Abstract
Objective:
This article analyzes the relationship between educational aspirations and fertility aspirations early in the life course in three different settings.
Background:
The negative relationship between women’s educational attainment and childbearing is one of the most consistent associations in social science. Family scholars have a more limited understanding of the relationship between educational aspirations and fertility aspirations before childbearing or union formation.
Method:
The authors use data collected in Jalisco, Mexico; Gaza, Mozambique; and Chitwan Valley, Nepal as part of the Family Migration and Early Life Outcomes project. They estimate nested Poisson regressions to model the relationship between adolescent educational aspirations and desired family size, controlling for individual- and household-level sociodemographic variables as well as adolescent beliefs and values.
Results:
On average, adolescents who desire more education want fewer children in unadjusted models. In Mozambique and Nepal, this association is attenuated in models accounting for household characteristics. In Mexico, the association persists after incorporating these factors, but the inclusion of individual aspirations attenuates the relationship between educational aspirations and desired family size. In Mozambique, the association of educational aspirations with desired family size is moderated by gender.
Conclusion:
As young people enter adolescence, their desires for education and childbearing are inversely related, but the mechanisms driving this association vary across contexts. This variation may be related to linkages between education, social status, and family values.
Keywords: Fertility, desired family size, education, adolescence, social context, cross-national
The negative relationship between women’s educational attainment and childbearing is one of the most consistent associations in the social sciences: it is found in rich and poor countries, in low and high fertility contexts, across urban and rural areas and across more than a century of available demographic data. Although data on men’s fertility is less extensive, the education-fertility relationship is negative for men as well in many contexts. Multiple explanations have been proposed for the negative relationship between education and childbearing, including incompatibility between childbearing and school enrollment or the kind of jobs more educated people are likely to have; increased autonomy or efficacy related to education (especially for women); and changing goals and values as a result of education (Axinn & Barber, 2001; Basu, 2002; Musick et al., 2009). The negative education-fertility relationship may also be partially due to selection into education, as people who choose to obtain high levels of schooling may already have been inclined to have smaller families (Brand & Davis, 2011).
The relationship between completed education and completed fertility results from a sequence of interconnected decisions and experiences that evolve across multiple stages of the life course. Yet we have limited understanding of how these decisions occur during the transition to adulthood, when key decisions about schooling and the timing and sequencing of family formation behaviors are made. Studying intentions, plans, and desires at specific life course stages can shed light on how people make decisions at these junctures and the relative importance of individual preferences and structural constraints in shaping outcomes. But existing evidence on the role of preferences in explaining the education-fertility relationship is mixed. Some studies find that people with more education want fewer children, while other show similar fertility aspirations across education levels (Behrman, 2015; Bongaarts, 2010; Morgan & Rackin, 2010; Quesnel-Vallée & Morgan, 2003; Testa & Stephany, 2017). Few studies consider the relationship between educational aspirations and fertility aspirations in adolescence before childbearing or union formation begins.
In this paper, we consider the relationship between educational aspirations and fertility aspirations early in the life course, drawing on data from three very different contexts and including both boys and girls to understand possible variation in this relationship. We use unique data collected in Jalisco province, Mexico; Gaza province, Mozambique; and Chitwan Valley, Nepal as part of the Family Migration and Early Life Outcomes (FAMELO) project. We use nested Poisson regressions to model the relationship between adolescent educational aspirations and desired family size, controlling for individual- and household-level sociodemographic variables as well as adolescents’ attitudes and values. We find that educational aspirations and desired family size are negatively correlated, but associations vary across contexts and differ for boys and girls. Documenting these associations sheds light on the way adolescents understand the meaning of education and childbearing and potential conflict or complementarity between them. Because adolescent aspirations shape their decisions about schooling and family formation, studying adolescent aspirations also contributes to knowledge about socioeconomic and gender inequality in these outcomes.
THE PERSISTENT ASSOCIATION BETWEEN EDUCATION AND FERTILITY
In general, women with higher levels of education have fewer children than those with less education (Axinn & Barber, 2001; Bongaarts, 2003; Kravdal, 2002; Kravdal & Rindfuss, 2008; Lutz & Samir, 2011; Martin, 1995; Skirbekk, 2008; Yoo, 2014), although this relationship may be reversing at higher parities in some low fertility contexts (Kravdal, 1992; Testa, 2014). The association between education and fertility is less consistent for men. In some settings where education is associated with higher or more stable earnings and men with stable earnings are more attractive partners, men who are more highly educated have more children (Kravdal & Rindfuss, 2008; Nisén et al., 2018). In other contexts, however, education is negatively correlated with fertility for men as well as for women (DeRose & Ezeh, 2005; Lakomý, 2017). And, across contexts, being enrolled in school is a strong negative predictor of short-term fertility for both men and women, and both men and women with more education have later first births than their less-educated counterparts (Dribe & Stanfors, 2008; Glick et al., 2006; National Research Council and Institute of Medicine, 2004; Rindfuss et al., 1996).
Many of the explanations for these associations are grounded in economic arguments, i.e., an argument that education raises the costs (broadly construed) of bearing and raising children, thus leading parents to have fewer of them. Education may increase parents’ aspirations for their own children’s education, making raising children more expensive (Axinn & Barber, 2001; Caldwell, 1976; Lareau, 2002). Education may also indirectly raise the costs of children by increasing the value of women’s time in paid labor and thus the opportunity costs of childbearing (Becker, 1981). Yet the consistency of the negative association between education and childbearing across settings with different levels of economic development and educational expansion suggests that ideational factors may also play an important role.
Education carries social meanings as well as economic value. Young people and their families across socioeconomic backgrounds and contexts aspire to high levels of education as means of monetary and moral worth (Detering, 2015; Frye, 2012; Nielsen, 2015; Silva & Snellman, 2018). Formal schooling is linked to cultural schemas incorporating individualism, autonomy, and secularism (Baker, 2014; Johnson-Hanks, 2005; Lesthaeghe, 1983). The association between formal education and individualist values and ideals may be particularly salient in the Global South, where education is also linked to schemas of modernization and economic development (Thornton, 2004). Parenting, in contrast, is associated with values of familism and collectivism (Hayford & Morgan, 2008; Hays, 1998; Lesthaeghe, 1983; Shepherd & Marshall, 2018; Smith, 2001) – values that are often understood in opposition to individualist values. This ideational opposition may contribute to the negative association between education and fertility. Implicit cognitive associations, such as a negative association between the world of work and the world of family, shape attitudes and values and, in turn, influence decision-making about childbearing and other demographic outcomes (Bachrach & Morgan, 2013; Johnson-Hanks et al., 2011; Shepherd & Marshall, 2018).
ADOLESCENT ASPIRATIONS FOR EDUCATION AND CHILDBEARING
Aspirations may differ from intentions (concrete plans for action) or expectations (what people think will actually happen) if people perceive constraints that prevent them from achieving their desires or if they have mixed or ambivalent desires (Fishbein & Ajzen, 2009; Miller, 1994). Nonetheless, adolescent preferences for marriage, childbearing, education, migration, and work help predict later outcomes (B. L. Barber et al., 2001; Eccles et al., 2004; Quesnel-Vallée & Morgan, 2003; Sassler & Schoen, 1999; Schoon, 2001; Sewell et al., 1969). Aspirations reflect individual values but are also a product of shared social and cultural norms and thus incorporate both structural and ideational constraints (Fishbein & Ajzen, 2009; Frye, 2012; Mische, 2009). For example, an ideal family size of two children is widespread in industrialized countries (Sobotka & Beaujouan, 2014) in part because social, economic, and family systems make it challenging to have larger families, and individuals incorporate these challenges into their ideals. Studying individual aspirations can be a way of understanding shared cultural ideas and shared understandings of structural constraints and can shed light on how people make decisions at different life course stages.
Educational aspirations are an interaction of perceived opportunities and individual self-efficacy or agency (Hartas, 2016; Roubeni et al., 2015). In low- and middle-income countries, such as the ones where this study takes place, family background, including parental education and household socioeconomic status, is a strong predictor of children’s educational aspirations (Altamirano et al., 2010; Chiapa et al., 2016; Tsung & Gao, 2012). Educational aspirations also reflect shared values (Frye, 2012). Family size preferences are informed by social norms and expectations about childbearing (Ryder, 1973, 2010; Sobotka & Beaujouan, 2014). They vary according to family background characteristics, most notably sibship size, and individual values such as religiosity (Hayford & Morgan, 2008; Morgan & Rackin, 2010; Regnier-Loilier & Depledge, 2006). More broadly, desired family size is a marker of aspirations within the family domain (Hayford & Morgan, 2008; Marshall & Shepherd, 2018; Thornton et al., 2012).
THE RELATIONSHIP BETWEEN EDUCATION ASPIRATIONS AND FERTILITY ASPIRATIONS
Research on the association between educational attainment and fertility desires or intentions finds a much less consistent relationship than that found between education and achieved fertility. In low- and middle-income countries, more educated women tend to want fewer children, although the strength of this association varies (Behrman, 2015; Bongaarts, 2010; Kebede et al., 2021; Martin, 1995). But studies in low-fertility contexts report a null or even positive relationship between education and fertility intentions (Musick et al., 2009; Nitsche & Hayford, 2020; Testa, 2014). Testa (2014) speculates that this may be a result of social systems that make it easier for more educated women to access paid family leave and childcare.
A much more limited literature has focused on the relationship between educational aspirations and fertility aspirations. Fertility goals change over the life course; in particular, people tend to reduce their desired or intended family size as they obtain more education (Hayford, 2009; Yeatman et al., 2013). Early childbearing can both constrain later educational opportunities (Hofferth et al., 2001; Raymo et al., 2015) and alter goals for future childbearing (Yeatman et al., 2013). As such, selection into higher education levels may contribute to the relationship between educational attainment and fertility aspirations. We focus on aspirations for both education and fertility early in the life course to understand the ideational relationship between goals in these domains before young people have direct experience of higher education or parenting. We expect there may be variation in this relationship across contexts both because of differences in (perceived or actual) conflict between education and childbearing and because of differences in the cultural associations between education and childbearing.
It is possible that there is no association between educational aspirations and childbearing aspirations or that any observed association between aspirations in these two domains is the result of correlation with some other characteristic. For example, children with more educated parents may have both more ambitious educational aspirations and smaller desired family size. Further, some research suggests that both educational attainment and fertility outcomes are partly genetically determined (Nisén et al., 2013); aspirations in these domains may also be determined in part by intrinsic personality factors. In our analyses, we control for a set of observable factors, but cannot fully account for confounding due to unobserved factors.
It is also possible that education and childbearing conflict, or are perceived to conflict, in ways that affect aspirations in these two domains. These conflicts depend on social systems and institutions, and thus we would expect variation across contexts. Strong norms and institutional structures discourage childbearing while enrolled in school, especially for girls (Frye, 2017; Luker, 1997; Mollborn & Jacobs, 2011). These norms and structures are widely observed around the world, although their strength varies. In contexts where there are also strong norms about the timing of marriage and family formation, it may be more difficult to combine ambitious educational aspirations with a desire for a large family. Educational aspirations may also be tied to goals for employment and career achievement. In contexts without social structures that support working parents, it can be difficult to combine work and parenting, and thus high educational aspirations may be understood as incompatible with high desired family size.
Educational aspirations and desired family size may also be negatively associated if aspirations reflect cultural values associated with education and childbearing that imply opposition between those domains. Two widely-held cultural models promote this opposition. The first is the breadwinner-homemaker model of the family. This model, originating during the 19th century in Europe and its settler colonies, was grounded in the growth of the industrial economy, which brought economic production out of the home and separated the world of work from the domestic sphere (Coontz, 1993; Hays, 1998). According to this model, family responsibilities should be clearly divided along gendered lines, with fathers responsible for financial provision and work outside the home and mothers responsible for raising children and work inside the home. Although this division of labor has rarely, if ever, been common in practice, it remains a culturally salient idea in much of the industrialized (and post-industrial) West (Blair-Loy, 2003; Collins, 2019; Gerson, 2009; Hays, 1998; Townsend, 2002). Because the breadwinner-homemaker model emphasizes a divide between domestic responsibilities and roles outside the home, this model may also imply conflict between educational aspirations and fertility aspirations, particularly for women.
The second cultural model opposing education and childbearing is the idea that education is “modern”, while having large families is not. Thornton (2004) proposed the label of “developmental idealism” (DI) for a set of beliefs that connects economic development, individual education, and a westernized nuclear family system as mutually causal. These ideas are introduced into less-developed settings by colonization, economic exchange, and cultural products such as movies and television, and they imply a connection between economic change, political change, and family change. According to the DI framework, as proposed by Thornton and elaborated in a series of studies in multiple contexts across the Global South (Allendorf & Thornton, 2019; Thornton et al., 2012, 2014), people who subscribe more strongly to DI beliefs will be likely to desire both more education and smaller families, because both of these goals are associated with economic progress according to DI schemas. Notably, the DI framework does not predict differences in this relationship by gender – both men and women subscribe to DI beliefs, and gender equality within the family is a component of DI beliefs.
The breadwinner-homemaker model and developmental idealism suggest slightly different patterns of association between educational aspirations and desired family size. As outlined above, according to the breadwinner-homemaker model, education and parenting are incompatible for women, but not for men, while DI ideas link education and small families for everyone. In addition, these two models may not be equally salient in all parts of the world. Because the breadwinner-homemaker model assumes a physical separation between work and family, it may be less influential in parts of the world where economic production relies on subsistence agriculture or other informal labor. DI ideals seek to explain the process of economic development and are less likely to guide aspirations and behavior in contexts that have already made the transition to an industrialized economy.
CURRENT STUDY
In this paper, we seek to better understand the ideational dimensions of the education-fertility relationship by examining the relationship between desires in these two domains among adolescents (ages 11–17) in three different contexts. We assess differences in this relationship between boys and girls. The transition to mass education and the expansion of schooling across less developed settings is associated with lower fertility and lower desired family size (Axinn & Barber, 2001; Behrman, 2015; Bongaarts et al., 2017; Caldwell, 1980). But the extant literature provides less guidance on how individual adolescents assess their options, see educational opportunities as complementary or competing with other domains, and set education and fertility goals across a variety of social, economic, and developmental contexts. The degree to which adolescents perceive conflict between education and childbearing may vary depending on the availability of paid work outside the home (for both men and women), the availability of support for combining work outside the home and childcare responsibilities, the accessibility of education (including gender differences in access), and other factors.
In the absence of existing theory or evidence identifying specific elements of social or cultural context, we take a case study approach, analyzing three distinct settings to generate data for future theory-building. By choosing settings that exhibit variation in schooling levels, fertility rates, economic systems, and gender and family systems, we attempt to maximize the possible variation in observed outcomes and associations, providing a sense of the range of possible relationships between educational and fertility aspirations. This follows the conceptual approach of the diverse case selection method described by Seawright and Gerring (2008).
Settings
The current study draws on data from a household-based survey that included interviews with children and adolescents from Jalisco, Mexico; Gaza, Mozambique; and Chitwan Valley, Nepal. The data come from the Family Migration and Early Life Outcomes (FAMELO) project, a project designed for comparative research on children and adolescents in settings with significant variation in social and economic development. All three focal settings have experienced rapid social and economic development, including educational expansion and new infrastructure (roads, transport networks, etc.).
The three settings provide variation in levels of development and changing family dynamics in which to investigate the connection between educational aspirations and desired family size. Based on the United Nations’ Human Development Index (HDI), a summary measure of health and well-being (with a scale of 0 to 1), Mexico is the most developed with an HDI of 0.767; Nepal follows with an HDI of 0.579; and Mozambique is comparatively the least developed with an HDI of 0.446 (United Nations Human Development Reports, 2018).
Educational aspirations may have different meanings in contexts where opportunities are constrained or where there are gendered norms about who should receive an education. Across our three sites, secondary enrollment is highest in Mexico, somewhat lower in Nepal, and lowest in Mozambique. In Mexico and Nepal, secondary enrollment is higher for girls than for boys. Gross secondary education enrollment rates are 93% for boys and 108% for girls in Mexico; 37% for boys and 33% for girls in Mozambique; and 67% for boys and 75% for girls in Nepal (World Bank, 2019). Note that gross enrollment ratios greater than 100% imply enrollment of “overage” children, that is, children who are older than the expected age for grade because of grade repetition or late entry into school.
Gender differentiation in family roles and the public domain also vary. Overall, Nepal has the lowest female labor force participation rate (26.3% of women 15 and older), Mexico follows (43.5%), and Mozambique has the highest female labor participation rate (79.8%) (World Bank, 2019). Much of this work, however, is in the informal sector or in subsistence agriculture.
In conjunction with educational expansion and increased access to health care including contraception, the total fertility rate (TFR) in all countries has been steadily declining since the early-to-mid 1970s. While the TFR in Mexico in 1970 was 6.8 births per woman, the current TFR in Mexico is 2.2 births per woman. The total fertility rate for Mozambique has also been on the decline during this window, although the change has been slower. The TFR for Mozambique in 1970 was 6.7 births per woman, while the current TFR is 4.9 births per woman (World Bank, 2019). The corresponding changes in TFR for Nepal were 5.9 average births per woman in 1970, compared with the current average of 2.0 births per woman. Differences in desired family size across the three countries are largely consistent with differences in TFR. Among women age 15–19, the average desired family size is 2.1 in Mexico (Instituto Nacional de Estadística y Geografía (INEGI), 2018), 1.9 in Nepal (Ministry of Health, Nepal et al., 2017), and 3.8 in Mozambique (Ministerio da Saude (MISAU) et al., 2013).
In summary, our settings provide unique case studies with distinct combinations of economic development, educational expansion, and relative gender parity in schooling.
Analytic Approach and Hypotheses
This paper asks whether higher educational aspirations are consistently associated with lower desired family size and considers which individual, family, and household characteristics may be particularly relevant to understanding the connection between both educational and fertility aspirations. Our primary focus is not on baseline levels of educational or fertility aspirations, but rather on the relationship between these two domains at the individual level across settings.
We analyze the degree to which educational aspirations predict fertility aspirations in a regression framework with desired family size as the outcome. In reality, adolescent preferences in these two domains are likely to be jointly determined and mutually dependent. We examine educational aspirations as the predictor because, for adolescents, educational aspirations are relatively short-term and require actions taken in the near future (Alexander & Cook, 1979; Glick & White, 2004; Perreira & Spees, 2015), while fertility aspirations represent preferences over a longer time horizon. Further, schooling has already begun for all adolescents in our sample, whereas fewer adolescents in the data have initiated romantic relationships, union formation, or sexual debut. Because adolescents have more direct experience with education than with family formation, educational aspirations are more likely to have been shaped by direct experience and knowledge than fertility aspirations.
We hypothesize that adolescent educational aspirations are negatively correlated with desired family size (H1). Some of this negative association may be attributable to household and family characteristics that influence aspirations in both domains. We further hypothesize that some of the association is due to perceived conflict between education and childbearing and so will persist net of household and family controls (H1a). If this association is attributable, in part, to shared cultural schemas about the meaning of education and parenthood, then measures of children’s own beliefs and values in family-related domains should serve as mechanisms connecting aspirations in these two domains, because these beliefs and values are also potentially part of shared schemas about the meaning of family and education. We hypothesize that accounting for beliefs and values will partially account for any negative association between educational aspirations and desired family size (H1b).
Boys and girls may anticipate different family roles and responsibilities and thus different potential conflicts between education and childbearing. We hypothesize that any negative association between educational aspirations and desired family size will be larger for girls (H2).
There is little in the extant literature to guide specific expectations for differences across settings in the relationship, so we do not propose formal hypotheses about differences across the three countries. In the discussion section, we consider possible explanations for cross-national differences as a potential guide for future comparative research.
METHOD
Data and Sample
We use data from the Family Migration and Early Life Outcomes (FAMELO) project, a multi-national household-based survey carried out in study sites in Mexico, Mozambique, and Nepal. In Mexico, the study area was rural municipalities in Jalisco province with high migration intensity based on the 2010 Mexican census. In Mozambique, the study area was Chibuto district in Gaza province. In Nepal, the study area was Chitwan district. The sites are not representative of their countries overall but were selected as settings of economic and social change, as well as a reliance on internal and international migration.
All three sites used a two-stage sampling approach, with communities selected with probability proportional to size and individuals selected within communities. Within each community, households with at least one child aged 5–17 and an adult relative who had primary responsibility for caring for the child were eligible for inclusion. In eligible households, either one or two children were randomly selected. The number of children per household was also randomly determined. Both the child(ren) and the adult caregiver were interviewed. We use children as the unit of analysis for this paper; measures of household characteristics are taken from the adult interview.
Only adolescents ages 11–17 were asked about future aspirations. This included 1,567 adolescents in Mexico; 1,366 adolescents in Mozambique; and 1,898 adolescents in Nepal. We excluded adolescents with missing data on the main outcome variable (fertility aspirations); other missing values were imputed (see details below). In addition, adolescents with non-numeric desired family size (see Measures section) were excluded from analysis. The final analytic samples for each country are: 1,425 adolescents from 1,168 households in Mexico; 1,253 adolescents from 1,093 households in Mozambique; and 1,890 adolescents from 1,559 households in Nepal. All analyses are run separately for each country.
Dependent Variable
Desired family size is the main outcome in our analysis. All adolescents were asked, “If it were up to you to decide, how many children would you like to have in your life?” Respondents who were unsure were prompted, “It’s ok if you’re not sure. We just want to get an idea of what you think you might want now.” This question specifically asks about fertility desires rather than intentions or expectations. Although the original question in the survey allowed for non-numeric responses (e.g., up to God, depends on partner), few adolescents chose those options. In Mexico, 142 adolescents gave non-numeric responses, most of which (n=124) were “don’t know.” In Mozambique, 106 adolescents gave non-numeric responses; of these, 55 answered “don’t know” and 45 adolescents responded that the decision was “up to God, partner, or family, etc.” In Nepal, eight adolescents gave non-numeric responses; of these, all responded “don’t know.” Because the number of non-numeric responses was relatively small and was distributed across different response categories, we limited analysis to numeric responses.
Independent Variables
The main independent variable in this study is adolescents’ educational aspirations. Respondents were asked, “If you could get as much schooling as you want, how far in school would you go?” Educational aspirations are originally measured in an average of 20 country-specific categories (20 for Mexico, 23 for Mozambique, and 17 for Nepal), ranging from “no school at all” to “doctorate degree.” Based on these distributions, we combined responses into a low, medium, and high category in each country. In Mexico and Nepal, we used the same categories: less than a post-secondary degree (including some post-secondary education without a completed degree), post-secondary degree, and post-graduate or higher. In Mozambique, overall educational aspirations were lower. The three categories we used for educational aspirations in this context were lower secondary or less, any upper secondary (including completed secondary), and any post-secondary.
Differences between boys and girls are a key focus of this analysis; all models control for gender, and we estimate additional models with an interaction term between gender and educational aspirations for all countries. All models account for age, measured as a continuous variable with a range from 11 through 17 years old to adjust for changes in aspirations as children get older. We then sequentially incorporate household/family-level and individual-level controls. We control for sibship size as a predictor of both desired family size and educational aspirations; sibship size is negatively associated with children’s educational attainment across multiple contexts (Choi et al., 2020). Total number of siblings is a continuous variable ranging from 0 to 9 siblings in Mexico, 0 to 21 siblings in Mozambique, and 0 to 6 siblings in Nepal. Caregiver characteristics are included as a proxy for caregiver expectations for the child, which may influence children’s own aspirations in multiple domains. We include the education of the primary caregiver as a categorical variable with three categories. For Mexico and Nepal, the categories are primary school or less, any lower secondary, and upper secondary or more. For Mozambique, where overall education levels are lower, the categories are preschool or less, any primary, and lower secondary or higher.
We use a binary variable measuring parent’s religiosity, where “1” indicates weekly or higher religious attendance and “0” indicates less than weekly religious attendance. This measure is only available for Mexico and Mozambique. In Nepal, we included a measure of the caste of the household head, which reflects social status and opportunities as well as religious practice. This includes the caste groups high caste Hindu, Hill Tibetoburmese, lower caste Hindu, Newar, Terai Tibetoburmese, and other. Models are not identical across settings as we prioritize inclusion of appropriate controls within each country over statistical comparability across countries.
We also created measures of household socioeconomic status based on the results of a principal components analysis (PCA) within each country, following the approach developed in (Filmer & Pritchett, 2001) and widely used in research in low- and middle-income countries. The score is mean-centered based on the country-specific mean.
To capture adolescents’ ideology and related future plans, we include variables that measure multiple dimensions of marital aspirations and views on traditional gender roles in the family. We directly measure adolescents’ views about gender roles in marriage using a continuous measure created from an item where respondents were asked to indicate on a scale of 1–5 whether they “strongly disagree” to “strongly agree” with the statement, “It is best if the husband makes decisions for the household and the wife follows his decisions.” A higher number indicates more traditional views. To measure marital aspirations, we use a continuous measure of adolescents’ ideal age at marriage/union formation, as adolescents’ desired number of children may be tempered by the expected timing of long-term unions (Zahra, 2020). We also include the desired education level of adolescents’ future partners, using the same categories as the education aspirations variables. This variable captures another dimension of marital aspirations, including desired partner characteristics and future resources available in the household.
Missing Data
To account for missing data on independent variables, we use multiple imputation using chained equations with the ice command in Stata. We run 25 imputations for the data, following empirical recommendations (Royston & White, 2011). For Mexico, most variables were missing less than 5% of observations, with only ideal age at marriage missing at 6.7% and desired education of partner missing at 7.8%. In Mozambique, most variables were missing at 5% of observations or less, with religiosity missing on about 11%, ideal age at marriage missing about 10.9% of observations, and educational aspirations for future partner missing at 13.6% of observations. Very little data was missing in Nepal, with most variables missing data at less than 2% of observations.
Analytic Strategy
We first present descriptive results on the distribution of educational and childbearing aspirations in order to understand how these desires are related to other child and family characteristics in Mexico, Mozambique, and Nepal. We then estimate multivariable regression models for each country. Because our dependent variable (desired family size) is a count variable, we use Poisson regressions to understand the relationship between educational aspirations and desired family size (Coxe et al., 2009). We also tested the fit of negative binomial regression but found that Poisson regressions fit the models better. For Nepal, given the high concentration of responses at one or two children, we also tested logistic regression models; results were substantively similar to the Poisson regression, which we prefer for easier discussion of results across countries.
We present four models for each country. We estimate a baseline model accounting for only child age and sex, a second model accounting for variables about the family of origin, a third model with adolescent aspirations and beliefs, and a final model which includes an interaction term between gender and educational aspirations. To compare coefficients across nested Poisson models, we present results as average marginal effects (AMEs) with corresponding standard errors. The AME can be interpreted as the average change in predicted desired number of children associated with a one-unit change in the focal independent variable, with the average calculated across individuals in the sample given observed values of other independent variables. We also use seemingly unrelated estimations to test whether coefficients are significantly different across models (Mize, 2019; Mize et al., 2019).
Sensitivity Analyses
In additional models (not shown here but available upon request), we estimated these models separately for adolescents ages 14–17; estimated models separately for boys and girls; estimated models controlling for current school enrollment; estimated models excluding caste in Nepal; and estimated models varying the order in which we incorporated independent variables. Results from all of these robustness checks were substantively similar to our preferred models.
RESULTS
Descriptive Results
Figure 1 shows the distribution of desired family size across countries; for ease of presentation, responses are top-coded at six. (In the multivariable analyses, the full range of responses is included.) The modal desired family size in Mexico is two children (46% of adolescents), but there are also substantial proportions of adolescents who want three or four children. Desired family size is more evenly distributed in Mozambique, with most adolescents wanting between two and four children, and a non-negligible number desiring larger families – in line with our expectations. There is less variation in Nepal where fertility aspirations are clustered at the lower end of the distribution, with most adolescents desiring two children (71%). Only 1.9% of adolescents in Nepal reported a desired family size of more than two children, and only 0.6% of adolescents reported wanting no children.
Figure 1.
Adolescent Desired Family Size by Country
Notes: Wave I (2017–2018) of Family and Early Life Outcomes (FAMELO) Project. Proportions calculated for analytic sample sizes in each country (see Data and Sample).
Distributions of educational aspirations are shown in Figure 2, grouped into a smaller number of categories based on the International Standard Classification of Education (ISCED) for clarity of presentation and comparability between countries. Overall, educational aspirations are high in the sample, though the specific educational attainment desired varies by country, reflecting both the availability of education and level of expansion within each country. In Mexico, where educational expansion has been ongoing for several decades, most adolescents (66%) aspire to a bachelor’s degree. Educational aspirations are lowest in Mozambique, reflecting the weaker educational system. Still, 65% of adolescents want to finish upper secondary school. In Nepal, the modal desired education is finishing upper secondary school (35%), but most adolescents have higher aspirations – about 28% of the sample wants the equivalent of a bachelor’s degree, and about 25% desire a master’s degree.
Figure 2.
Adolescent Educational Aspirations by Country
Notes: Wave I (2017–2018) of Family and Early Life Outcomes (FAMELO) Project. Proportions calculated for all adolescents with non-missing information on educational aspirations, with the following sample sizes: 1,425 adolescents in Mexico; 1,191 adolescents in Mozambique; and 1,890 adolescents in Nepal. Low, medium, and high educational aspirations for Mexico and Nepal are: less than a post-secondary degree, post-secondary degree, and post-graduate or higher; for Mozambique, they are: lower secondary or less, any upper secondary, and any post-secondary. *Post-secondary non-tertiary and short-cycle tertiary categories were combined with Bachelor’s or equivalent due to small category sizes.
Table 1 shows distributions of the key independent variables across the three countries. We focus discussion in the text on measures of adolescents’ characteristics. In all three countries, the sample is about half girls, and the average age of children in the sample is between 13 and 14. Adolescents in Mexico and Mozambique are less likely than adolescents in Nepal to agree that husbands should have more decision-making power than wives in a family, with an average value of 2.5 (on a scale of 1–5, with 5 representing the least egalitarian views) in Mexico and Mozambique compared to 3.9 in Nepal. The ideal age at marriage is between 24 and 26 in all three countries, and the distribution of desired education for spouse is similar to the distribution of the child’s own desired education. Despite the different cultural contexts of the focal settings, gender role and marital expectations are remarkably similar among adolescents.
Table 1.
Descriptive Statistics
| Mexico (N=1,567) | Mozambique (N=1,366) | Nepal (N=1,898) | ||||
|---|---|---|---|---|---|---|
| Mean/Prop. (SD) | n | Mean/Prop. (SD) | n | Mean/Prop. (SD) | n | |
| Desired family size | 2.5 (1.0) | 1,425 | 3.8 (1.9) | 1,253 | 1.7 (0.5) | 1,890 |
| Child educational aspirationsa | 1,531 | 1,270 | 1,897 | |||
| Low | .22 | 0.26 | 0.35 | |||
| Medium | .69 | 0.64 | 0.28 | |||
| High | .09 | 0.10 | 0.37 | |||
| Girls (vs. Boys) | .49 | 1,567 | 0.48 | 1,364 | 0.49 | 1,898 |
| Age | 13.8 (1.9) | 1,567 | 13.6 (1.9) | 1,366 | 14.2 (2.0) | 1,898 |
| Total number of siblings | 2.1 (1.4) | 1,541 | 3.6 (2.3) | 1,323 | 1.4 (0.8) | 1,885 |
| Primary caregiver educationb | 1,543 | 1,344 | 1,885 | |||
| Low | .36 | 0.27 | 0.46 | |||
| Medium | .48 | 0.62 | 0.20 | |||
| High | .16 | 0.11 | 0.34 | |||
| Weekly religious attendance | .65 | 1,502 | 0.91 | 1,199 | - | - |
| Caste | 1,885 | |||||
| High-caste Hindu | - | - | - | - | 0.47 | |
| Hill Tibetoburmese | - | - | - | - | 0.23 | |
| Lower-caste Hindu | - | - | - | - | 0.11 | |
| Newar | - | - | - | - | 0.04 | |
| Terai Tibetoburmese | - | - | - | - | 0.14 | |
| Other | - | - | - | - | 0.01 | |
| Household wealth index (range: −10 to 10) | 0.1 (1.9) | 1,543 | 0.2 (2.2) | 1,326 | −0.1 (2.1) | 1,885 |
| Believes that husbands should make decisions in household (range: 1–5) | 2.3 (1.1) | 1,556 | 2.7 (1.2) | 1,264 | 3.9 (0.9) | 1,898 |
| Ideal age at marriage | 25.7 (3.5) | 1,403 | 24.3 (5.5) | 1,167 | 25 (2.9) | 1,858 |
| Desired education for partnerc | 1,405 | 1,121 | 1,871 | |||
| Low | .26 | 0.30 | 0.41 | |||
| Medium | .69 | 0.61 | 0.30 | |||
| High | .05 | 0.09 | 0.28 | |||
Notes.
= Educational aspirations for Mexico and Nepal are: less than a post-secondary degree, post-secondary degree, and post-graduate or higher. Mozambique: lower secondary or less, any upper secondary, and any post-secondary.
= Parent education for Mexico and Nepal: primary school or less, any lower secondary, and upper secondary or more. Mozambique: Preschool or less, primary or less, and lower secondary or more. c = Desired education for partner for Mexico and Nepal are: less than a post-secondary degree, post-secondary degree, and post-graduate or higher. Mozambique: lower secondary or lower, any upper secondary, any post-secondary. All statistics drawn from non-imputed data.
= Desired education for partner for Mexico and Nepal are: less than a post-secondary degree, post-secondary degree, and post-graduate or higher. Mozambique: lower secondary or lower, any upper secondary, any post-secondary. All statistics drawn from non-imputed data.
Multivariable Results
Next, we describe the results of our multivariable regressions within each individual country. Results for Mexico, Mozambique, and Nepal are included in Tables 2, 3, and 4, respectively. For each country, we first present a base model (Model 1) to test Hypothesis 1 - that educational and fertility aspirations are negatively correlated. Model 2 adds household-level control variables to test H1a that this association is robust to family characteristics. Model 3 adds ideational measures to the model to address H1b, that the negative correlation between education and fertility aspirations is partially explained by adolescent attitudes related to other family behaviors. Model 4 then includes the interaction term between gender and educational aspirations addressing H2, that the negative association is larger for girls than for boys.
Table 2.
Poisson Regressions Predicting Adolescent Desired Family Size in Mexico
| Model 1 | Model 2 | Model 3 | Model 4e | |
|---|---|---|---|---|
| AME (SE) | AME (SE) | AME (SE) | AME (SE) | |
| Child educational aspirations (vs. low, less than a post-secondary degree) | ||||
| Medium (post-secondary degree) | −0.03 | 0.01 | 0.06 | |
| (0.07) | (0.07) | (0.07) | ||
| High (post-graduate or more) | −0.41*** | −0.33**a | −0.22b | |
| (0.12) | (0.13) | (0.13) | ||
| Effect of educational aspirations on desired family size for boys and girls in interacted models | ||||
| Boys: Post-secondary degree | - | - | - | 0.04 |
| (0.09) | ||||
| Girls: Post-secondary degree | - | - | - | 0.09 |
| (0.11) | ||||
| Boys: Post-graduate or more | - | - | - | −0.05 |
| (0.19) | ||||
| Girls: Post-graduate or more | - | - | - | −0.37* |
| (0.17) | ||||
| Girls (vs. Boys) | −0.18** | −0.18** | −0.19**b | |
| (0.05) | (0.05) | (0.06) | ||
| Age | 0.03* | 0.03* | 0.04* | 0.01* |
| (0.01) | (0.01) | (0.01) | (0.01) | |
| Number of siblings | 0.03 | 0.02 | 0.03 | |
| (0.02) | (0.02) | (0.02) | ||
| Primary caregiver education (vs. low, primary school or less) | ||||
| Medium (any lower-secondary school) | 0.02 | 0.02 | 0.02 | |
| (0.07) | (0.07) | (0.07) | ||
| High (upper-secondary school or more) | 0.06 | 0.07 | 0.07 | |
| (0.08) | (0.08) | (0.08) | ||
| Weekly religious attendance | 0.11 | 0.11 | 0.11 | |
| (0.06) | (0.06) | (0.06) | ||
| Household SES | −0.07*** | −0.06***b | −0.06*** | |
| (0.02) | (0.02) | (0.02) | ||
| Husbands should make decisions | 0.06* | 0.06* | ||
| (0.03) | (0.03) | |||
| Ideal age at marriage | −0.03** | −0.03** | ||
| (0.01) | (0.01) | |||
| Desired education of future partner (vs. low, less than a post-secondary degree) | ||||
| Medium (post-secondary degree) | −0.01 | −0.01 | ||
| (0.07) | (0.07) | |||
| High (post-graduate or more) | −0.10 | −0.09 | ||
| (0.16) | (0.16) | |||
| Observations | 1,425 | 1,425 | 1,425 | 1,425 |
Note: Results derived from imputed data. Average marginal effects (AMEs) presented, with standard errors in parentheses.
p<.001,
p<.01,
p<.05.
Coefficients significantly differ between Model 1 and 2 (p<.05).
Coefficients significantly differ between Model 2 and 3 (p<.05).
Coefficients significantly differ between Model 3 and 4 (p<.05).
The effect of education is significantly different for boys and girls.
Model 4 allows for the gender difference to vary according to educational aspirations. This variation is not theoretically central, so we report the averaged effect across all groups.
Table 3.
Poisson Regressions Predicting Adolescent Desired Family Size in Mozambique
| Model 1 | Model 2 | Model 3 | Model 4e | |
|---|---|---|---|---|
| AME (SE) | AME (SE) | AME (SE) | AME (SE) | |
| Child educational aspirations (vs. low, lower secondary or less) | ||||
| Medium (any upper secondary) | −0.23 | 0.02a | 0.02 | |
| (0.13) | (0.13) | (0.13) | ||
| High (any post-secondary) | −0.78*** | −0.24a | −0.30 | |
| (0.15) | (0.17) | (0.18) | ||
| Effect of educational aspirations on desired family size for boys and girls in interacted models | ||||
| Boys: Post-secondary degree | - | - | - | −0.22d |
| (0.18) | ||||
| Girls: Post-secondary degree | - | - | - | 0.32 |
| (0.18) | ||||
| Boys: Post-graduate or more | - | - | - | −0.69**d |
| (0.22) | ||||
| Girls: Post-graduate or more | - | - | - | 0.22 |
| (0.25) | ||||
| Girls (vs. Boys) | −1.21*** | −1.20*** | −1.22*** | |
| (0.10) | (0.09) | (0.09) | ||
| Age | −0.01 | −0.02 | −0.03 | −0.03 |
| (0.03) | (0.03) | (0.04) | (0.04) | |
| Number of siblings | 0.11* | 0.11* | 0.11* | |
| (0.04) | (0.05) | (0.04) | ||
| Primary caregiver education (vs. low, pre-school or less) | ||||
| Medium (primary or less) | −0.10 | −0.10 | −0.10 | |
| (0.12) | (0.12) | (0.12) | ||
| High (lower secondary or more) | −0.57** | −0.58** | −0.58** | |
| (0.17) | (0.17) | (0.17) | ||
| Weekly religious attendance | −0.40 | −0.41 | −0.41 | |
| (0.21) | (0.21) | (0.21) | ||
| Household SES | −0.10*** | −0.10*** | −0.10*** | |
| (0.02) | (0.03) | (0.02) | ||
| Husbands should make decisions | 0.03 | 0.02 | ||
| (0.05) | (0.05) | |||
| Ideal age at marriage | −0.01 | −0.01 | ||
| (0.01) | (0.01) | |||
| Desired education of future partner (vs. low, lower secondary or less) | ||||
| Medium (any upper secondary) | 0.02 | 0.04 | ||
| (0.12) | (0.12) | |||
| High (any post-secondary) | 0.16 | 0.15 | ||
| (0.24) | (0.24) | |||
| Observations | 1,253 | 1,253 | 1,253 | 1,253 |
Note: Results derived from imputed data. Average marginal effects (AMEs) presented, with standard errors in parentheses.
p<.001,
p<.01,
p<.05.
Coefficients significantly differ between Model 1 and 2 (p<.05).
Coefficients significantly differ between Model 2 and 3 (p<.05).
Coefficients significantly differ between Model 3 and 4 (p<.05).
The effect of education is significantly different for boys and girls.
Model 4 allows for the gender difference to vary according to educational aspirations. This variation is not theoretically central, so we report the averaged effect across all groups.
Table 4.
Poisson Regressions Predicting Adolescent Desired Family Size in Nepal
| Model 1 | Model 2 | Model 3 | Model 4e | |
|---|---|---|---|---|
| AME (SE) | AME (SE) | AME (SE) | AME (SE) | |
| Child educational aspirations (vs. low, less than a post-secondary degree) | ||||
| Medium (post-secondary degree) | 0.00 | 0.03a | 0.04 | |
| (0.03) | (0.03) | (0.03) | ||
| High (post-graduate or more) | −0.06* | −0.02a | −0.01 | |
| (0.03) | (0.03) | (0.03) | ||
| Effect of educational aspirations on desired family size for boys and girls in interacted models | ||||
| Boys: Post-secondary degree | - | - | - | 0.05 |
| (0.04) | ||||
| Girls: Post-secondary degree | - | - | - | 0.03 |
| (0.04) | ||||
| Boys: Post-graduate or more | - | - | - | 0.00 |
| (0.04) | ||||
| Girls: Post-graduate or more | - | - | - | −0.01 |
| (0.05) | ||||
| Girls (vs. Boys) | −0.11*** | −0.13***a | −0.15*** | |
| (0.02) | (0.02) | (0.03) | ||
| Age | 0.01 | 0.01 | 0.01 | 0.01 |
| (0.01) | (0.01) | (0.01) | (0.01) | |
| Number of siblings | 0.04** | 0.04** | 0.04** | |
| (0.01) | (0.01) | (0.01) | ||
| Primary caregiver education (vs. low, primary school or less) | ||||
| Medium (any lower-secondary school) | −0.04 | −0.03 | −0.03 | |
| (0.03) | (0.03) | (0.03) | ||
| High (upper-secondary school or more) | −0.06 | −0.06 | −0.05 | |
| (0.03) | (0.03) | (0.03) | ||
| Caste (vs. High-caste Hindu) | ||||
| Hill Tibetoburmese | −0.04 | −0.04 | −0.04 | |
| (0.03) | (0.03) | (0.03) | ||
| Lower-caste Hindu | 0.04 | 0.04 | 0.04 | |
| (0.04) | (0.04) | (0.04) | ||
| Newar | 0.14* | 0.15* | 0.15* | |
| (0.06) | (0.06) | (0.06) | ||
| Terai Tibetoburmese | −0.01 | −0.02 | −0.02 | |
| (0.04) | (0.04) | (0.04) | ||
| Other | 0.10 | 0.08 | 0.08 | |
| (0.15) | (0.15) | (0.15) | ||
| Household SES | −0.02** | −0.01* | −0.01* | |
| (0.01) | (0.01) | (0.01) | ||
| Husbands should make decisions | 0.05** | 0.05** | ||
| (0.01) | (0.01) | |||
| Ideal age at marriage | −0.01** | −0.01** | ||
| (0.00) | (0.00) | |||
| Desired education of future partner (vs. low, less than a post-secondary degree) | ||||
| Medium (Post-secondary degree) | −0.02 | −0.02 | ||
| (0.03) | (0.03) | |||
| High (post-graduate or more) | 0.02 | 0.02 | ||
| (0.04) | (0.04) | |||
| Observations | 1,890 | 1,890 | 1,890 | 1,890 |
Note: Results derived from imputed data. Average marginal effects (AMEs) presented, with standard errors in parentheses.
p<.001,
p<.01,
p<.05.
Coefficients significantly differ between Model 1 and 2 (p<.05).
Coefficients significantly differ between Model 2 and 3 (p<.05).
Coefficients significantly differ between Model 3 and 4 (p<.05).
The effect of education is significantly different for boys and girls.
Model 4 allows for the gender difference to vary according to educational aspirations. This variation is not theoretically central, so we report the averaged effect across all groups.
In Mexico (Table 2), the baseline model controlling for only age and gender shows a negative association between educational aspirations and desired family size (H1). On average, respondents who aspired a post-graduate degree or more wanted −0.41 fewer children than those who wanted less than a post-secondary degree (p<.001), a relatively large effect size given that the mean desired family size in Mexico is about 2.5 children (Table 1). The difference in desired family size between adolescents who aspire to a post-secondary degree vs. less than a post-secondary degree is small and not statistically different from zero. In Model 2, controlling for household and caregiver characteristics, the association between high educational aspirations and desired family size remains relatively large and statistically significant, although slightly smaller than in the baseline model (H1a). This difference in magnitude across models is statistically significant. After adding measures of adolescent attitudes in Model 3, the AME of high educational aspirations becomes smaller (about half the magnitude of the AME in Model 1) and is no longer statistically different from zero (H1b). Again, this change in magnitude is statistically significant. Taken together, results from Models 1–3 suggest that the negative association between educational aspirations and desired family size is partially, but not primarily, attributable to family characteristics that predict both educational and fertility goals. The association is also partially related to a broader set of attitudes toward marriage and family.
Model 4 tests for an interaction between educational aspirations and gender. Consistent with Hypothesis 2, the average marginal effect for high educational aspirations is larger for girls than for boys, though the difference between these values does not reach statistical significance (p=.19). In the model with the interaction, high educational aspirations are associated with an average desired family size of 0.37 fewer children versus low educational aspirations (p<.05) for girls. For boys, the AME is close to zero and not statistically significant.
Associations of other independent variables with desired family size are largely consistent across the four models. In all models, girls want fewer children than boys, and older adolescents want more children than younger ones. Household wealth is negatively associated with desired family size; other household characteristics, such as sibship size and caregiver education, are not significantly associated with desired family size. Adolescents with less gender egalitarian views want more children on average, although the magnitude of this association is small (0.06 more children for a one-point increase on the five-point scale). A later desired age of marriage is associated with smaller desired family size.
In Mozambique (Table 3), educational aspirations are negatively associated with desired family size in the baseline model, providing support for the first hypothesis. Compared to adolescents who aspire to complete lower secondary school or less, those who aspire to any post-secondary education want 0.78 fewer children (p<.01) and those who aspire to any upper secondary education want 0.23 fewer children, although this difference is not statistically significant. Hypothesis 1a is not supported; controlling for household and family characteristics in Model 2 largely accounts for these associations. In Model 2, the AME for both medium educational aspirations and high educational aspirations are significantly smaller than in Model 1, and neither are significantly different from zero. In Model 3, high educational aspirations are not significantly associated with lower desired family size and this AME is not significantly different from Model 2. Thus, H1b is also not supported in this case.
We do not find support for hypothesis 2 in Mozambique – in fact, gender differences are in the opposite direction from the hypothesized relationship. Model 4 shows that the negative association between educational aspirations and desired family size is larger for boys than for girls, and that this difference is statistically significant. Boys who have high educational aspirations want significantly fewer children, on average, than boys who want a low level (AME = −0.69, p<.01). For girls, the AME is positive but not statistically different from zero.
As in Mexico, girls in Mozambique want fewer children than boys. Sibship size, caregiver education, and household SES are also associated with adolescent desired family size. Having more siblings is associated with a higher desired family size, while adolescents with more educated caregivers and those living in wealthier households want fewer children, on average. Unlike in Mexico, adolescent attitudes toward other family behaviors are not significantly related to desired family size in these multivariable models.
In Nepal (Table 3), educational aspirations and fertility aspirations are negatively correlated, consistent with hypothesis 1, but this relationship is small in magnitude. The average difference between adolescents with high and low educational aspirations is only −0.06 children (p<.05). The difference between adolescents with medium and low educational aspirations is even smaller and not statistically different from zero. Recall that the range of desired family size is very compressed in Nepal (Figure 2); thus, a relatively smaller effect size is not unexpected. Controlling for household and family characteristics reduces the magnitude of the AME even further, and differences across educational aspirations are not statistically significant in Model 2, yielding a lack of support for H1a. Adolescent family attitudes do not further contribute to explaining this association (Model 3). Thus, H1b is not supported in Nepal. In Model 4, the AME of educational aspirations is small and not statistically significant for both boys and girls, and there is no statistically significant difference between the AME for boys and girls. Thus, H2 is not supported.
Across all models, girls want fewer children than boys in Nepal. Adolescents living in households where the household head is in the Newar caste (who tend to be socially disadvantaged) have larger desired family sizes, on average, than those with household heads who are high-caste Hindu. Adolescents in wealthier households have smaller desired family sizes. As in Mexico, adolescents with less gender egalitarian views want larger families, and those who want to marry later want smaller families.
For ease of interpretation and comparison across countries, a summary of all results can be found in Table 5. We describe and contextualize similarities and differences across countries in the discussion section.
Table 5.
Summary of Results for All Countries
| Hypothesis | Mexico | Mozambique | Nepal |
|---|---|---|---|
| H1: Educational aspirations and desired family size are negatively correlated. | Supported | Supported | Supported |
| H1a: This relationship persists when accounting for household and family characteristics. | Supported | Not supported | Not supported |
| H1b: This relationship is partially explained by individual beliefs and values. | Supported | Not supported | Not supported |
| H2: This relationship is stronger for girls than for boys. | Supported | Not supported: stronger for boys | Not supported; no difference |
| Other independent variables significantly associated with desired family size? | Household SES (−); Gender inegalitarian beliefs (+); Desired age at marriage (−) | Sibship size (+); Caregiver education (−); Household SES (−) | Sibship size (+); Household head caste; Household SES (−); Gender inegalitarian beliefs (+); Desired age at marriage (−) |
Notes: + and - signs after listed variables denotes the direction of the association between the respective variable and desired family size.
DISCUSSION AND CONCLUSION
Adolescent aspirations for education and childbearing reflect their values about these two domains of adult life. In this article, we examine the relationship between educational aspirations and desired family size among adolescents in three disparate contexts, assessing differences between boys and girls, in order to understand how values related to these two domains are correlated. We find that in all three study contexts, on average, adolescents who desire more education want fewer children. In baseline models with minimal controls, educational aspirations are negatively associated with desired family size in all three settings, supporting H1. However, how this relationship varies for boys and girls and whether the association between educational aspirations and desired family size is robust to household factors, family characteristics, and adolescent beliefs and values varies by country (see Table 5). We did not propose formal hypotheses about cross-national differences, given the small evidence base of comparative research to build on. However, it is possible to speculate post-hoc about possible explanations for these differences as a potential direction for future research.
Mexico is the most economically developed of our three settings. Educational expansion is furthest along, and jobs outside of subsistence agriculture are most widely available. Paid work – especially the kind of work that highly educated people would seek – takes place outside the home. Practical conflicts between paid work and childbearing may influence aspirations, and cultural conflicts between education and childbearing stemming from the breadwinner-homemaker model may be salient in this context, particularly for women. Paternalistic interactions and policies are still relatively commonplace within professional workplaces, and women are less likely to be perceived as fitting the requirements of an ideal worker (Brumley, 2014; Ruiz Castro, 2012). Thus, girls may see ambitious educational and career goals as in conflict with their family formation goals. These conflicts are consistent with the robust negative association between educational aspirations and desired family size, the role of beliefs and values in explaining this association, and the fact that the association is stronger for girls than for boys in the Mexican context.
Mozambique is the least developed of our three settings. Access to education, especially secondary education, remains limited, and many households rely on subsistence agriculture or small-scale family production. The division of labor is highly gendered, but this division does not map on to the division between domestic work and paid work envisioned by the breadwinner-homemaker model; rather, most work takes place as part of family-based agricultural production. In this context, conflicts between education and fertility may come from economic constraints in accessing education rather than role conflict or ideological conflict. Both the costs of adolescents’ education and the costs of their potential future children’s education may be salient in linking educational goals with childbearing goals (Axinn & Barber, 2001; Caldwell, 1976, 1980). Further, prior research in the sub-Saharan African context shows that economic hardships are negatively associated with both desired family size and achieved fertility (Agadjanian, 2005; Smith, 2020). Consistent with this research, we find that the relatively large negative association between educational aspirations and desired family size in Mozambique is mostly explained by household characteristics, with caregiver education and household economic conditions playing a key role. That is, adolescents living in households with more resources to support their education are likely to desire both more education and fewer children. Our results show that the negative association between educational aspirations and desired family size is stronger for boys than for girls in this context, perhaps because boys are more likely to anticipate having financial responsibility for their children’s schooling.
Nepal provides an interesting in-between case. Schooling is widely available in Nepal, and educational aspirations among adolescents are extremely high. At the individual level, the association between educational aspirations and desired family size is weak even in models with minimal controls and almost entirely explained by household characteristics. Adolescent beliefs and values are correlated with desired family size but do not mediate the relationship with educational aspirations. Average desired family size is almost universally small in Nepal – that is, at the aggregate level, there does seem to be a strong negative correlation between educational aspirations and fertility aspirations. It may be that this aggregate relationship is so strong that there is relatively little individual-level variation to explain.
Our findings point to the joint importance of economic systems and gendered family systems as contextual factors that shape adolescent aspirations. McDonald (2000) proposed that fertility declines initially as gender equity increases in non-family institutions and women take on responsibilities outside of childbearing. If gender equity in family institutions also increases, both men and women can combine work and childcare, and fertility stabilizes; if gender equity in the family does not increase, fertility continues to drop to lowest-low levels. Our findings hint at a similar interaction between gender equity in multiple institutions (education, work, family), but also call attention to the fact that the distinction between “family” and “non-family” institutions may depend on a certain level of economic development and differentiation, and that in many agricultural economies, economic production takes place within rather than outside the family. Based on these findings, future research with a larger number of countries could test systematic variation in the relationship between educational aspirations and fertility aspirations according to the level of men’s and women’s participation in the formal (non-family) labor force; the economic returns to schooling for men and for women; and the degree to which parenthood is negatively correlated with labor force participation for men and for women.
Across study sites, educational attainment is associated with lower completed fertility and lower desired fertility (J. S. Barber & Axinn, 2004; Kebede et al., 2021; Martin, 1995; Satyavada & Adamchak, 2000). (At least among women – data on men’s fertility is much more limited.) Our finding of mixed associations between educational aspirations and fertility aspirations early in the life course points to potentially complex patterns of how and for whom educational aspirations are realized. It is possible that childbearing makes it difficult to achieve early educational aspirations, or individuals change their goals as they gain experience both in school and in family relationships. Longitudinal data are needed to fully understand these processes.
Though the relationship between educational aspirations and desired family size is not mediated by gender across countries, we do find consistent gender differences in desired family size overall. In all study sites, girls report a smaller desired family size than boys, in models with and without individual and household controls. Women bear the physical consequences of pregnancy and birth, and women in these settings are responsible for a larger share of childcare than men are. Adolescent girls are likely aware of these differential burdens. Yet because most childbearing takes place in heterosexual couples, women will have, on average, the same number of children as men. As adolescents grow up, boys and girls will be differentially able to carry out their aspirations, and their early preferences may evolve in different ways. Differences in how aspirations are converted to outcomes is a possible dimension of gender inequality that evolves over the life course. In the United States, boys and girls have similar goals for work, family, and work-family balance in adolescence and early adulthood, and these goals are largely egalitarian (Gerson, 2009; Hayford & Hardie, 2021). However, when young people are asked not only about their goals but about what they think will happen if they are unable to achieve their goals, more gender differences appear. Gerson (2009) found that both boys and girls wanted egalitarian relationships with shared responsibilities for work and childcare. Boys’ “second choice” option was a breadwinner-homemaker model, while girls’ fallback moved toward more, not less, economic responsibility and independence for mothers. As the adolescent boys and girls in our sample age, their aspirations may evolve and diverge as well.
Our analysis draws on data from three distinct study settings. This comparative component is a key strength, but also a potential limitation; each study site is a unique case with its own particularities, and the comparative approach prevents a full investigation of these specificities. Our analysis attempted to balance specificity with comparability, for example by incorporating site-specific variables such as caste in Nepal and by using different categorizations of high, medium, and low education in Mozambique than in the other two sites. Future research in a wider set of settings, as well as larger-scale comparative research, is necessary to develop an understanding of how context shapes aspirations more fully in different domains and their intersection. As noted above, our cross-sectional data is another limitation of our approach. We are unable to disentangle possible mutual causation between educational and fertility aspirations, and we can analyze only a snapshot of their association rather than the evolving picture. Although we control for some of the most salient family, household, and individual characteristics that contribute to both educational aspirations and fertility aspirations, we cannot fully analyze the way that aspirations in these domains are jointly determined.
Despite these limitations, our study provides an important advancement in the literature and addresses how educational aspirations are related to childbearing aspirations. As aspirations likely affect future plans, desires for family and education will have important implications on adolescents as they transition to adulthood. Our study underscores the need to analyze fertility aspirations for adolescents and how these aspirations are generated in conjunction with other desires for the future, particularly education. Our results also have important implications for understanding the conflicts – real and perceived – between educational aspirations and desired future family size. Given that fertility preferences are informed by societal values (Ryder, 1973, 2010), girls living within a context in which rapid, if uneven, educational expansion coexists with largely gendered family roles may see higher educational aspirations as incompatible with having a large family. The negative relationship between educational and fertility aspirations in our study, rather than completed education or fertility, highlights the salience of gender inequality even at early stages of the life course. Even with increasing women’s labor force participation, women are the primary caregivers for their children and simultaneously perform different roles, including worker, spouse, and mother (Schieman et al., 2009; Voydanoff, 2004). As adolescents transition to adulthood, they will increasingly have greater agency to attempt to meet their aspirations – while also becoming increasingly aware of the constraints of their life circumstances. Given adolescents’ awareness of both interpersonal and broad gender inequities in their societies, we may also expect that ambitious aspirations may be attenuated or adjusted as adolescents approach adulthood. Continuing to study how adolescents conceive of their future opportunities and how these conceptions influence their future decisions will help us better understand the relationship between fertility desires, actual fertility, and education.
Acknowledgements:
We gratefully acknowledge support from the National Institutes of Health, Eunice Kennedy Shriver Institute of Child Health and Human Development via grants P01 HD080659 (Glick, PI), 3P01HD080659-04S1 (Glick, PI; Alcaraz), OSU’s Institute for Population Research (P2C-HD058484), and PSU’s Population Research Institute, P2C-HD041025. We also acknowledge seed grant funding from the Department of Sociology at Ohio State University. An earlier version of this paper was presented at the 2019 Population Association of America Annual Meeting. We thank John Casterline for providing us with information based on calculations from Demographic and Health Surveys.
Contributor Information
Melissa Alcaraz, Brigham Young University.
Sarah R. Hayford, Ohio State University
Jennifer E. Glick, Pennsylvania State University
REFERENCES
- Agadjanian V (2005). Fraught with Ambivalence: Reproductive Intentions and Contraceptive Choices in a Sub-Saharan Fertility Transition. Population Research and Policy Review, 24(6), 617–645. [Google Scholar]
- Alexander KL, & Cook MA (1979). The Motivational Relevance of Educational Plans: Questioning the Conventional Wisdom. Social Psychology Quarterly, 42(3), 202–213. [Google Scholar]
- Allendorf K, & Thornton A (2019). New Research on Developmental Idealism: Introduction to the Special Issue. Sociology of Development, 5(3), 225–228. [Google Scholar]
- Altamirano A, Lopez-Calva LF, & Soloaga I (2010). Inequality and Teenagers’ Educational Aspirations in Urban Mexico. SSRN Electronic Journal. [Google Scholar]
- Axinn WG, & Barber JS (2001). Mass Education and Fertility Transition. American Sociological Review, 66(4), 481–505. [Google Scholar]
- Bachrach CA, & Morgan SP (2013). A Cognitive–Social Model of Fertility Intentions. Population and Development Review, 39(3), 459–485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Baker D (2014). The Schooled Society: The Educational Transformation of Global Culture. Stanford University Press. [Google Scholar]
- Barber BL, Eccles JS, & Stone MR (2001). Whatever Happened to the Jock, the Brain, and the Princess?: Young Adult Pathways Linked to Adolescent Activity Involvement and Social Identity. Journal of Adolescent Research, 16(5), 429–455. [Google Scholar]
- Barber JS, & Axinn WG (2004). New ideas and fertility limitation: The role of mass media. Journal of Marriage and Family, 66(5), 1180–1200. [Google Scholar]
- Basu AM (2002). Why does Education Lead to Lower Fertility? A Critical Review of Some of the Possibilities. World Development, 30(10), 1779–1790. [Google Scholar]
- Becker GS (1981). Altruism in the Family and Selfishness in the Market Place. Economica, 48(189), 1–15. [Google Scholar]
- Behrman JA (2015). Does Schooling Affect Women’s Desired Fertility? Evidence From Malawi, Uganda, and Ethiopia. Demography, 52(3), 787–809. [DOI] [PubMed] [Google Scholar]
- Blair-Loy M (2003). Competing Devotions: Career and Family among Women Executives. Harvard University Press. [Google Scholar]
- Bongaarts J (2003). Completing the fertility transition in the developing world: The role of educational differences and fertility preferences. Population Studies, 57(3), 321–335. [DOI] [PubMed] [Google Scholar]
- Bongaarts J (2010). The causes of educational differences in fertility in Sub-Saharan Africa. Vienna Yearbook of Population Research, 8, 31–50. [Google Scholar]
- Bongaarts J, Mensch BS, & Blanc AK (2017). Trends in the age at reproductive transitions in the developing world: The role of education. Population Studies, 71(2), 139–154. [DOI] [PubMed] [Google Scholar]
- Brand JE, & Davis D (2011). The Impact of College Education on Fertility: Evidence for Heterogeneous Effects. Demography, 48(3), 863–887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Brumley KM (2014). ‘You care for your work; I’ll care for your family’: Perceptions of paternalistic managerial actions and employee commitment in Mexico. Community, Work & Family, 17(4), 467–485. [Google Scholar]
- Caldwell JC (1976). Toward A Restatement of Demographic Transition Theory. Population and Development Review, 2(3/4), 321–366. [Google Scholar]
- Caldwell JC (1980). Mass Education as a Determinant of the Timing of Fertility Decline. Population and Development Review, 6(2), 225–255. [Google Scholar]
- Chiapa C, Prina S, & Parker A (2016). The Effects of Financial Inclusion on Children’s Schooling, and Parental Aspirations and Expectations. Journal of International Development, 28(5), 683–696. [Google Scholar]
- Choi S, Taiji R, Chen M, & Monden C (2020). Cohort Trends in the Association Between Sibship Size and Educational Attainment in 26 Low-Fertility Countries. Demography, 57(3), 1035–1062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Collins C (2019). Making Motherhood Work. Princeton Unviersity Press. [Google Scholar]
- Coontz S (1993). The Way We Never Were: American Families And The Nostalgia Trap (Reprint edition). Basic Books. [Google Scholar]
- Coxe S, West SG, & Aiken LS (2009). The Analysis of Count Data: A Gentle Introduction to Poisson Regression and Its Alternatives. Journal of Personality Assessment, 91(2), 121–136. [DOI] [PubMed] [Google Scholar]
- DeRose LF, & Ezeh AC (2005). Men’s influence on the onset and progress of fertility decline in Ghana, 1988–98. Population Studies, 59(2), 197–210. [DOI] [PubMed] [Google Scholar]
- Detering NM (2015). Instrumental and Expressive Education: College Planning in the Face of Poverty. Sociology of Education, 88(4), 284–301. [Google Scholar]
- Dribe M, & Stanfors M (2008). Education, Work and Parenthood: Comparing the Experience of Young Men and Women in Sweden. Journal of Family and Economic Issues, 30(1), 32. [Google Scholar]
- Eccles JS, Vida MN, & Barber B (2004). The Relation of Early Adolescents’ College Plans and Both Academic Ability and Task-Value Beliefs to Subsequent College Enrollment. The Journal of Early Adolescence, 24(1), 63–77. [Google Scholar]
- Filmer D, & Pritchett LH (2001). Estimating Wealth Effects without Expenditure Data-or Tears: An Application to Educational Enrollments in States of India. Demography, 38(1), 115–132. [DOI] [PubMed] [Google Scholar]
- Fishbein M, & Ajzen I (2009). Predicting and Changing Behavior: The Reasoned Action Approach. Psychology Press. [Google Scholar]
- Frye M (2012). Bright Futures in Malawi’s New Dawn: Educational Aspirations as Assertions of Identity. American Journal of Sociology, 117(6), 1565–1624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Frye M (2017). Cultural Meanings and the Aggregation of Actions: The Case of Sex and Schooling in Malawi. American Sociological Review, 82(5), 945–976. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gerson K (2009). The Unfinished Revolution: Coming of Age in a New Era of Gender, Work, and Family (1 edition). Oxford University Press. [Google Scholar]
- Glick JE, Ruf SD, White MJ, & Goldscheider F (2006). Educational Engagement and Early Family Formation: Differences by Ethnicity and Generation. Social Forces, 84(3), 1391–1415. [Google Scholar]
- Glick JE, & White MJ (2004). Post-secondary school participation of immigrant and native youth: The role of familial resources and educational expectations. Social Science Research, 33(2), 272–299. [Google Scholar]
- Hartas D (2016). Young people’s educational aspirations: Psychosocial factors and the home environment. Journal of Youth Studies, 19(9), 1145–1163. [Google Scholar]
- Hayford SR (2009). The evolution of fertility expectations over the life course. Demography, 46(4), 765–783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hayford SR, & Hardie JH (2021). Gender Differences in Adolescents’ Work and Family Orientations in the United States. The Sociological Quarterly, 62(3), 488–509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hayford SR, & Morgan SP (2008). Religiosity and Fertility in the United States: The Role of Fertility Intentions. Social Forces, 86(3), 1163–1188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hays S (1998). Cultural Contradictions of Motherhood. Yale University Press. [Google Scholar]
- Hofferth SL, Reid L, & Mott FL (2001). The Effects of Early Childbearing on Schooling over Time. Family Planning Perspectives, 33(6), 259–267. [PubMed] [Google Scholar]
- Instituto Nacional de Estadística y Geografía (INEGI). (2018). Encuesta Nacional de la Dinámica Demográfica (ENADID) 2018 [Data file]. Instituto Nacional de Salud Pública & Consejo Nacional de Población. [Google Scholar]
- Johnson-Hanks JA (2005). Uncertain Honor: Modern Motherhood in an African Crisis. University of Chicago Press. [Google Scholar]
- Johnson-Hanks JA, Bachrach CA, Morgan SP, & Kohler H-P (2011). The Theory of Conjunctural Action. In Johnson-Hanks JA, Bachrach CA, Morgan SP, & Kohler H-P (Eds.), Understanding Family Change and Variation: Toward a Theory of Conjunctural Action (pp. 1–22). Springer; Netherlands. [Google Scholar]
- Kebede E, Striessnig E, & Goujon A (2021). The relative importance of women’s education on fertility desires in sub-Saharan Africa: A multilevel analysis. Population Studies, 0(0), 1–20. [DOI] [PubMed] [Google Scholar]
- Kravdal Ø (1992). The Emergence of a Positive Relation Between Education and Third Birth Rates in Norway with Supportive Evidence from the United States. Population Studies, 46(3), 459–475. [Google Scholar]
- Kravdal Ø (2002). Education and fertility in sub-Saharan africa: Individual and community effects. Demography, 39(2), 233–250. [DOI] [PubMed] [Google Scholar]
- Kravdal Ø, & Rindfuss RR (2008). Changing Relationships between Education and Fertility: A Study of Women and Men Born 1940 to 1964. American Sociological Review, 73(5), 854–873. [Google Scholar]
- Lakomý M (2017). The role of values and of socioeconomic status in the education-fertility link among men and women. Vienna Yearbook of Population Research, 15, 121–141. [Google Scholar]
- Lareau A (2002). Invisible Inequality: Social Class and Childrearing in Black Families and White Families. American Sociological Review, 67(5), 747–776. [Google Scholar]
- Lesthaeghe R (1983). A Century of Demographic and Cultural Change in Western Europe: An Exploration of Underlying Dimensions. Population and Development Review, 9(3), 411–435. [Google Scholar]
- Luker K (1997). Dubious Conceptions: The Politics of Teenage Pregnancy. Harvard University Press. [Google Scholar]
- Lutz W, & Samir KC (2011). Global Human Capital: Integrating Education and Population. Science, 333(6042), 587–592. [DOI] [PubMed] [Google Scholar]
- Marshall EA, & Shepherd H (2018). Fertility Preferences and Cognition: Religiosity and Experimental Effects of Decision Context on College Women. Journal of Marriage and Family, 80(2), 521–536. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Martin TC (1995). Women’s Education and Fertility: Results from 26 Demographic and Health Surveys. Studies in Family Planning, 26(4), 187–202. [PubMed] [Google Scholar]
- McDonald P (2000). Gender Equity in Theories of Fertility Transition. Population and Development Review, 26(3), 427–439. [Google Scholar]
- Miller WB (1994). Childbearing motivations, desires, and intentions: A theoretical framework. Genetic, Social, and General Psychology Monographs, 120(2), 223–258. [PubMed] [Google Scholar]
- Ministerio da Saude (MISAU), Instituto Nacional de Estatística (INE), & ICF International (ICFI). (2013). Moçambique Inquérito Demográfico e de Saúde 2011. MISAU, INE e ICFI. [Google Scholar]
- Ministry of Health, Nepal, New ERA, & ICF. (2017). Nepal Demographic and Health Survey 2016. Ministry of Health, Nepal. [Google Scholar]
- Mische A (2009). Projects and Possibilities: Researching Futures in Action. Sociological Forum, 24(3), 694–704. [Google Scholar]
- Mize TD (2019). Best Practices for Estimating, Interpreting, and Presenting Nonlinear Interaction Effects. Sociological Science, 6, 81–117. [Google Scholar]
- Mize TD, Doan L, & Long JS (2019). A General Framework for Comparing Predictions and Marginal Effects across Models. Sociological Methodology, 49(1), 152–189. [Google Scholar]
- Mollborn S, & Jacobs J (2011). “We’ll Figure a Way”: Teenage Mothers’ Experiences in Shifting Social and Economic Contexts. Qualitative Sociology, 35(1), 23–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Morgan SP, & Rackin H (2010). The Correspondence Between Fertility Intentions and Behavior in the United States. Population and Development Review, 36(1), 91–118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Musick K, England P, Edgington S, & Kangas N (2009). Education Differences in Intended and Unintended Fertility. Social Forces, 88(2), 543–572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- National Research Council and Institute of Medicine. (2004). Growing Up Global: The Changing Transitions to Adulthood in Developing Countries.
- Nielsen K (2015). ‘“Fake It ”til You Make It”: Why Community College Students’ Aspirations ‘“Hold Steady”’. Sociology of Education, 88(4), 265–283. [Google Scholar]
- Nisén J, Martikainen P, Kaprio J, & Silventoinen K (2013). Educational Differences in Completed Fertility: A Behavioral Genetic Study of Finnish Male and Female Twins. Demography, 50(4), 1399–1420. [DOI] [PubMed] [Google Scholar]
- Nisén J, Martikainen P, Myrskylä M, & Silventoinen K (2018). Education, Other Socioeconomic Characteristics Across the Life Course, and Fertility Among Finnish Men. European Journal of Population, 34(3), 337–366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nitsche N, & Hayford SR (2020). Preferences, Partners, and Parenthood: Linking Early Fertility Desires, Marriage Timing, and Achieved Fertility. Demography, 57(6), 1975–2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Perreira KM, & Spees L (2015). Foiled Aspirations: The Influence of Unauthorized Status on the Educational Expectations of Latino Immigrant Youth. Population Research and Policy Review, 34(5), 641–664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Quesnel-Vallée A, & Morgan SP (2003). Missing the Target? Correspondence of Fertility Intentions and Behavior in the U.S. Population Research and Policy Review, 22(5), 497–525. [Google Scholar]
- Raymo JM, Carlson MJ, VanOrman A, Lim S, Perelli-Harris B, & Iwasawa M (2015). Educational differences in early childbearing: A cross-national comparative study. Demographic Research, 33, 65–92. [PMC free article] [PubMed] [Google Scholar]
- Regnier-Loilier A, & Depledge R (2006). Influence of Own Sibship Size on the Number of Children Desired at Various Times of Life. Population, Vol. 61(3), 165–194. [Google Scholar]
- Rindfuss RR, Morgan SP, & Offutt K (1996). Education and the Changing Age Pattern of American Fertility: 1963–1989. Demography, 33(3), 277–290. [PubMed] [Google Scholar]
- Roubeni S, De Haene L, Keatley E, Shah N, & Rasmussen A (2015). “If We Can’t Do It, Our Children Will Do It One Day”: A Qualitative Study of West African Immigrant Parents’ Losses and Educational Aspirations for Their Children. American Educational Research Journal, 52(2), 275–305. [Google Scholar]
- Royston P, & White IR (2011). Multiple Imputation by Chained Equations (MICE): Implementation in Stata. Journal of Statistical Software, 45(1), 1–20. [Google Scholar]
- Ruiz Castro M (2012). Time Demands and Gender Roles: The Case of a Big Four Firm in Mexico. Gender, Work & Organization, 19(5), 532–554. [Google Scholar]
- Ryder NB (1973). A Critique of the National Fertility Study. Demography, 10(4), 495–506. [PubMed] [Google Scholar]
- Ryder NB (2010). Norman Ryder on the Sociology of Fertility Reduction. Population and Development Review, 36(3), 607–620. [Google Scholar]
- Sassler S, & Schoen R (1999). The Effect of Attitudes and Economic Activity on Marriage. Journal of Marriage and Family, 61(1), 147–159. [Google Scholar]
- Satyavada A, & Adamchak DJ (2000). Determinants of current use of contraception and children ever born in Nepal. Social Biology, 47(1–2), 51–60. [DOI] [PubMed] [Google Scholar]
- Schieman S, Milkie MA, & Glavin P (2009). When Work Interferes with Life: Work-Nonwork Interference and the Influence of Work-Related Demands and Resources. American Sociological Review, 74(6), 966–988. [Google Scholar]
- Schoon I (2001). Teenage job aspirations and career attainment in adulthood: A 17-year follow-up study of teenagers who aspired to become scientists, health professionals, or engineers. International Journal of Behavioral Development, 25(2), 124–132. [Google Scholar]
- Seawright J, & Gerring J (2008). Case Selection Techniques in Case Study Research: A Menu of Qualitative and Quantitative Options. Political Research Quarterly, 61(2), 294–308. [Google Scholar]
- Sewell WH, Haller AO, & Portes A (1969). The Educational and Early Occupational Attainment Process. American Sociological Review, 34(1), 82–92. [Google Scholar]
- Shepherd H, & Marshall EA (2018). The implicit activation mechanism of culture: A survey experiment on associations with childbearing. Poetics, 69, 1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Silva JM, & Snellman K (2018). Salvation or Safety Net? Meanings of “College” among Working- and Middle-Class Young Adults in Narratives of the Future. Social Forces, 97(2), 559–582. [Google Scholar]
- Skirbekk V (2008). Fertility trends by social status. Demographic Research, 18, 145–180. [Google Scholar]
- Smith DJ (2001). Romance, Parenthood, and Gender in a Modern African Society. Ethnology, 40(2), 129–151. [Google Scholar]
- Smith DJ (2020). Masculinity, Money, and the Postponement of Parenthood in Nigeria. Population and Development Review, 46(1), 101–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sobotka T, & Beaujouan É (2014). Two Is Best? The Persistence of a Two-Child Family Ideal in Europe. Population and Development Review, 40(3), 391–419. [Google Scholar]
- Testa MR (2014). On the positive correlation between education and fertility intentions in Europe: Individual- and country-level evidence. Advances in Life Course Research, 21, 28–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Testa MR, & Stephany F (2017). The educational gradient of fertility intentions: A meta-analysis of European studies. Vienna Yearbook of Population Research, 15, 293–330. [Google Scholar]
- Thornton A (2004). Reading history sideways: The fallacy and enduring impact of the developmental paradigm on family life. University of Chicago Press. [Google Scholar]
- Thornton A, Binstock G, Yount KM, Abbasi-Shavazi MJ, Ghimire D, & Xie Y (2012). International Fertility Change: New Data and Insights From the Developmental Idealism Framework. Demography, 49(2), 677–698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Thornton A, Pierotti RS, Young-DeMarco L, & Watkins S (2014). Developmental Idealism and Cultural Models of the Family in Malawi. Population Research and Policy Review, 33(5), 693–716. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Townsend N (2002). Package Deal. Temple University Press. [Google Scholar]
- Tsung L, & Gao F (2012). What accounts for the underachievement of South Asians in Hong Kong? The voices of Pakistani and Nepalese parents. Educational Research, 54(1), 51–63. [Google Scholar]
- United Nations Human Development Reports. (2018). United Nations Human Development Index (HDI) | Human Development Reports.
- Voydanoff P (2004). The Effects of Work Demands and Resources on Work-to-Family Conflict and Facilitation. Journal of Marriage and Family, 66(2), 398–412. [Google Scholar]
- World Bank. (2019). World Development Indicators. https://data.worldbank.org/indicator
- Yeatman S, Sennott C, & Culpepper S (2013). Young Women’s Dynamic Family Size Preferences in the Context of Transitioning Fertility. Demography, 50(5), 1715–1737. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yoo SH (2014). Educational differentials in cohort fertility during the fertility transition in South Korea. Demographic Research, 30, 1463–1494. [Google Scholar]
- Zahra F (2020). High Hopes, Low Dropout: Gender Differences in Aspirations for Education and Marriage, and Educational Outcomes in Rural Malawi. Comparative Education Review, 64(4), 670–702. [Google Scholar]