Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Mar 1.
Published in final edited form as: Stud Fam Plann. 2013 Mar;44(1):67–84. doi: 10.1111/j.1728-4465.2013.00344.x

The Influence of Neighbors’ Family Size Preference on Women’s Progression to Higher Parity Births

Elyse A Jennings 1, Jennifer S Barber 2
PMCID: PMC3621985  NIHMSID: NIHMS443182  PMID: 23512874

Abstract

Large families can have a negative impact on the health and well-being of women, children, and their communities. Although 60 percent of people in our rural Nepalese sample report that two children is their ideal number, nearly half of the married women continue childbearing after their second child. We explore the influence of women’s and their neighbors’ family size preferences on women’s progression to higher parity birth, comparing this influence across two cohorts. We find that a) neighbors’ family size preferences influence women’s fertility, b) older cohorts of women are more influenced by their neighbors’ preferences than younger cohorts of women, and c) the influence of neighbors’ preferences is independent of women’s own preferences.

Keywords: Neighbor Influences, Parity Progression, Attitudes, Fertility, South Asia

Introduction

Most countries have experienced rapid fertility decline in recent decades (Caldwell 2001), in part due to the global dissemination of Western ideals that favor smaller families. This dissemination has occurred both through the non-deliberate diffusion of ideas within and between populations (Casterline 2001; Hornik and McAnany 2001; Thornton 2005; Watkins 1987) and the more deliberate family planning initiatives that focused on changing people’s preferences and making those preferences achievable (Thornton et al. 2012). These values have been successfully circulated in rural Nepal, where sixty percent of people in the Southern region of Chitwan say that two children is the ideal.1, 2 Yet fertility behavior in this area is slower to change: about half of all married women have more than two children.

It is important to understand why women may choose to have more than two children, as having a large family can have negative consequences for the health and well-being of both mothers and children. Compared to small families, larger families tend to have more closely spaced births. With more mouths to feed and less time for a mother to recuperate between births, mothers and children of these growing families may be at risk of nutritional deprivation or other negative health outcomes (Cleland and Sathar 1984; Curtis, Diamond, and McDonald1993; Desai 1995; Horton 1986; Winnikoff 1983). Furthermore, parents have finite resources to distribute among their children, and the resources available to each child are reduced as family size increases (Powell and Steelman 1990; Zajonc and Mullally 1997). For example, parents may invest less in each child’s education when they have more children (Blake 1981; Kessler 1991; Knodel, Havanon, and Sittitrai 1990; Knodel 1991). In addition, larger families may reduce parental emotional investment per child (Kidwell 1981), which can impede social and emotional growth. These negative consequences are particularly likely if one or more of the births is unintended (Barber and East 2009; Barber and East 2011), and higher parity births are more likely to be unintended (Eggleston 1999; Marston and Cleland 2003).3

On the other hand, women who prefer large families may choose to have more than two children despite the potential negative consequences. Substantial evidence links individual fertility preferences to behaviors in Western settings (e.g., Ajzen 1988; Barber 2001; Fishbein and Ajzen 1975). In this rural South Asian setting, where lives are embedded in families and communities, it is likely that neighbors’ preferences also have a great deal of influence on women’s behavior (Moursund and Kravdal 2003). Neighbors are often aware of individuals’ actions and these actions may even have consequences for neighbors’ well-being, providing an incentive to encourage people in their community, or neighborhood, to behave in certain ways. Furthermore, we would expect individuals in this setting to be aware of their neighbors’ preferences and to feel pressure to behave in ways that maintain good standing with them. Recent social changes, however, may have reduced the influence of neighbors among younger cohorts, who have greater exposure than older cohorts to more individualistic ideas through having spent more time outside of their homes and communities (Axinn and Yabiku 2001; Ghimire et al. 2006; Yabiku 2005).

This study explores the influence of neighbors’ family size preferences on women’s progression to higher parity – that is, a third or fourth live birth – and compares this influence across cohorts to assess whether younger women are less responsive than older women to their neighbors’ preferences. We use unique data on attitudes and preferences from a rural Nepalese setting. Such measures are rare in such South Asian contexts. Because these data include preference measures at the individual-level garnered from interviews with every member of each sampled neighborhood, they allow us to construct a measure of preferences at the neighborhood-level (average preference across residents) and investigate this neighborhood-level influence on individual behavior. We investigate influences of neighbors’ family size preferences on individual fertility behavior in the decade after these preferences were measured. In doing so, we consider whether these neighbor influences operate via women’s own preferences or net of women’s own preferences.

Below we outline the theoretical reasons to expect that neighbors’ attitudes and preferences may be important in determining women’s behavior.

Conceptual Framework

Many models of behavior share the assumption that behavior results from a reasoned process in which individuals consider their options, evaluate the consequences, and make decisions about how to act. For example, the reasoned action and planned behavior frameworks (Fishbein and Ajzen 1975) have been used to refine demographic theories of fertility decline, such as the concept of the KAP-gap (Freedman et al. 1974), in which the gap between the desire to stop childbearing and the lack of contraceptive use was referred to as “discrepant behavior.” Lesthaeghe’s “Ready, Willing, and Able” (Lesthaeghe and Vanderhoeft 2001) and Coale’s “Three Conditions for Fertility Decline” (Coale 1973) also employed the reasoned action and planned behavior frameworks to posit that people’s attitudes must be favorable toward smaller families before they choose to adopt family planning methods.4

There are many reasons to expect that neighbors’ preferences will influence women’s behavior in this setting. In fact, the theory of planned behavior posits that individuals are directly influenced by consideration of others’ attitudes (Fishbein and Ajzen 1975). For example, ample evidence shows that parents influence the behaviors of their offspring (Axinn and Thornton 1993; Axinn and Thornton 1996; Barber 2000; Bengtson 1975). Friends and peers can also influence behavior, through social norms and stigma that directly guide behavior and may shape preferences (Rutenberg and Watkins 1997; Stuber, Galea, and Link 2008). We expect neighbors to have a similar influence. In this rural Nepalese context, neighbors are likely to be among an individual’s most important social networks. Neighborhoods in Nepal tend to be fairly ethnically homogenous and residents tend to be stable. Because of the way land is acquired – almost solely through inheritance – people usually reside within the neighborhood in which their parents (in the case of males) or in-laws (in the case of females) reside. Individuals’ lives are exposed to their neighbors, with housing being quite open: windows are not covered with glass, front doors often are left open, and whole families sometimes sleep in the yard in hot weather. In fact, when one is sitting on the front porch visiting a Nepalese family, it is not uncommon to see the toothbrushes, birth control pills, and the family’s other personal items stuffed into the thatched roof of the porch for all to view. In addition, neighbors interact with one another on a daily basis during their routine activities, such as collecting water or firewood (Cameron 1998). As a result, neighbors’ preferences and behaviors are quite likely to be known and to be a part of every resident’s daily life (Barber 2004).

Mechanisms of Neighbors’ Influence

Because people tend to be aware of their neighbors’ preferences in this setting, neighbors’ preferences have ample opportunity to influence individual women’s behaviors. This influence can occur via neighborhood socialization, in which neighbors’ preferences influence women’s behavior through women’s own preferences (Barber 2000; Katz, Joiner Jr., and Kwon 2002). Neighbors’ close and constant interaction allows for new ideas to diffuse between them (Barber 2004; Behrman, Kohler, and Watkins 2002; Bongaarts and Watkins 1996; Cleland 2001; Lesthaeghe 1978). This diffusion process may lead women to internalize their neighbors’ preferences and develop or change their own preferences to align with their neighbors’.

The influence of neighbors’ preferences on women’s behavior can also occur through social pressure, in which neighbors’ preferences have a direct influence on women’s behavior (Barber 2000; Fishbein and Ajzen 2010; Troyer and Younts 1997). Neighbors may exert social pressure on women by embodying and enforcing social norms, which guide women’s behavioral choices. Neighbors also have the power to punish, via stigmatization, if someone in the neighborhood does not behave as deemed appropriate (Coleman 1990; Stuber et al. 2008). These forces may lead women to set aside their own desires in order to appease their neighbors. For example, a woman may want a large family, but she may be aware that this is socially undesirable in her neighborhood. Therefore, she may limit her fertility to avoid being stigmatized or even ostracized. Although social pressure acts like socialization in influencing women to align their behavior to the preferences of those around them, social pressure is different from socialization in that its influence on behavior may run counter to a woman’s own preferences. Although distinct, these two mechanisms may operate simultaneously.

Cohort Differences in Neighbors’ Influence

The influence of neighbors’ preferences relative to individuals’ own preferences may change over historic time, as collective values become less common. This setting of rural Nepal has experienced especially rapid social change in recent decades (Axinn and Yabiku 2001; Yabiku 2005), including the spread of schools, easier access to the city via the bus system, an increased number of markets, and increased opportunities for employment. As a result of this improved access to nonfamily institutions and experiences, people spend more time outside of their home communities than they did in the past. Given the timing of these social changes, younger cohorts of women have had greater exposure to nonfamily experiences and experiences outside of their immediate community than older cohorts, who spent much of their lives with limited access to schools, public transportation, nonfamily employment, or city life. Additionally, younger cohorts have had more exposure to the new ideas and values that accompany these “outside” experiences. As a result, younger cohorts of women may be less influenced than their older counterparts by their neighbors’ preferences, choosing to follow ideas acquired from outside of their neighborhoods.

Older cohorts of women, on the other hand, grew up during a time when there were fewer opportunities or reasons to venture outside of their neighborhood. These older, more community-centric, cohorts are likely to have placed more importance on behaving in accordance with their neighbors’ preferences and thus to have been more susceptible to both neighborhood socialization and social pressure. These women had less exposure to the types of experiences and ideas from outside of their community that may compete with the preferences of their own neighbors. For example, a woman from an older cohort may have perceived that her neighbors preferred large families and, therefore, bear many children of her own. On the other hand, a younger woman living in the same neighborhood may weigh the ideas she has acquired from school or media more heavily and opt to stop her childbearing after she has had only one or two children.

Fertility and Family Size Preference

Similar to other countries in South Asia, there was a rapid decline in the fertility rate of Nepal after family planning initiatives began in the late 1950s. The total fertility rate fell from 6.1 in the early 1950s to 4.41 in the late 1990s, and then down to 3.0 by 2011 (Population Reference Bureau 2011; Thornton et al. 2012; United Nations 2011). Couples mainly adopt contraception as a means to stop childbearing. The most common method is vasectomy, which can be obtained easily and at no cost (Labrecque et al. 2005; Tuladhar 1987). In deciding when to stop childbearing, couples often value sons for their permanency in the natal home – as compared to daughters, who move to their husbands’ home upon marriage – and their role of caring for elderly parents (Bennett 1983; Cameron 1998). Still, even sons can present a cost to their parents, who must provide them with needed fees and supplies for their education (Caldwell 1982). Furthermore, mothers are responsible for providing childcare, even as they work in the fields to contribute to their household subsistence (Cameron 1998). These characteristics of childbearing and childrearing may play into fertility preferences at both the individual and community levels.

Identifying how family size preferences -- women’s and their neighbors’ – influence childbearing behavior has important implications in this context. Family planning initiatives were introduced to decrease both the desire for and achievement of large families (Thornton et al. 2012), and it is unclear whether remaining high fertility is due to women’s own childbearing desires, or to structural impediments to achieving their preferences for smaller families. Women with neighbors who prefer larger families may be motivated to have more children than women whose neighbors prefer smaller families. This pressure or desire to have a larger family is likely to lead women to have children at a faster rate.

Of course, beyond women’s own and their neighbors’ preferences, there are other confounding factors that could influence their progression to higher parity birth. For example, women’s access to contraception may limit their ability to implement their preferences. Women living in neighborhoods that are far from a health center, where contraception is distributed, may not be able to achieve a small family, even if that is what they or their neighbors prefer. In addition to neighbor-level confounders, women’s family size preferences may be confounded by their preference for sons. In this setting, couples typically prefer at least one son, so they may continue to have children until their goal for sons is reached (Dahal, Padmadas, and Hinde 2008; Leone, Matthews, and Zuanna 2003; Stash 1996), regardless of their or their neighbors’ overall family size preference.

We investigate the influences of neighbors’ preferences on transitions to higher parity births across the two cohorts of women, exploring whether older women followed their neighbors’ preferences more than younger women, and whether the influence of neighbors’ preferences decreased over time. Finally, we investigate whether the influence of neighbors operates through neighborhood socialization or through social pressure.

Hypotheses

We approach our analyses with three main hypotheses, following from the theoretical framework described above.

  1. Neighbors’ family size preferences will be related to individual childbearing behavior. Specifically, women whose neighbors prefer more children will have more higher parity births than women whose neighbors prefer smaller families.

  2. Neighbors’ preferences will have a greater influence on the behavior of older than of younger cohorts of women.

  3. Neighbors’ preferences will influence women independent of their own preferences.

Data

To test our hypotheses we use data from the Chitwan Valley Family Study (CVFS), conducted in rural Nepal. The data were collected from people in 171 neighborhoods, sampled from three strata of varying distance from the city. These neighborhoods typically consist of naturally occurring clusters of 5 to 15 households, surrounded by farmland. (Where a neighborhood consisted of more than 15 households, one contiguous section of the neighborhood was chosen.) The CVFS includes a baseline interview (averaging 72 minutes), consisting of a structured questionnaire and a semi-structured Life History Calendar interview, which was conducted in 1996. These interviews collected the information on both the attitudinal and experiential measures used here. A Neighborhood History Calendar was also used in 1996 to document characteristics of the neighborhoods, such as distance to the nearest health center. The CVFS interviewed all members, aged 15–59 and their spouses (even if outside this age range or living elsewhere), of every household in the sampled neighborhoods. The overall response rate for the survey was 97 percent. Monthly follow-up interviews were conducted with people in 151 neighborhoods, beginning in 1997.5 These monthly interviews collected information about household members on a range of demographic events, including childbearing. We use 147 months (12.25 years) of data from these interviews. This prospective design allows us to accurately model the influence of neighbors’ childbearing preferences on women’s subsequent behavior. The study is particularly well-suited for studying community influences on behavior, as it includes data from interviews with each resident in each of the sampled neighborhoods.

Our analytic sample consists of all women ages 15 to 34 in 1996 who were at risk of having another birth after their second or third live birth (N=594). Of these women, 446 fall into the younger cohort (ages 15–24 in 1996) and 148 fall into the older cohort (ages 25–34 in 1996). Our independent variable – a measure of family size preference – comes from the baseline study conducted in 1996, while the dependent variable for higher parity births comes from the monthly interviews that began in 1997.

Measures

Dependent variable

The dependent variable is a monthly time-varying dichotomous variable indicating whether the respondent had a higher parity birth. This variable is coded as 0 for every month up to the ninth month prior to the birth, and 1 in the ninth month prior to the birth. Respondents do not contribute to person-months of exposure to risk of birth for the eight months prior to the birth month and for three months after the birth.

Family Size Preference

We measure family size preference using the Coombs scale (Coombs 1974, 1979). This measure allows for variance in respondents’ reports of family size preference. This is useful because, as mentioned above, the majority of respondents in our sample stated a preference for two children. The Coombs scale allows us to differentiate between those respondents who want two children at most and those who want two children at least. Respondents were first asked “If you could have exactly the number of children you want, how many children would you want to have?” Next, respondents were asked how many children they would like to have if they could not have their first choice. (Respondents who already had children were asked how many children they would like to have if they could start life over.) Finally, they were asked how many children they would have if they could have neither of their first two choices. Originally, this item was coded on a scale of 1 to 25, as displayed in Figure 1. We have collapsed the item into 3 categories, since few neighbors fall below a Coombs scale value of 6 and few people fall above a value of 8. See Table 1 for descriptive statistics of neighbors’ and individual women’s family size preferences for each cohort.

Figure 1.

Figure 1

Open in a new tab

Coding scheme used for coombs scale measure of family size preference

Table 1.

Descriptive statistics

Covariates Mean Standard deviation Minimum value Maximum value
Family size preference
  Neighbor, cohorts combined 6.72 0.68 6.00 8.00
  Neighbor, younger cohort 6.72 0.70 6.00 8.00
  Neighbor, older cohort 6.71 0.61 6.00 8.00
  Individual, cohorts combined 6.41 0.68 6.00 8.00
Control variables
 Demographics
 Older cohort 0.25 0.43 0.00 1.00
 Age (first month of hazard) 24.76 3.61 17.00 38.00
 Brahmin/Chettri 0.49 0.50 0.00 1.00
 Dalit 0.09 0.29 0.00 1.00
 Newar 0.06 0.24 0.00 1.00
 Hill Indigenous 0.15 0.36 0.00 1.00
 Terai Indigenous 0.21 0.41 0.00 1.00
 Fertility experiences
 Parity (first month of hazard) 2.15 0.36 2.00 3.00
 Number of sons born 0.56 0.77 0.00 3.00
 Son Preference 2.78 0.86 1.00 4.00
 Number of kids that died 0.07 0.28 0.00 2.00
 Age at first birth 20.28 2.91 13.00 33.00
 Nonfamily experiences
 Years of education 5.04 3.94 0.00 14.00
 Ever had a wage labor job 0.45 0.50 0.00 1.00
 Ever lived away from family 0.07 0.25 0.00 1.00
 Ever member of youth club 0.03 0.17 0.00 1.00
 Sum of ever exposed to radio, TV, movie 2.76 0.55 0.00 3.00
 Perceived availability of contraception 0.91 0.20 0.00 1.67
 Neighborhood contraceptive use (mean) 0.44 0.17 0.00 1.00
 Health center within five-minute walk 0.26 0.44 0.00 1.00
 Length of exposure
 Duration of residence in neighborhood, years 7.98 6.92 0.00 30.00
 Time (first month of hazard) 40.73 41.62 0.00 142.00
Sample and dependent variable description
 Total women in sample 594
 Total births 229
 Proportion of women having third parity birth 0.26
 Proportion of women having fourth parity birth 0.12
 Proportion of women having third and fourth parity birth 0.05
Open in a new tab

For neighbors’ preferences, we employ the same measure and coding scheme. We constructed a neighborhood-level average from each neighbor sampled in each of the151 neighborhoods. Thus, the average childbearing preference for each neighborhood is constructed by summing the values of the measure for each resident and dividing by the number of residents. The respondent’s own preferences, household members’ preferences, and neighboring relatives’ preferences are not included in these averages; thus the neighborhood average differs for each respondent in each neighborhood. These measures are described in the following paragraphs.

Controls

To properly specify our models, we control for characteristics of the respondents that may influence both the family size preference and the likelihood of higher parity birth. First, to control for fecundity, we use a monthly time-varying covariate of respondents’ age (in years). Next, we control for ethnicity. Ethnicity in Nepal is complex, multifaceted, and related to religion. A full description of the ethnic groups residing in this setting is beyond the scope of this paper (for detailed descriptions of these groups see Acharya and Bennett 1981; Bennett 1983; Cameron 1997; Fricke 1986; Gellner and Quigley 1995; Guneratne 1994; Gurung 1980; MacFarlane 1976). We control for five classifications of ethnicity, coded as dummy variables, because of their different propensities to have large families: Brahmin/Chettri (high-caste Hindu), Dalit (low-caste Hindu), Newar, Terai Indigenous, and Hill Indigenous. Brahmin/Chettri ethnicity is the omitted category; influences of the other four groups are relative to this group.

Next, we control for respondents’ experiences with childbearing. We include a time-varying covariate for the respondent’s monthly parity status (i.e., whether they have had two or three live births). We also include a measure of the number of boys the respondent had given birth to as of 1996, as number of sons may influence respondents’ childbearing preferences as well as their subsequent childbearing behavior (Dahal et al. 2008). Many women will have an additional child in an attempt to reach their preferred number of sons – usually one or two – in spite of having met their family size preference. Therefore, we also control for respondents’ son preference. This preference measure comes from a survey item, specifically designed for this Nepalese population, in which respondents were asked to agree or disagree with a common Nepali phrase: “Yota aka, ke aka? Yota chora, ke chora?”. This roughly translates to “Having only one son is the same as having only one eye,” meaning it is good to have an extra son, just in case. Responses are coded from a scale of 1 to 4: strongly disagree, disagree, agree, and strongly agree.

Next, we include a measure to indicate the number of children the respondent gave birth to that subsequently died as of 1996. Experiencing the death of a child may motivate women to exceed their family size preference to help ensure that infant mortality does not cause them to fall short of their completed family size goal. We also control for respondents’ age at the time of their first birth, as women who began childbearing later in life may be inclined to speed the succession of their births.

Additionally, we control for respondents’ nonfamily experiences. Exposure to activities and ideas outside of the family home can influence the value women place on having a (large) family and their fertility experiences (Barber and Axinn 2004; Ghimire et al. 2006). We include a measure of the respondents’ accumulated years of education in 1996. We also include a dummy variable for whether respondents have ever had a wage labor job as of 1996, coded 0 to indicate that they never had a wage labor job and 1 to indicate that they ever had a wage labor job. Similarly, we include a variable for whether respondents ever lived away from their family as of 1996 (coded 1 for yes, 0 for no). We also include a dummy variable for whether respondents were ever a member of a youth club (coded 1 for yes, 0 for no). As a final indicator of exposure to nonfamily ideas we include a measure that is the sum of three dummy variables: ever listened to the radio, ever watched TV, and ever watched a movie.

In order to account for the potential confounding influence of access to contraception, we include three indicators of the respondents’ individual-level and neighborhood-level access. First, we control for respondents’ perceived access to contraception. We use the mean value of responses to eight survey items, asking whether it is easy (coded as 1) or difficult (coded as 0) to get eight different methods of contraception: birth control pills, IUD, Norplant, Depo Provera, foam, condom, male sterilization, female sterilization, and one ‘other’ method the respondent was given the opportunity to mention. Second, we control for an indicator of neighbors’ contraceptive use. This is coded as the mean value of whether neighbors (excluding household members and related neighbors) have ever used any of the eight methods of contraception. Third, we control for the distance of the neighborhood to the nearest health center. We use a measure that indicates whether the nearest health center is within a five-minute walk from respondent’s neighborhood, coded 1 if this is the case and 0 if it takes more than five minutes to walk from the neighborhood to the nearest health center.

To account for the length of exposure to members of women’s current neighborhood we control for the duration of respondent’s residence, in years, as of 1996. Finally, to account for the duration of the exposure to birth risk, we control for the time-varying duration of time, in months, since the first monthly interview.

Analytic Method

We use event history methods to model the risk of having a third or fourth birth with 147 months of data. Because the data are precise to the month, we use discrete-time methods to estimate these models, with person-months of exposure as the unit of analysis. We consider women to be at risk of a higher parity birth after they are married and have two or three children. Women are removed from the risk set during the months that they are not exposed to the risk of becoming pregnant with their third or fourth child. Women who have a third child are removed for the eight months following the first month of their pregnancy and for a three-month period of amenorrhea following the birth. Women pregnant with their fourth child are removed completely as of the eighth month prior to the birth. Sterilization is treated as a competing risk: women who are sterilized or whose husbands are sterilized after the start of the hazard cease to contribute to the person-months of exposure to risk of birth as of the first month of sterilization.6

We use logistic regression to estimate the discrete-time hazard models. The discrete time approach yields results similar to a continuous approach because the incidence of birth in any one month is quite low, but the discrete time approach allows us to avoid making any parametric assumptions regarding the distribution of the underlying baseline hazard (Yamaguchi 1991). Our time-varying measures of respondent characteristics are lagged by one month.

Women who are at risk of a third- or fourth-parity birth are included in our sample, allowing for repeatable events in the hazard. In the sample, 201 women had at least one birth (either third or fourth birth), and, of these, 28 women had two births (a third and fourth birth), for a total of 229 births. These data allow for parity variation both within and between individuals (Teachman 2011). The repeated birth events can introduce potential bias in the estimates. To account for this potential bias, we estimate three-level models: births nested within individuals, nested within neighborhoods. We use one-tailed tests of significance to investigate our unidirectional theory for the influence of family size preferences and two-tailed tests to investigate the influence of the control measures. We discuss the results as additive influences on the log odds of having a birth.

Results

Table 2 displays the relationship between neighbors’ preferences and the log-odds of a higher parity birth. In this table we test three hypotheses: that neighbors’ preferences influence women’s behavior, that the influence of neighbors is greater among the older cohort of women, and that the influence of neighbors’ is independent of women’s own preferences. Model 1 displays the model of neighbors’ preferences. This model does not support our first hypothesis: neighbors’ preferences are not significantly related to women’s parity progression, overall.

Table 2.

Log odds from logistic regression estimates of neighbors’ attitudes predicting the hazard of women’s higher parity birth (3rd or 4th birth) in rural Nepal

Model 1 Model 2 Model 3
Family size preference
 Neighbor † −0.33 (0.36) −0.73 (0.40) −0.74 (0.39)
 Neighbor * Older cohort † 2.01** (0.80) 2.25** (0.80)
 Individual † 1.12*** (0.32)
Controls
 Demographics
  Older Cohort −11.95* (5.50) −13.74* (5.49)
  Age 0.21*** (0.05) 0.13+ (0.07) 0.14* (0.07)
  Ethnic Group (ref: Brahnim/Chettri)
   Dalit −0.13 (0.83) 0.40 (0.85) 0.42 (0.84)
   Newar −0.42 (1.10) −1.47 (1.13) −0.47 (1.10)
   Hill Indigenous −1.71* (0.73) −1.65* (0.74) −1.54* (0.73)
   Terai Indigenous 1.03+ (0.71) 1.01 (0.72) 0.89 (0.71)
 Fertility Characteristics
  Parity −4.88*** (0.19) −4.93*** (0.19) −4.94*** (0.19)
  Number of sons born −0.42 (0.35) −0.57 (0.36) −0.61+ (0.36)
  Son Preference 0.03 (0.26) 0.02 (0.26) −0.01 (0.26)
  Number of children who died 3.33*** (0.71) 3.16*** (0.72) 3.11*** (0.72)
  Age at First Birth −0.31*** (0.09) −0.36*** (0.09) −0.38*** (0.09)
 Nonfamily Experiences
  Years of education −0.35*** (0.08) −0.34*** (0.08) −0.30*** (0.08)
  Ever had a wage labor job −0.74 (0.47) −0.67 (0.48) −0.79+ (0.48)
  Ever lived away from family 1.21 (0.86) 1.01 (0.88) 0.98 (0.87)
  Ever member of youth club 0.39*** (0.06) 0.39*** (0.06) 0.39*** (0.06)
  Sum of ever exposed to radio, TV, movie −0.47* (0.40) −0.49 (0.41) −0.46 (0.40)
 Contraceptive Access
  Perceived availability of contraception 0.97 (1.12) 0.93 (1.14) 1.06 (1.13)
  Neighborhood contraceptive use (mean) −0.13 (1.33) −0.39 (1.36) −0.26 (1.33)
  Health center within five-minute walk −1.31* (0.57) −1.26* (0.58) −1.11+ (0.57)
 Length of Exposure
  Duration of residence in neighborhood −0.10** (0.03) −0.11** (0.04) −0.10** (0.03)
  Time 0.02*** (0.00) 0.03*** (0.01) 0.03*** (0.01)
N (person-months) 37655 37655 37655
N (persons) 594 594 594
Open in a new tab

Note: Estimates are presented as log odds. Standard errors are given in parentheses.

†

denotes one-tailed tests, otherwise two-tailed tests were used,

+

p<.10

*

p<.05

**

p<.01

***

p<.001

In model 2 we interact neighbors’ preferences with each cohort. We find that neighbors’ family size preferences are related to older women’s parity progression, but have no significant relationship with younger women’s parity progression. Older women with neighbors who prefer larger families have more higher parity births. Specifically, women in the older cohort have 1.28 greater log odds of experiencing a higher parity birth for each one-unit increase in their neighbors’ family size preference. Model 2 offers support for the hypothesis that neighbors’ preferences are more strongly related to older women’s higher parity births than to younger women’s higher parity births.7

In Model 3, we add women’s individual preferences.8 Adding individual preferences to the model does not reduce the significant link between neighbors’ preferences and individual behavior, compared to Model 2. In Model 3, women of the older cohort have 1.51 greater log odds of experiencing a higher parity birth for each one-unit increase in their neighbors’ family size preference. Individuals’ own preferences are related to their parity progression behavior, even net of neighbors’ preferences, as well: women have 1.12 greater log odds of experiencing a higher parity birth for each one-unit increase in their family size preference. However, individual preferences for family size do not explain the relationship between neighbors’ preferences and individual behavior.

Many of the control measures are significantly related to women’s parity progression, as well. First, the baseline hazard, represented by age, is related to the risk of a higher parity birth. During the years in our analytic sample – ages 15 through 34 – women experience a linearly increasing risk of a higher parity birth. This is consistent with previous research on women in this age group.

Second, various group memberships are related to women’s risk of higher parity birth. Women of Hill Indigenous ethnicity have fewer higher parity births, relative to Brahmins/Chettris. Fertility experiences generally operate as would be expected – women with three children have fewer additional births than women with only two children. Additionally, women who have more sons and who were older at the time of their first birth have fewer higher parity births. Also consistent with expectations, women who experienced the death of a child had more higher parity births than women without this experience.

Nonfamily experiences do not consistently influence parity progression as expected. More educated women, women who ever worked for wages, and women with more media exposure have fewer high parity births than their counterparts, as would be expected. However, women who were ever members of a youth club have more high parity births – an unexpected result.

Discussion and Conclusion

We extend evidence of the well-established link between attitudes and behaviors to this non-Western context in rural Nepal. We move beyond the individual attitude-behavior link to the influence of neighbors’ preferences on behavior, and whether those neighbors’ preferences are net of women’s own preferences. Our results offer evidence that women respond to neighbors’ preferences. Specifically, neighbors’ family size preferences appear to independently, positively, and significantly influence women’s progression to larger families.

Neighbors’ influence on higher parity births emerge when examining these relationships separately by cohort. In fact, we only find a significant relationship between neighbors’ preferences among the older cohort of women. We suggest that this difference may be at least partly explained by the rapidly changing social context of Chitwan, Nepal, which has provided increasing access to opportunities, experiences, and ideas emanating from outside of the immediate community (Axinn and Yabiku 2001; Barber 2004). During their youth, and cumulatively throughout their life course, the younger cohort of women has had more exposure to these new experiences, while the older cohort was raised in a context in which there was less contact with people, organizations, or ideas outside of their immediate community. The daily social interactions of the older cohort were more confined to their own neighborhood, likely causing them to pay more attention to their neighbors’ attitudes and preferences. Over time and across cohorts, as the social context changed, family size preference may have become more individualized.

Furthermore, women’s preferences do not explain the relationship between neighbors’ preferences and women’s behavior. This supports the social pressure hypothesis: women may respond to neighbors’ preferences, regardless of what they, themselves, desire. The relationship neighbors’ family size preferences on behavior is both independent of and stronger than women’s own preferences. By extension, this does not support neighborhood socialization theory – that is, we find no evidence that neighbors’ preferences influence behavior via their influence on women’s preferences. When we include measures of women’s own preferences in our model, the relationship between neighbors’ preferences and individual behavior is not attenuated.

While this study suggests important neighbor influences on individual fertility behaviors, there are important limitations. Studies of the influence of neighbors on individuals inevitably face issues of selection and directionality of influence. In this Nepalese setting, selection may be a lesser issue, as there is very little mobility in housing (due to inheritance practices) and people are usually selected into their neighborhoods based on family lineage. However, it is the case that the direction of influence can go both ways: individuals have the same potential to influence their neighbors as their neighbors have to influence them. Additionally, there are a number of community-level factors that could explain the relationship that we attribute, in this paper, to neighbors’ preferences. We have attempted to account for many of these factors by including direct measures – for example, neighborhood-level distance to health clinics and neighbors’ contraceptive use – to account for individuals’ access to contraception. However, we cannot account for every potential confounder in the model, and therefore the results should be interpreted with caution. Additionally, the study that produced the data used here began more than 15 years ago. Thus, the assessments of individuals’ and neighbors’ preferences are from 1996. However, in order to assess the relationship between preferences and subsequent behavior throughout the childbearing years, it is necessary to have a relatively long period of observation. Fortunately, the data on preferences are combined with a household registry system that includes observations of more than a decade of subsequent fertility behavior, with observations into the year 2009. Nonetheless, although fertility preferences tend to be stable (Freedman, Coombs, and Bumpass 1965), preferences are subject to change (Krosnick and Alwin 1989), and we do not have measures of these changes. It is quite possible that, if we had time-varying information about neighbors’ preferences, the relationships in our models would be even stronger. Finally, this sample is limited to a single district in Nepal and is not generalizable to neighbor influences in other settings. Despite these issues, we hope that our analyses motivate further research on neighbor influences, and the potential decline of this influence over time.

These findings may be relevant to policymakers who aim to further decrease fertility in this region. Promoting favorable attitudes toward smaller families at the community-level may influence fertility at the individual level. But this community-level influence may diminish over time, as young women are increasingly exposed to other messages from outside of their communities. Investing in attitude transmission via schools and the media, for example, may have greater payoff for women who are just entering, or recently entered, their reproductive years.

Overall, these results may indicate a decline in collectivism in South Asia, perhaps particularly in rural areas where historically isolated people have been rather suddenly exposed to outside influences via increased access to cities, education, employment, and the media. Older women – whose exposure to experiences and ideas, prior to their reproductive years, was largely constrained to their own neighborhoods – tend to be very much influenced by their neighbors’ preference for family size. But for younger women, the influence of neighbors’ ideas, preferences, and attitudes may have been partially replaced by ideas that accompany the exposure to social life outside of the immediate community (Barber 2004). With these changes, neighborhood collectivism, socialization, and social pressure may have a decreasing influence on individual childbearing preferences and behavior. Instead, over time, women may respond more to infiltration of Western ideas and values, as their immediate surroundings and social interactions become only a fraction of the stimuli to which they are exposed.

Acknowledgments

This research was supported by the National Institute of Child Health and Human Development (grant numbers HD032912, HD041028, and HD007339) and the National Science Foundation (grant number OISE 0729709). We would like to thank the Institute for Social and Environmental Research in Chitwan, Nepal for collecting the data used here; William Axinn and Arland Thornton for helpful comments on an earlier version of this paper; and Cathy Sun and Daniel Thompson for assisting with data management. All errors and omissions remain the responsibility of the authors.

Footnotes

1

This percentage is based on data from the 1996 baseline Chitwan Valley Family Study (CVFS). The data consist of 5271 men and women between ages 15–59, and their spouses, who were living in 171 sampled neighborhoods.

2

Reported family size preference might be influenced by social desirability bias, particularly in this context, where the government has been promoting a two-child family for decades. However, achieved fertility in Nepal has yet to match the reported desired fertility (Dahal et al. 2008).

3

Due to the prevalence of son preference in Nepal, this may not be true for couples who have only daughters and continue to have births so that they can have a son (Dahal et al. 2008; Leone et al. 2003; Stash 1996). In these cases, couples may have intended to have only two children, but continue to give birth in the hopes of having a son.

4

This is not necessarily true in the case of couples who adopt contraception to defer a first birth or to space births, while still intending to have many children.

5

Twenty of the 171 neighborhoods sampled in the 1996 baseline CVFS were selected as an oversample for ethnic representation. Only the original 151 neighborhoods were followed for the monthly interviews. Therefore, we limit our investigation to these 151 neighborhoods.

6

We also ran the models including women who were sterilized or whose husbands were sterilized after the 1996 interview through the end of the hazard. Additionally, we ran separate models with a time-varying control for whether respondents were sterilized after 1996. The results were very similar to the results we obtain when we treat sterilization as a competing risk, though slightly diluted because women who had made the choice to remove themselves from the risk of birth were treated as continuing to be at risk. We believe it is appropriate to remove these women from the risk: a couple’s choice to become sterilized after 1996 is influenced by their attitudes in 1996, just as their choice to have a birth is influenced by their attitudes.

7

We also investigated interactions of attitudes with respondent demographics, fertility experiences, nonfamily experiences, and duration of residence in neighborhood. These did not reveal consistent significant influences.

8

Note that neighbors’ and individual women’s family size preference are correlated at only 0.07.

Contributor Information

Elyse A. Jennings, University of Michigan

Jennifer S. Barber, University of Michigan

References

  1. Ajzen Icek. Attitudes, Personality, and Behavior. Chicago, IL: Dorsey Press; 1988. [Google Scholar]
  2. Acharya Meena, Bennett Lynn. The Status of Women in Nepal: The Rural Women of Nepal. Kathmandu, Nepal: Tribhuvan University; 1981. [Google Scholar]
  3. Axinn William G, Thornton Arland. Mothers, children, and cohabitation: the intergenerational effects of attitudes and behavior. American Sociological Review. 1993;58(2):233–246. [Google Scholar]
  4. Axinn William G, Thornton Arland. The influence of parents’ marital dissolutions on children’s attitudes toward family formation. Demography. 1996;33(1):66–81. [PubMed] [Google Scholar]
  5. Axinn William G, Yabiku Scott T. Social change, the social organization of families, and fertility limitation. The American Journal of Sociology. 2001;106(5):1219–1261. [Google Scholar]
  6. Barber Jennifer S. Intergenerational influences on the entry into parenthood: mothers’ preferences for family and nonfamily behavior. Social Forces. 2000;79(1):319–48. [Google Scholar]
  7. Barber Jennifer S. Ideational influences on the transition to parenthood: attitudes toward childbearing and competing alternatives. Social Psychology Quarterly. 2001;64(2):101–127. [Google Scholar]
  8. Barber Jennifer S. Community social context and individualistic attitudes toward marriage. Social Psychology Quarterly. 2004;67(3):236–256. [Google Scholar]
  9. Barber Jennifer S, Axinn William G. New ideas and fertility limitation: the role of mass media. Journal of Marriage and Family. 2004;66(5):1180–1200. [Google Scholar]
  10. Barber Jennifer S, East Patricia L. Home and parenting resources available to siblings depending on their birth intention status. Child Development. 2009;83(3):921–939. doi: 10.1111/j.1467-8624.2009.01306.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Barber Jennifer S, East Patricia L. Children’s experiences after the unintended birth of a sibling. Demography. 2011;48(1):101–125. doi: 10.1007/s13524-010-0011-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Behrman Jere R, Kohler Hans-Peter, Watkins Susan Cotts. Social networks and changes in contraceptive utilize over time: evidence from a longitudinal study in rural Kenya. Demography. 2002;39(4):713–738. doi: 10.1353/dem.2002.0033. [DOI] [PubMed] [Google Scholar]
  13. Bengtson Vern L. Generation and family effects in value socialization. American Sociological Review. 1975;40:358–71. [Google Scholar]
  14. Bennett Lynn. Dangerous Wives and Sacred Sisters: Social and Symbolic Roles of High-Caste Women in Nepal. New York, NY: Columbia University Press; 1983. [Google Scholar]
  15. Blake Judith. Family size and the quality of children. Demography. 1981;18(4):421–442. [PubMed] [Google Scholar]
  16. Bongaarts John, Watkins Susan C. Social interactions and contemporary fertility transitions. Population and Development Review. 1996;22(4):639–682. [Google Scholar]
  17. Cameron Mary M. On the Edge of Auspicious: Gender and Caste in Nepal. Chicago, IL: University of Chicago Press; 1998. [Google Scholar]
  18. Caldwell John C. Theory of Fertility Decline. London: Academic Press; 1982. [Google Scholar]
  19. Caldwell John C. The globalization of fertility behavior. Population and Development Review. 2001;27:93–115. Supplement: Global Fertility Transition. [Google Scholar]
  20. Casterline John B. Diffusion processes and fertility transition: introduction. In: Casterline JB, editor. Diffusion Processes and Fertility Transition. Washington, DC: National Academy Press; 2001. pp. 1–38. [PubMed] [Google Scholar]
  21. Cleland John G. Potatoes and pills: an overview of innovation-diffusion contributions to explanations of fertility decline. In: Casterline JB, editor. Diffusion Processes and Fertility Transition. Washington, DC: National Academy Press; 2001. pp. 1–38. [Google Scholar]
  22. Cleland John G, Sathar Zeba A. The effect of birth spacing on childhood mortality in Pakistan. Population Studies. 1984;38(3):401–418. doi: 10.1080/00324728.1984.10410300. [DOI] [PubMed] [Google Scholar]
  23. Coale Ansley J. The demographic transition reconsidered. International Population Conference; Liege. 1973; Liege: IUSSP; 1973. [Google Scholar]
  24. Coleman James. Foundations of Social Theory. Cambridge: Harvard University Press; 1990. [Google Scholar]
  25. Coombs Lolagene C. The measurement of family size preferences and subsequent fertility. Demography. 1974;11(4):587–611. doi: 10.2307/2060472. [DOI] [PubMed] [Google Scholar]
  26. Coombs Lolagene C. Reproductive goals and achieved fertility: a fifteen-year perspective. Demography. 1979;16:523–534. [PubMed] [Google Scholar]
  27. Curtis Sian L, Diamond Ian, McDonald John W. Birth interval and family effects on postneonatal mortality in Brazil. Demography. 1993;30(1):33–43. [PubMed] [Google Scholar]
  28. Dahal Govinda P, Padmadas Sabu S, Andrew Hinde PR. Fertility-limiting behavior and contraceptive choice among men in Nepal. International Family Planning Perspectives. 2008;34(1):6–14. doi: 10.1363/ifpp.34.006.08. [DOI] [PubMed] [Google Scholar]
  29. Desai Sonalde. When are children in large families disadvantaged? Evidence from cross-national analysis. Population Studies. 1995;49:195–210. [Google Scholar]
  30. Eggleston Elizabeth. Determinants of unintended pregnancy among women in Ecuador. International Family Planning Perspectives. 1999;25(1):27–33. [Google Scholar]
  31. Fishbein Martin, Ajzen Icek. Beliefs, Attitude, Intention, and Behavior: An Introduction to Theory and Research. Reading, MA: Addison-Wesley; 1975. [Google Scholar]
  32. Fishbein Martin, Ajzen Icek. Predicting and Changing Behavior: The Reasoned Action Approach. New York: Psychology Press; 2010. [Google Scholar]
  33. Freedman Ronald, Coombs Lolagene C, Bumpass Larry. Stability and Change in Expectations about Family Size: A Longitudinal Study. Demography. 1965;2:250–275. [Google Scholar]
  34. Freedman Ronald, Coombs Lolagene, Chang Ming-Cheng, Sun Te-Hsiung. Trends in fertility, family size preferences, and practice of family planning: Taiwan, 1965–1973. Studies in Family Planning. 1974;5(9):270–288. [PubMed] [Google Scholar]
  35. Fricke Thomas E. Himalayan Households: Tamang Demography and Domestic Processes. Ann Arbor, MI: UMI Research Press; 1986. [Google Scholar]
  36. Gellner David N, Quigley Declan., editors. Contested Hierarchies: A Collaborative Ethnography of Caste among the Newars of the Kathmandu Valley, Nepal. New York: Oxford University Press; 1995. [Google Scholar]
  37. Ghimire Dirgha J, Axinn William G, Yabiku Scott A, Thornton Arland. Social change, premarital non-family experience and spouse choice in an arranged marriage society. American Journal of Sociology. 2006;111(4):1181–1219. [Google Scholar]
  38. Guneratne Upali Arjun. Doctoral Dissertation. University of Chicago; 1994. The Tharus of Chitwan: Ethnicity, Class and the State in Nepal. [Google Scholar]
  39. Gurung Harka B. Vignettes of Nepal. Kathmandu, Nepal: Sajha Prakashan; 1980. [Google Scholar]
  40. Hornik Robert, McAnany Emile. Mass media and fertility change. In: Casterline J, editor. Diffusion Processes and Fertility Transition. Washington: National Academy Press; 2001. pp. 208–239. [Google Scholar]
  41. Horton Susan. Child nutrition and family size in the Philippines. Journal of Development Economics. 1986;23(1):161–176. [Google Scholar]
  42. Katz Jennifer, Joiner Thomas E, Jr, Kwon Paul. Membership in a devalued social group and emotional well-being: developing a model of personal self-esteem, collective self-esteem, and group socialization. Sex Roles. 2002;47(9/10):419–431. [Google Scholar]
  43. Kessler Daniel. Birth order, family size, and achievement: family structure and wage determination. Journal of Labor Economics. 1991;9(4):413–426. doi: 10.1086/298275. [DOI] [PubMed] [Google Scholar]
  44. Kidwell Jeannie S. Number of siblings, sibling spacing, and birth order: their effects on perceived parent-adolescent relationships. Journal of Marriage and the Family. 1981;43(2):315–332. [Google Scholar]
  45. Knodel John. Family size and children’s education in Thailand: evidence from a national sample. Demography. 1991;28(1):119–131. [PubMed] [Google Scholar]
  46. Knodel John, Havanon Napaporn, Sittitrai Werasit. Family size and the education of children in the context of rapid fertility decline. Population and Development Review. 1990;16(1):31–62. [Google Scholar]
  47. Krosnick Jon A, Alwim Duane F. Aging and susceptibility to attitude change. Journal of Personality and Social Psychology. 1989;57(3):416–425. doi: 10.1037//0022-3514.57.3.416. [DOI] [PubMed] [Google Scholar]
  48. Labrecque Michel, Pile John, Sokal David, Kaza Ramachandra CM, Rahman Mizanur, Bodh SS, Bhattarai Jeewan, Bhatt Ganesh D, Vaidya Tika Man. Vasectomy techniques in South and South East Asia. BMC Urology. 2005;5(10) doi: 10.1186/1471-2490-5-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Leone Tiziana, Matthew Zoë, Zuanna Gianpiero Dalla. Impact and determinants of sex preference in Nepal. International Family Planning Perspectives. 2003;29(2):69–75. doi: 10.1363/ifpp.29.069.03. [DOI] [PubMed] [Google Scholar]
  50. Lesthaeghe Ron L. The decline of Belgian fertility, 1800–1970. Princeton: Princeton University Press; 1978. [Google Scholar]
  51. Lesthaeghe Ron J, Vanderhoeft Camille. Ready, willing, and able: a conceptualization of transitions to new behavioral forms. In: Casterline JB, editor. Diffusion Processes and Fertility Transition. Washington: National Academy Press; Washington, D.C: National Academy Press; 2001. pp. 240–264. [Google Scholar]
  52. MacFarlane Alan. Resources and population: a study of the Gurungs of Nepal. Cambridge, MA: Cambridge University Press; 1976. [Google Scholar]
  53. Marston Cicely, Cleland John. Do unintended pregnancies carried to term lead to adverse outcomes for mother and child? An assessment in five developing countries. Population Studies. 2003;57(1):77–93. doi: 10.1080/0032472032000061749. [DOI] [PubMed] [Google Scholar]
  54. Moursund Anne, Kravdal Øystein. Individual and community effects of women’s education and autonomy on contraceptive use in India. Population Studies. 2003;57(3):285–301. doi: 10.1080/0032472032000137817. [DOI] [PubMed] [Google Scholar]
  55. Population Reference Bureau. World Population Data Sheet. Washington, DC: Population Reference Bureau; 2011. [Google Scholar]
  56. Powell Brian, Steelman Lala Carr. Beyond sibship size: sibling density, sex composition, and educational outcomes. Social Forces. 1990;69(1):181–206. [Google Scholar]
  57. Rutenberg Naomi, Watkins Susan Cotts. The buzz outside the clinics: conversations and contraception in Nyanza Province, Kenya. Studies in Family Planning. 1997;28(4):290–307. [PubMed] [Google Scholar]
  58. Stash Sharon. Ideal-family-size and sex-composition preferences among wives and husbands in Nepal. Studies in Family Planning. 1996;27(2):107–118. [PubMed] [Google Scholar]
  59. Stuber Jennifer, Galea Sandro, Link Bruce G. Smoking and the emergence of a stigmatized social status. Social Science & Medicine. 2008;67(3):420–430. doi: 10.1016/j.socscimed.2008.03.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Teachman Jay. Modeling Repeatable Events Using Discrete-Time Data: Predicting Marital Dissolution. Journal of Marriage and Family. 2011;73:525–540. [Google Scholar]
  61. Thornton Arland. Readings History Sideways: The Fallacy and Enduring Impact of the Developmental Paradigm on Family Life. Chicago: The University of Chicago Press; 2005. [Google Scholar]
  62. Thornton Arland, Binstock Georgina, Yount Kathryn, Abbasi-Shavazi Mohammad Jalal, Ghimire Dirgha, Xie Yu. International fertility change: new data and insights from the developmental idealism framework. Demography. 2012 doi: 10.1007/s13524-012-0097-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Troyer Lisa, Wesley Younts C. Whose expectations matter? The relative power of first- and second-order expectations in determining social influence. American Journal of Sociology. 1997;103:692–732. [Google Scholar]
  64. Tuladhar Jayanti Man. Effect of family planning availability and aceessibility on contraceptive use in Nepal. Studies in Family Planning. 1987;18(1):49–53. [PubMed] [Google Scholar]
  65. United Nations, Department of Economic and Social Affairs, Population Division. World Population Prospects: The 2010 Revision. New York: United Nations; 2011. [Google Scholar]
  66. Watkins Susan Cotts. The fertility transition: Europe and the third world compared. Sociological Forum. 1987;2(4):645–673. [Google Scholar]
  67. Winnikoff Beverly. The effects of birth-spacing on child and maternal health. Studies in Family Planning. 1983;14:231–245. [PubMed] [Google Scholar]
  68. Yabiku Scott. The effect of non-family experiences on age of marriage in a setting of rapid social change. Population Studies. 2005;59(3):339–354. doi: 10.1080/00324720500223393. [DOI] [PubMed] [Google Scholar]
  69. Yamaguchi Kazuo. Event History Analysis. Newbury Park, CA: Sage Publications, Inc; 1991. [Google Scholar]
  70. Zajonc Robert B, Mullally Patricia R. Birth order: reconciling conflicting effects. American Psychologist. 1997;52(7):685–699. [Google Scholar]

RESOURCES