following calls to operationalize timing attitudes as a range (Liefbroer and Billari 2010).
These results have implications for widespread changes in marital timing. Age at marriage is rising around the world with both men and women increasingly delaying marriage (United Nations 2016; Ortega 2014). Developmental idealism theory suggests an important cause of this trend is the spread of values and beliefs about family behaviors and their causal links to socioeconomic development (Thornton 2001, 2005). Specifically, among the schemas found within developmental idealism are the valuation of older ages at marriage and beliefs that marrying later is inherently good, modern, and related to development. Our finding that timing attitudes do influence marriage provides empirical support for developmental idealism theory and the potential role of timing attitudes in the global trend toward later marriage (Allendorf et al. 2019).
These results further suggest that marital timing attitudes have important consequences for individuals’ lives and well-being. As noted above, shifting marriage forward or back by just a few years can have important life-long consequences for education, employment, and health. When young people and parents value older ages at marriage, those young people are likely, on average, to stay in school longer, earn more income in the labor market, and be less likely to experience maternal mortality and other adverse health outcomes due to early childbearing. On the other hand, they may also incur substantially higher dowry costs, damage their reputations, and marry poorer quality spouses.
It is important to reinforce, however, that our analyses have limitations. Most importantly, while our estimates of the influence of timing attitudes on marriage benefit from clear temporal ordering they may still be inflated by other unknown factors related to both timing attitudes and marital timing. Estimates of the influence of parents’ timing attitudes on children’s attitudes and the influence of timing attitudes on expectations could further be inflated by reverse causation. Also, due to the small sample size of person-months at older ages, we were unable to rigorously assess the influence of it being “too late” to marry. Future studies should address these limitations by following individuals for a period longer than 7 years and identifying an instrumental variable for timing attitudes or using other more rigorous identification strategies. We also reinforce that intergenerational influences we identify work through transmission of parents’ attitudes to children, as well as parents’ social control of children. Future studies should do more to disentangle the relative role of social control versus ideational transmission.
Future research should also extend and elaborate on these findings by evaluating this conceptual model in other settings. We expect that timing attitudes have similar influences well beyond Chitwan; we expect that when it is the “right time” to marry, marriage is, in general, more likely than when it is “too early.” Beyond this simple connection though, there may well be important differences in these links across contexts. The size of intergenerational influences likely varies according to the extent of parental involvement in children’s marriages. The effect of parents’ timing attitudes may be larger in Northern India, where parents have more control over children’s marriages than in Chitwan, but smaller in the USA, where parents have little to no control over children’s marriages. We do not expect these intergenerational effects to be nonexistent though. As noted above, Axinn and Thornton (1992) previously found that mother’s timing attitudes influence children’s entrance into marriage in the USA.
Future studies should also assess whether cohabitation conditions these influences. If cohabitation is not acceptable, the available choices are to marry now or wait and forgo partnership entirely until marrying later. In places where cohabitation is widely acceptable, there are more options. When it is “too early” for marriage, individuals may cohabit until it is the “right time” and, if it is “too late,” they may opt to forgo marriage entirely in favor of cohabitation. If so, the influence of timing attitudes about marriage—but not union formation in general—would be more influential in contexts where cohabitation is an acceptable alternative to marriage.
Acknowledgements
We thank the Institute for Social and Environmental Research in Chitwan, Nepal, which collected and processed the data. The content of the article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Institute for Social and Environmental Research.
Funding
The data collection, management, and analysis for this article were supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health under Grant R01-HD-054847 and R24-HD041028.
Compliance with Ethical Standards
Conflict of interest
The authors declare that they have no conflict of interest.
Human and Animal Rights
This research does involve research with human participants that was approved by the appropriate Institutional Review Boards.
Informed Consent
All research with human participants included informed consent.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- Addo FR. Debt, cohabitation, and marriage in young adulthood. Demography. 2014;51(5):1677–1701. doi: 10.1007/s13524-014-0333-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ahearn LA. Invitations to love: Literacy, love letters and social change in Nepal. Ann Arbor: The University of Michigan Press; 2001. [Google Scholar]
- Allendorf K. Schemas of marital change: From arranged marriages to eloping for love. Journal of Marriage and Family. 2013;75(2):453–469. doi: 10.1111/jomf.12003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Allendorf K. Conflict and compatibility? Developmental idealism and gendered differences in marital choice. Journal of Marriage and Family. 2017;79(2):337–355. doi: 10.1111/jomf.12339. [DOI] [Google Scholar]
- Allendorf K, Pandian RK. The decline of arranged marriage? Marital change and continuity in India. Population and Development Review. 2016;42(3):435–464. doi: 10.1111/j.1728-4457.2016.00149.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Allendorf K, Thornton A, Mitchell C, Young-DeMarco L. The influence of developmental idealism on marital attitudes, expectations, and timing. Journal of Family Issues. 2019;40(17):2359–2388. doi: 10.1177/0192513X19856642. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Allendorf K, Thornton A, Mitchell C, Young-DeMarco L, Ghimire DJ. Early women, late men: Timing attitudes and gender differences in marriage. Journal of Marriage and Family. 2017;79(5):1478–1496. doi: 10.1111/jomf.12426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Andrist L, Banerji M, Desai S. Negotiating marriage: Examining the gap between marriage and cohabitation in India. In: Kaur R, Palriwala R, editors. Marrying in South Asia: Shifting concepts, changing practices in a globalising world. Delhi: Orient Black Swan; 2014. pp. 116–140. [Google Scholar]
- Axinn WG, Clarkberg ME, Thornton A. Family influences on family-size preferences. Demography. 1994;31(1):65–79. doi: 10.2307/2061908. [DOI] [PubMed] [Google Scholar]
- Axinn WG, Ghimire DJ, Barber JS. The influence of ideational dimensions of social change on family formation in Nepal. In: Jayakody R, Thornton A, Axinn WG, editors. International family change: Ideational perspectives. New York: Lawrence Erlbaum Associates; 2008. pp. 251–280. [Google Scholar]
- Axinn WG, Thornton A. The influence of parental resources on the timing of the transition to marriage. Social Science Research. 1992;21(3):261–285. doi: 10.1016/0049-089X(92)90008-5. [DOI] [Google Scholar]
- Axinn WG, Thornton A. Mothers, children, and cohabitation—The intergenerational effects of attitudes and behavior. American Sociological Review. 1993;58(2):233–246. doi: 10.2307/2095968. [DOI] [Google Scholar]
- Bachrach CA. Culture and demography: From reluctant bedfellows to committed partners. Demography. 2014;51(1):3–25. doi: 10.1007/s13524-013-0257-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bandura A. Social foundations of thought and action. Englewood Cliffs, NJ: Prentice Hall; 1986. [Google Scholar]
- Barber JS. Community social context and individualistic attitudes toward marriage. Social Psychology Quarterly. 2004;67(3):236–256. doi: 10.1177/019027250406700302. [DOI] [Google Scholar]
- Barber JS, Axinn WG. Gender role attitudes and marriage among young women. Sociological Quarterly. 1998;39(1):11–31. doi: 10.1111/j.1533-8525.1998.tb02347.x. [DOI] [Google Scholar]
- Barber JS, Shivakoti GP, Axinn WG, Gajurel K. Sampling strategies for rural settings: A detailed example from Chitwan Valley Family Study, Nepal. Nepal Population Journal. 1997;6(5):193–203. [Google Scholar]
- Billari FC, Liefbroer AC. Should I stay or should I go? The impact of age norms on leaving home. Demography. 2007;44(1):181–198. doi: 10.1353/dem.2007.0000. [DOI] [PubMed] [Google Scholar]
- Buttenheim AM, Nobles J. Ethnic diversity, traditional norms, and marriage behaviour in Indonesia. Population Studies-a Journal of Demography. 2009;63(3):277–294. doi: 10.1080/00324720903137224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Caltabiano M, Castiglioni M. Changing family formation in Nepal: Marriage, cohabitation and first sexual intercourse. International Family Planning Perspectives. 2008;34(1):30–39. doi: 10.1363/3403008. [DOI] [PubMed] [Google Scholar]
- Chari AV, Heath R, Maertens A, Fatima F. The causal effect of maternal age at marriage on child wellbeing: Evidence from India. Journal of Development Economics. 2017;127:42–55. doi: 10.1016/j.jdeveco.2017.02.002. [DOI] [Google Scholar]
- Cherlin AJ. The deinstitutionalization of American marriage. Journal of Marriage and Family. 2004;66(4):848–861. doi: 10.1111/j.0022-2445.2004.00058.x. [DOI] [Google Scholar]
- Chodorow N. The reproduction of mothering. Berkeley, CA: University of California Press; 1978. [Google Scholar]
- Chowdhry P. Contentious marriages, eloping couples: Gender, caste, and patriarchy in northern India. New Delhi: Oxford University Press; 2007. [Google Scholar]
- Clarkberg M, Stolzenberg RM, Waite LJ. Attitudes, values, and entrance into cohabitational versus marital unions. Social Forces. 1995;74(2):609–632. doi: 10.2307/2580494. [DOI] [Google Scholar]
- De Valk HAG, Liefbroer AC. Timing preferences for women’s family-life transitions: Intergenerational transmission among migrants and Dutch. Journal of Marriage and Family. 2007;69(1):190–206. doi: 10.1111/j.1741-3737.2006.00353.x. [DOI] [Google Scholar]
- Dhar D, Jain T, Jayachandran S. Intergenerational transmission of gender attitudes: Evidence from India. The Journal of Development Studies. 2019;55(12):2572–2592. doi: 10.1080/00220388.2018.1520214. [DOI] [Google Scholar]
- Efevbera Y, Bhabha J, Farmer PE, Fink G. Girl child marriage as a risk factor for early childhood development and stunting. Social Science and Medicine. 2017;185:91–101. doi: 10.1016/j.socscimed.2017.05.027. [DOI] [PubMed] [Google Scholar]
- Elder GH. Time, human agency, and social change - perspective on the life course. Social Psychology Quarterly. 1994;57(1):4–15. doi: 10.2307/2786971. [DOI] [Google Scholar]
- Elder, G. H., Shanahan, M. J., & Jennings, J. A. (2015). Human development in time and place. In M. H. Bornstein, & T. Leventhal (Eds.), Ecological settings and processes in developmental systems (Vol. Handbook of Child Psychology and Developmental Science, pp. 6–54). Hoboken, NJ: Wiley.
- Field E, Ambrus A. Early marriage, age of menarche, and female schooling attainment in Bangladesh. Journal of Political Economy. 2008;116(5):881–930. doi: 10.1086/593333. [DOI] [Google Scholar]
- Fishbein M, Ajzen I. Predicting and changing behavior: The reasoned action approach. New York: Psychology Press, Taylor & Francis Group; 2010. [Google Scholar]
- Ghimire DJ, Axinn WG, Yabiku ST, Thornton A. Social change, premarital nonfamily experience, and spouse choice in an arranged marriage society. American Journal of Sociology. 2006;111(4):1181–1218. doi: 10.1086/498468. [DOI] [Google Scholar]
- Goldscheider F, Kaufman G, Sassler S. Navigating the “new” marriage market: How attitudes toward partner characteristics shape union formation. Journal of Family Issues. 2009;30(6):719–737. doi: 10.1177/0192513x09331570. [DOI] [Google Scholar]
- Grover S. Marriage, love, caste, and kinship support: Lived experiences of the urban poor in India. New Delhi: Social Science Press; 2011. [Google Scholar]
- Guzzo KB. Marital intentions and the stability of first cohabitations. Journal of Family Issues. 2009;30(2):179–205. doi: 10.1177/0192513x08323694. [DOI] [Google Scholar]
- Hahn Y, Nuzhat K, Yang HS. The effect of female education on marital matches and child health in Bangladesh. Journal of Population Economics. 2018;31(3):915–936. doi: 10.1007/s00148-017-0673-9. [DOI] [Google Scholar]
- Jennings EA, Axinn WG, Ghimire DJ. The effect of parents’ attitudes on sons’ marriage timing. American Sociological Review. 2012;77(6):923–945. doi: 10.1177/0003122412464041. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Johnson-Hanks JA, Bachrach CA, Morgan SP, Kohler H-P. Understanding family change and variation: Toward a theory of conjuntural action. Dordrecht: Springer; 2011. [Google Scholar]
- Killewald A, Gough M. Does specialization explain marriage penalties and premiums? American Sociological Review. 2013;78(3):477–502. doi: 10.1177/0003122413484151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kim HK, McKenry PC. The relationship between marriage and psychological well-being—A longitudinal analysis. Journal of Family Issues. 2002;23(8):885–911. doi: 10.1177/019251302237296. [DOI] [Google Scholar]
- Koball HL. Crossing the threshold: Men’s incomes, attitudes toward the provider role, and marriage timing. Sex Roles. 2004;51(7–8):387–395. doi: 10.1023/B:SERS.0000049228.77597.1a. [DOI] [Google Scholar]
- Lehrer EL. Religion as a determinant of economic and demographic behavior in the United States. Population and Development Review. 2004;30(4):707–726. doi: 10.1111/j.1728-4457.2004.00038.x. [DOI] [Google Scholar]
- Lesthaeghe R. The unfolding story of the second demographic transition. Population and Development Review. 2010;36(2):211–251. doi: 10.1111/j.1728-4457.2010.00328.x. [DOI] [PubMed] [Google Scholar]
- Liefbroer AC, Billari FC. Bringing norms back in: A theoretical and empirical discussion of their importance for understanding demographic behaviour. Population, Space and Place. 2010;16(4):287–305. doi: 10.1002/psp.552. [DOI] [Google Scholar]
- Liefbroer AC, Gerritsen L, Gierveld JD. The influence of intentions and life-course factors on union formation behavior of young adults. Journal of Marriage and the Family. 1994;56(1):193–203. doi: 10.2307/352713. [DOI] [Google Scholar]
- Ludwig V, Bruderl J. Is there a male marital wage premium? New evidence from the United States. American Sociological Review. 2018;83(4):744–770. doi: 10.1177/0003122418784909. [DOI] [Google Scholar]
- Mahmud, S., & Amin, S. (2006). Girls’ schooling and marriage in rural Bangladesh. In E. Hannum (Ed.), Childrens lives and schooling across societies (Vol. 15, pp. 71–99). Research in the Sociology of Education.
- Malhotra A, Tsui AO. Marriage timing in Sri Lanka: The role of modern norms and ideas. Journal of Marriage and the Family. 1996;58(2):476–490. doi: 10.2307/353511. [DOI] [Google Scholar]
- Malhotra A, Tsui AO. Work and marriage: Mother-daughter similarities in Sri Lanka. Journal of Comparative Family Studies. 1999;30(2):219–241. doi: 10.3138/jcfs.30.2.219. [DOI] [Google Scholar]
- Mensch BS, Singh S, Casterline J. Trends in the timing of first marriage among men and women in the developing world. In: Lloyd CB, Behrman JR, Stromquist NP, Cohen B, editors. The changing transitions to adulthood in developing countries: selected studies. Washington, DC: National Academies Press; 2005. pp. 118–171. [Google Scholar]
- Ministry of Health [Nepal], New ERA, & ICF . Nepal demographic and health survey 2016. Kathmandu, Nepal: Springer; 2017. [Google Scholar]
- Mody P. The intimate state: Love-marriage and the law in Delhi. Delhi: Routledge India; 2008. [Google Scholar]
- Moen P, Erickson MA, DempsterMcClain D. Their mother’s daughters? The intergenerational transmission of gender attitudes in a world of changing roles. Journal of Marriage and the Family. 1997;59(2):281–293. doi: 10.2307/353470. [DOI] [Google Scholar]
- Ortega JA. A characterization of world union patterns at the national and regional level. Population Research and Policy Review. 2014;33(2):161–188. doi: 10.1007/s11113-013-9301-x. [DOI] [Google Scholar]
- Petersen T. Recent advances in longitudinal methodology. Annual Review of Sociology. 1993;19:425–454. doi: 10.1146/annurev.so.19.080193.002233. [DOI] [Google Scholar]
- Santhya KG. Early marriage and sexual and reproductive health vulnerabilities of young women: A synthesis of recent evidence from developing countries. Current Opinion in Obstetrics and Gynecology. 2011;23(5):334–339. doi: 10.1097/GCO.0b013e32834a93d2. [DOI] [PubMed] [Google Scholar]
- Sassler S, Goldscheider F. Revisiting Jane Austen’s theory of marriage timing—Changes in union formation among American men in the late 20th century. Journal of Family Issues. 2004;25(2):139–166. doi: 10.1177/0192513x03257708. [DOI] [Google Scholar]
- Sassler S, Schoen R. The effect of attitudes and economic activity on marriage. Journal of Marriage and the Family. 1999;61(1):147–159. doi: 10.2307/353890. [DOI] [Google Scholar]
- Settersten RA. The salience of age in the life course. Human Development. 1997;40(5):257–281. doi: 10.1159/000278730. [DOI] [Google Scholar]
- Settersten RA. A time to leave home and a time never to return? Age constraints on the living arrangements of young adults. Social Forces. 1998;76(4):1373–1400. doi: 10.1093/sf/76.4.1373. [DOI] [Google Scholar]
- Settersten RA, Hägestad GO. What’s the latest? Cultural age deadlines for family transitions. The Gerontologist. 1996;36(2):178–188. doi: 10.1093/geront/36.2.178. [DOI] [PubMed] [Google Scholar]
- Settersten RA, Mayer KU. The measurement of age, age structuring, and the life course. Annual Review of Sociology. 1997;23:233–261. doi: 10.1146/annurev.soc.23.1.233. [DOI] [Google Scholar]
- Swidler A. Talk of love: How culture matters. Chicago: University of Chicago Press; 2001. [Google Scholar]
- Thornton A. The developmental paradigm, reading history sideways, and family change. Demography. 2001;38(4):449–465. doi: 10.2307/3088311. [DOI] [PubMed] [Google Scholar]
- Thornton A. Reading history sideways: The fallacy and enduring impact of the developmental paradigm on family life. Chicago: The University of Chicago Press; 2005. [Google Scholar]
- Thornton A, Axinn WG, Hill DH. Reciprocal effects of religiosity, cohabitation, and marriage. American Journal of Sociology. 1992;98(3):628–651. doi: 10.1086/230051. [DOI] [Google Scholar]
- Thornton A, Axinn WG, Xie Y. Marriage and cohabitation. Chicago: University of Chicago Press; 2008. [Google Scholar]
- Thornton A, Lin H-S. Social change and the family in Taiwan. Chicago: The University of Chicago Press; 1994. [Google Scholar]
- Tosi M. Age norms, family relationships, and home leaving in Italy. Demographic Research. 2017;36:281–306. doi: 10.4054/DemRes.2017.36.9. [DOI] [Google Scholar]
- Uecker JE, Stokes CE. Early marriage in the United States. Journal of Marriage and Family. 2008;70(4):835–846. doi: 10.1111/j.1741-3737.2008.00530.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- United Nations (2016). Changing patterns of marriage and unions across the world. In D. o. E. a. S. A. United Nations, Population Division (Ed.), Population Facts (pp.).
- Van Bavel J, Nitsche N. The proper age for parenthood and second birth rates in Europe. European Sociological Review. 2013;29(6):1149–1161. doi: 10.1093/esr/jct003. [DOI] [Google Scholar]
- Van de Kaa DJ. Europe’s second demographic transition. Population Bulletin. 1987;42(1):1–57. [PubMed] [Google Scholar]
- Waite LJ, Gallagher M. The case for marriage: Why married people are happier, healthier, and better off financially. New York: Doubleday; 2000. [Google Scholar]
- Waite LJ, Lehrer EL. The benefits from marriage and religion in the United States: A comparative analysis. Population and Development Review. 2003;29(2):255–275. doi: 10.1111/j.1728-4457.2003.00255.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Waller MR, McLanahan SS. “His” and “her” marriage expectations: Determinants and consequences. Journal of Marriage and Family. 2005;67(1):53–67. doi: 10.1111/j.0022-2445.2005.00005.x. [DOI] [Google Scholar]
- Yabiku ST. The effect of non-family experiences on age of marriage in a setting of rapid social change. Population Studies-a Journal of Demography. 2005;59(3):339–354. doi: 10.1080/00324720500223393. [DOI] [PubMed] [Google Scholar]
- Yabiku ST. Neighbors and neighborhoods: Effects on marriage timing. Population Research and Policy Review. 2006;25(4):305–327. doi: 10.1007/s11113-006-9006-5. [DOI] [Google Scholar]
- Yamokoski A, Keister LA. The wealth of single women: Marital status and parenthood in the asset accumulation of young baby boomers in the United States. Feminist Economics. 2006;12(1–2):167–194. doi: 10.1080/13545700500508478. [DOI] [Google Scholar]
- Yount KM, Crandall A, Cheong YF, Osypuk TL, Bates LM, Naved RT, et al. Child marriage and intimate partner violence in rural Bangladesh: A longitudinal multilevel analysis. Demography. 2016;53(6):1821–1852. doi: 10.1007/s13524-016-0520-8. [DOI] [PMC free article] [PubMed] [Google Scholar]