Gender Differences in Adolescents’ Work and Family Orientations in the United States

Abstract

We use the Education Longitudinal Study: 2002 to compare the perceived importance of work and family achievement among young women and men. We apply latent class analysis to identify distinct configurations of values, then examine associations between latent classes and educational and occupational expectations. Results show high ambitions for both work and family among both young women and men. Although young women are more likely than young men to report that marriage and family relationships are very important, differences are small. Young women are also more likely to value work-related outcomes and to hold high educational and occupational expectations.

Depsite sustained and substantial shifts in the gendered distribution of paid work and unpaid family caregiving in the United States, these responsibilities continue to be experienced differently by men and women. Profound changes in women’s economic opportunities have occurred over the past five decades in the United States, spurring increases in female labor force participation (Goldin 2006; Percheski 2008). Women make up nearly half the U.S. labor market (BLS 2015), and women’s labor force participation is expected to rise at a faster rate than men’s, although at a slower rate than in recent years (Toossi 2015). Yet despite these signs of movement toward gender equality, women continue to earn less and work in less prestigious occupations than men (Cohen, Huffman and Knauer 2009; Ryu 2010). Similarly, although men’s time spent in household and childcare has increased, women still provide the bulk of household labor in cohabiting and married couple households (Bianchi et al. 2000; Gupta 1999; Sassler and Miller 2017), particularly in families with children (Yavorsky, Dush, and Schoppe-Sullivan 2015; Sanchez and Thomson 1997). Women are also far more likely than men to live with children in single-parent households (U.S. Census 2019).

While several studies shed light on how adult men and women negotiate the potentially conflicting demands of work and family responsibilities and how inequalities in the two spheres may be jointly caused or mutually reinforcing (e.g. Blair-Loy 2001; Damaske 2011; Garey 1999; Gerson 1994, 2010; Jacobs and Gerson 2005; Johnson 2005), less is known about how adolescents anticipate balancing these two aspects of their future lives. Understanding adolescents’ work and family values (e.g., the relative importance they place on dimensions of work and family life) is important because these orientations early in the life course set the stage for education, employment, and family trajectories later in life (Bozick et al. 2010; Mortimer, Staff, and Lee 2005). To the extent that young women place greater importance on family life than young men (either in absolute terms or relative to work life), these values may contribute to later gender inequality in work and family. Furthermore, understanding adolescents’ values for family and work outcomes can shed light on how young people perceive potential conflicts between work and family and the extent to which they incorporate these conflicts into their own “imagined futures” (Frye 2012). That is, even if values were not predictive of behavior, they would be worthy of study as a reflection of cultural schemas.

In this paper, we consider gender differences in the relationship between work and family values early in the life course, drawing on data from the 2004 follow-up of the Education Longitudinal Study: 2002, a school-based survey of adolescents’ educational, occupational, and life plans. We study associations between the importance placed on future work and family activities among secondary school-age students, assess the gendered nature of these associations, and examine the association between combinations of work and family orientations and specific educational and occupational expectations. We apply latent class analysis to develop profiles of adolescents’ work and family values. Latent class analysis is ideal for our purposes because it utilizes a person-centered approach that classifies individuals into “types” based on their responses to a number of survey items, thus allowing “researchers to characterize individuals more holistically” (p. 62, Lanza and Cooper 2016) than is possible with variable-centered analyses. It is particularly well-suited for studying associations between work and family values, since previous research has theorized and observed distinct typologies of work-family orientations for both men and women (e.g., Damaske 2011; Gerson 1994; Hakim 2000). Thus, rather than just predict what characteristics are associated with valuing work or valuing family, we can identify typologies of young people who value primarily family, primarily work, both, neither, or other combinations, and then examine what characteristics are associated with these typologies and how they are associated with educational and occupational aspirations. This approach moves beyond previous research that has largely considered work orientations and family orientations as independent predictors of other aspirations. This research contributes to the literature on gender, work and family life, and adolescent aspirations.

Gender and work-family outcomes

Gender inequality in the labor market and in the home have multiple overlapping sources. “Demand-side” explanations of gender inequality argue that structural barriers make it more difficult for women to break into certain occupations, work long hours, and get ahead at work. According to these explanations, women encounter barriers in the pathways to high status jobs (Riegle-Crumb et al. 2012; Riegle-Crumb and Humphries 2012), discrimination in employment and at work (Bobbitt-Zeher 2011; McLaughlin, Uggen, and Blackstone 2012; Reskin and Roos 1990), and greater demands at home (Daminger 2019; Sassler and Miller 2017; Schneider 2011), all of which curtail the choices that they can make. A robust literature on gender inequality has established the importance of these kinds of structural barriers in explaining women’s and men’s work-family outcomes (e.g. Collins 2019; Damaske 2011; McRae 2003).

However, these structural barriers are not the only contributor to gender differences in work and family outcomes. Indeed, work on gender inequality in earnings and occupational gender segregation suggests that multiple factors are likely at play, including socialization and psychological factors, in addition to discrimination (Blau and Kahn 2017; England 2005). If men and women have different preferences and expectations for either family- or work-based activities, these “supply-side” factors may lead to different choices in both domains (Correll 2001; Eccles 1994; Hakim 2000). Preference theory (Hakim 2000) proposes that women’s preferences for work and family fall into distinct typologies. Some women are primarily work-oriented, some primarily family-oriented, and some prefer to combine the two. Negative associations between women’s work and family behavior are primarily driven by the distribution of women across these typologies. Like other supply-side theories, preference theory recognizes the importance of structural constraints: Both the distribution of women across preference groups and the strength of associations between preferences and outcomes vary depending on institutional conditions (e.g., availability of part-time work, affordability of child care) and cultural context (Hakim 2003). Still, this perspective posits a primary role for women’s and men’s individual values in driving gender differences in work and family.

Examining adolescents’ orientations toward work and family life, as well as how those orientations shape their future plans, will help evaluate how supply-side explanations of gender inequality hold up among young people today. In particular, we ask whether young women and men differ in the constellation of work-family values they deem important and in the connection between these values and future goals.

Work and family orientations in young adulthood

A major goal of research on young adult development and status attainment is to understand the social processes that influence young people’s future plans. Adolescents’ aspirations and values motivate behavior in the transition to adulthood, guiding young people toward particular educational and occupational trajectories and personal life decisions. For instance, Mortimer, Staff, and Lee (2005) found that goals and values expressed early in life shape work, school, and family trajectories through the transition to adulthood. Prior research has also found that occupational aspirations, while not perfect predictors of attainment, are highly correlated with occupational attainment in adulthood (Schoon 2001; Sewell, Haller, and Portes 1969) and influence educational plans and behavior (Schneider and Stevenson 1999). Similarly, attitudes about marriage and childbearing are strong but imperfect predictors of future family behavior (Barber 2001; Quesnel-Vallée and Morgan 2003; Sassler and Schoen 1999; Willoughby 2012). Notably, these studies focus on slightly different forms of young people’s future orientations. Some research analyzes specific goals, either expressed as aspirations (what young people “would like” to do) or as expectations (what they “expect” to do). Other studies examine values and preferences, indicators of how favorable young people are toward certain life domains such as having children, marrying, earning a good income, or being committed to their career. Both matter. Specific goals help young people make decisions about their next steps, such as college enrollment, course taking, and volunteering opportunities. Preferences and values can help select those goals by pointing young people towards a particular type of career over another.

This connection between preferences and values, on the one hand, and future plans, on the other, is perhaps best elucidated by Eccles’s expectancy-value theory (Eccles 1994; Frome et al. 2006; Wigfield et al. 2010). According to this model, young people’s goals—inclusive of outcomes like occupational aspirations, but including other types of outcomes—are shaped by their beliefs regarding how well they will do on particular tasks (or types of tasks) and how much they value those tasks. Their goals in these domains, in turn, predict achievement-related choices and performance. This conceptual model can be applied to both short- and long-term goals and outcomes (Eccles 1994; Wigfield and Eccles 2000). In the short term, values across multiple domains can shape occupational choices (Eccles 1987, 1994). For example, Eccles (1994) describes gender differences in the relative value young women and young men place on work contexts that allow time for family versus work contexts that provide extrinsic rewards. She hypothesizes that these differences could lead young women and young men to make plans for different types of occupations that are perceived as either accommodating to family or not, or lucrative and high prestige or not.

Past research shows substantial gender differences in family- and work-related values, which subsequently predicted differences in occupational and educational aspirations (e.g. Cooper, Arkkelin, and Tiebert 1994; Eccles 1994; Frome et al. 2006; Herzog 1982; Wijting, Arnold, and Conrad 1977). However, much of this research finding large gender differences is outdated. More recent studies find that gender gaps in values have narrowed or even reversed in recent years. For example, the Pew Research Center found that young women are now more likely to rate their future careers as very important or one of the most important things to them than young men are (Patten and Parker 2012). These gaps in the value young women and young men place on careers appear to be reflected in their specific occupational aspirations. Young women hold career aspirations that require more education, on average, than the careers young men aspire to, although their target jobs are similar in median earnings (Howard et al. 2011). Other research finds, however, that although young women and young men appear to equally value extrinsic job rewards such as pay and prestige, young women are more likely than young men to state preferences for jobs that offer intrinsic rewards (such as security, autonomy, and time for leisure) and that fulfill an altruistic motivation (Hitlin 2006; Johnson 2002; Marini et al. 1996).

The gendered nature of work and family values may be more apparent when considering these domains jointly than when analyzing them in isolation. In the United States, where there are few structural or policy supports for managing the potentially competing demands of paid work and family caregiving (Collins 2019), work and family commitments are often in conflict. The cultural meanings of workplace achievement and family relationships can also be seen as opposing (Blair-Loy 2001; Hays 1998), at least for women; for men, work and family roles may be more complementary (Townsend 2002). To the extent that young people, and particularly young women, perceive work-family conflicts as relevant to their own futures, we might expect their work and family goals and values to be negatively correlated. In the 1970s and 1980s, women’s work and family preferences were strongly and negatively correlated (Rexroat and Shehan 1984; Shaw and Shapiro 1987; Stolzenberg and Waite 1977; Waite and Stolzenberg 1976). More recent findings about correlations between goals for work and family are mixed, and there is limited evidence regarding possible distinctions between values and preferences and more focused expectations. For example, Gerson (2010) finds that young people generally share egalitarian values when discussing ideal work-family arrangements, but that their “fallback” positions often reflect more traditional gendered family arrangements. Emerging research also finds that young people hold more traditional preferences regarding work and family arrangments while being tolerant of more egalitarian arrangements (Dernberger and Pepin 2020). Other research finds that some women consistently prioritize family and prefer not to work for pay (Damaske 2011), but most express strong commitment to both work and family (Barnes 2015; Blair-Loy 2001; Damaske 2011; Garey 1999).

Substantial bodies of research have examined young people’s work and family preferences and values and how they differ by gender, but less is known about how attitudes in these two domains are related for young people and how they are associated with educational and occupational expectations. Structural and normative constraints on women’s labor force participation have declined rapidly, and young women’s and young men’s work and family values, including their orientations toward combining work and family, have likely changed as well. The degree to which they have truly aligned or remain differentiated and how this shapes their occupational and educational aspirations, however, is unclear.

Predictors of work-family orientations

Other factors besides gender also shape which women (and men) share similar work-family orientations. For example, as Dawn Dow writes about in Mothering while Black, African American mothers face different kinds of pressures than do white mothers, including the expectation that they will work outside the home (2019). Given the long tradition of Black women supporting families on their earnings, Black adolescents may be more likely than white adolescents to place equal value on home and work. On the other hand, given the high value placed on family in Hispanic communities (Landale and Oropesa 2007), Hispanic adolescents (both young women and young men) may value family outcomes relatively higher than do whites and Blacks.

Socioeconomic status, family structure, and school performance may also be associated with adolescents’ work and family values. Young people from more socioeconomically disadvantaged families may value work more highly than those from relatively advantaged families, both as an avenue out of economic disadvantage (Silva and Snellman 2018) and because they are more likely to have had work experiences in adolescence (Staff and Mortimer 2008). Prior research also suggests that attitudes toward family life are influenced by the type of family structure one grows up in (Amato and DeBoer 2001; Cunningham and Thornton 2006). Therefore, we expect young people living with both married parents will value family outcomes such as marriage more highly than those living with a single parent or step-parent. Finally, we expect that academic achievement and work values may be reciprocally associated with one another; academic achievement may prompt interest in demanding occupations, thus heightening youths’ interest in work, while caring deeply about one’s own occupational future may also influence young people’s dedication to academic performance.

Current study

In the current study, we ask, first, how gender shapes young people’s values for work and family outcomes, and then how both gender and typologies of work and family values are associated with specific educational and occupational expectations. To answer these questions, we use nationally representative survey data to examine associations between values for work and family outcomes among young (high school-age) women and men. We study adolescents’ values for their future work and family in order to better understand how they think about their adult lives and the degree to which they perceive work and family as conflicting. Results are relevant for understanding the gendered nature of work and family schemas (Blair-Loy 2001). Findings shed light on how young people’s values for work and family-related outcomes group together and whether there are differences in these groupings by gender. Additionally, we examine whether work and family value groupings are associated with educational and occupational expectations, and whether these associations differ for boys and girls, to assess how general values are related to more specific plans.

Data and methods

Data and sample

We use data from the Education Longitudinal Study (ELS) of 2002, a nationally representative survey of U.S. secondary students conducted by the National Center for Education Statistics (NCES). This study is one in a series of panel datasets that collect data on a random sample of young people, usually starting in high school, and follow them into young adulthood. The aims of the study are to “study the educational, vocational, and personal development of students at various stages in their educational careers and the personal, familial, social, institutional, and cultural factors that may affect that development” (p. 2, Ingels et al. 2014). The base year of data collection in 2002 sampled high school sophomores (students in their second year of secondary school). We use data from the first follow-up, in 2004, when most of the original sample members were seniors in high school (in their final year of secondary school). Respondents were between age 16 and age 21 at the time of interview, with 98% of the sample age 17–19. The follow-up sample was refreshed to ensure that it was nationally representative of seniors; in all, 14,930 adolescents were interviewed in 2002. The large majority of these respondents (n=13,370) were enrolled in grade 12, but the sample also includes students enrolled in other grades (n=278), homeschooled students (n=41), early graduates (n=559), and dropouts (n=682). We include the full sample in analyses; results are virtually identical when limited to students enrolled in grade 12 at the time of interview. We exclude cases with missing data using listwise or pairwise deletion (depending on the analysis). Our analysis has multiple components, with a different set of variables included in each component, and thus the number of cases excluded varies across components of the analysis. We describe the analytic sample for each component in the “Missing data” section below.

Measures

One section of the survey was devoted to plans and expectations about the future. In this section, students were asked, for a series of eighteen items, “How important is each of the following to you in your life?” Response options were not important, somewhat important, and very important. We focus on eight core items related to work and family values. (A full list of all eighteen items is provided in the Appendix). We labeled as “work-focused” five items that related to professional activities or achievement. The work-focused items included in the analysis are “Being successful in your line of work”, “Being able to find steady work”, “Becoming an expert in your field of work”, “Getting a good job”, and “Having lots of money.” Based on exploratory analysis, we excluded “Getting a good education” from the work-focused items because there was very low variation in responses to this question – virtually all surveyed students reported that education was very important. The family-focused items included in the analysis were “Finding the right person to marry and having a happy family life”, “Having children,” and “Living close to parents and relatives.” These items incorporate relationships with spouse, children, and family of origin. Respondents were not asked to compare values for work and family outcomes. However, because the importance of work-related items and family-related items are measured on the same scale, we can make inferences about the relative value placed on different domains.

Respondents were also asked about specific educational and occupational expectations. Educational expectations were measured by asking “As things stand now, how far in school do you think you will get?” We categorized educational expectations as less than a four-year college degree, a four-year college degree, more than a four-year college degree, or don’t know. Occupational expectations were derived by asking respondents the job they expected or planned to have at age 30. These responses were recorded verbatim. For respondents who could provide an answer, NCES classified occupations into broad categories (e.g., craftsperson, sales, laborer, professional) and then used a modified version of the Duncan prestige scale to rank the categories. The largest single categories were two categories of professional workers; approximately half of the analytic sample reported career goals that fell into one of these two categories. About 30% of the analytic sample responded “don’t know” to the question about occupational expectations; we treat this response as a separate category. (The proportion of don’t know responses is much higher than in the National Education Longitudinal Study of 1988, the previous NCES longitudinal survey (Staff et al. 2010). This very high level of “don’t know” responses may be attributable to the format of the question, which required respondents to write in an exact occupation rather than choosing categories as in the NELS.) The other occupational categories made up less than 5% of the sample each. In order to capture the broad characteristics of high school students’ career goals, we collapsed responses into five categories based on the prestige scale: advanced professional (requires advanced degree), professional (does not require advanced degree), middle (including categories such as managerial, clerical, proprietor), low (including categories such as laborer, craftsperson, and service), and “don’t know”.

In addition to these outcomes, we include predictors of adolescent work-family typologies. We measure explanatory factors related to family background (race-ethnicity, family structure, socioeconomic status) and academic skills (math scores). Race-ethnicity is measured as non-Hispanic white, non-Hispanic black, Hispanic (all races), and other race (non-Hispanic). Students reporting multiple races are included in the “other race” category. Family structure is dichotomized as intact (living with married biological mother and biological father) and non-intact (all other family structures). Socioeconomic status is measured in quartiles constructed by NCES based on a composite measure of parental occupation and education and family income. Based on the results of exploratory analysis, we collapsed quartiles into high (top two quartiles) and low (bottom two quartiles) SES. Math scores come from a cognitive test administered as part of the ELS questionnaire; we use quartiles of scores to represent ranges of math ability. (Although the base year ELS survey also included a test measuring reading ability, only math tests were included in the first follow-up).

Analysis

Our analysis begins with basic descriptive statistics on the distribution of work-related and family-related values and the correlations between variables for young women and young men. To understand the associations among the importance attributed to work-related values and family-related values, we next apply latent class analysis (LCA), incorporating predictors of class membership. We then use descriptive statistics to show asssociations between typologies of work and family values and specific educational and occupational goals.

LCA assumes that observed correlations between measures are explained by the presence of discrete groups (classes) within the population (Collins and Lanza 2010; McCutcheon 1987). That is, this method assumes the presence of distinct “types” of people characterized by different combinations of life values and goals. LCA is a person-centered approach, in that it considers configurations of attitudes as they appear within individuals, rather than a variable-centered approach that might generate latent variables representing degree of valuing family and degree of valuing work and estimate the association between these latent variables (see Howard and Hoffman 2018; Muthén and Muthén 2000; see also Kossek et al. 2012 for an example combining person-centered and variable-centered approaches). When conducting LCA, the analyst specifies a number of classes (or types). Given that number of classes, software is used to estimate the distribution of respondents across classes with the highest likelihood of having produced the observed distribution of variables. The outputs of the model are the proportion of the sample in each class, the distribution of each of the variables in each class, and the overall model fit, that is, how well the derived solution fits the observed data. Typically, models are estimated for a range of numbers of classes, and the analyst selects the model with the best combination of statistical fit and substantive plausibility (Collins and Lanza 2010). The number of classes in this final model is also considered an output of a latent class analysis. We estimate conditional models that take into account individual characteristics as well as values on the outcome measures in assigning the likelihood of various group membership possibilities. In exploratory analyses, we also estimated unconditional models; results did not differ substantively. We present conditional models here because these models are more efficient in estimating the associations between individual characteristics and group membership (Collins and Lanza 2010). After determining the best-fitting latent class model, we formally test whether the composition and distribution of classes are different for young women and young men and compare the distribution of young women and young men across classes. All models are estimated using a downloadable Stata add-on (http://methodology.psu.edu) in Stata 15.0.

Missing data

Overall, levels of missing data in the ELS are low. Many variables, including all of the variables used to predict class membership in the LCA conditional models, were multiply imputed by NCES statisticians and included in the public-use dataset. Where available, we use these imputed values. For variables without imputed values (work- and family-related values, educational and occupational expectations), our treatment of missing data varies depending on the type of analysis.

Our initial descriptive analyses of work- and family-related values use casewise deletion and pairwise deletion. For univariate descriptive statistics, we present distributions of responses among adolescents with valid responses. For correlations between measures, we use cases with valid responses on both measures. These approaches allow us to keep the greatest number of cases possible and keeps our sample size close to the sample sized used for LCA, which allows for the inclusion of incomplete cases (described below). Specific sample sizes for each item and each pairwise correlation are shown in Tables 1 and 2, respectively.

Table 1.

Percent of respondents reporting each item is “very important”

	# of valid cases		% responding very important
	Young men	Young women	Young men	Young women
Work-related items
Being successful in your line of work	7391	7445	89.3	92.2	***
Being able to find steady work	7366	7413	85.5	88.8	***
Becoming an expert in your field of work	7361	7423	76.9	73.3	***
Getting a good job	7347	7410	92.1	94.7	***
Having lots of money	7387	7445	43.3	28.1	***
Family-related items
Finding the right person to marry and having a happy family life	7387	7440	79.5	81.1
Having children	7363	7413	44.2	52.6	***
Living close to parents and relatives	7350	7404	23.7	28.4	***

Source: Education Longitudinal Study: 2002; followup 1, 2004, student questionnaire.

^*:p<.05;

^**:p<.01;

^***:p<.001; chi-square tests of differences between young women and young men. Weighted using cross-sectional weights.

Table 2.

Correlations between work-related and family-related items (pairwise sample size in parentheses)

	Finding the right person to marry, happy family	Having children	Living close to parents, relatives
	Young men
Being successful in your line of work	.17	.10	.05
	(7378)	(7354)	(7341)
Being able to find steady work	.18	.12	.08
	(7357)	(7334)	(7322)
Becoming an expert in your field of work	.06	.10	.10
	(7349)	(7333)	(7318)
Getting a good job	.15	.11	.09
	(7334)	(7318)	(7303)
Having lots of money	.01	.06	.14
	(7371)	(7347)	(7339)
	Young women
Being successful in your line of work	.10	.04	.02
	(7431)	(7406)	(7397)
Being able to find steady work	.10	.08	.05
	(7400)	(7377)	(7366)
Becoming an expert in your field of work	.03	.02	.07
	(7410)	(7391)	(7376)
Getting a good job	.06	.06	.05
	(7397)	(7376)	(7363)
Having lots of money	.02	.01	.07
	(7432)	(7405)	(7396)

Source: Education Longitudinal Study: 2002; followup 1, 2004, student questionnaire.

The central component of our analysis is the identification of work-family typologies using latent class analysis. LCA uses maxiumum likelihood estimation methods; thus, cases with incomplete data can be included in the analysis. Only cases with missing data on all measures of work- and family-related values (i.e., cases that contribute no information to the estimation of class structure and membership; n=68, 0.46%) are excluded from the LCA, resulting in an analytic sample of 14,862 respondents for the LCA.

To examine the correlation of class membership with educational and occupational expectations, we use casewise deletion to exclude cases with missing values for expectations. Because there are substantial numbers of respondents who do not know their educational and occupational goals, we treat these responses as meaningful outcomes rather than missing data and include them as a category, as described above in the section on measures. In the analytic sample of respondents included in the LCA, 45 (0.30%) had missing data on educational expectations. For occupational expectations, levels of missing data were higher. In the LCA sample, there were 380 respondents (2.56%) without valid responses. In addition, respondents who wanted to be in military service (n=141, or 0.95% of the full sample), those who wanted to be homemakers (n=11, 0.07%), and those who reported that they did not plan to be working at age 30 (n=97, 0.65%) were not assigned a status ranking and so could not be compared to other responses. Because these categories were so small, it is difficult to draw conclusions about distributions across latent classes in these groups, and we excluded them from the table.

Results

Descriptive results

Table 1 shows the proportion of young women and young men reporting “very important” for each of the work- and family-related items. Many of the items are highly valued by most respondents; overall, work-related items are reported to be very important more often than family-related items. The highest rating for importance is for “getting a good job” – 94.7% of young women and 92.1% of young men responded “very important” for this item. Being successful at work was also rated very important by most adolescents. Among the family-related items, around 80% of both young women and young men believed that finding the right person to marry and having a happy family life was very important. Fewer respondents felt it was very important to have children (52.6% of young women and 44.2% of young men), and even fewer said it was very important to live close to parents and relatives (28.4% of young women and 23.7% of young men). These measures reflect values, not specific aspirations and expectations. Thus, for example, although relatively few young men and women say that having children is very important, nationally representative data from another source show that for a similar age group at this time point, about 90% of young men and 87% of young women intended to have at least one (more) child at some point in their life (Chandra et al. 2005; Martinez et al. 2006).

On the whole, young women express more enthusiasm for both work and family outcomes than young men. Young women are significantly more likely than young men to respond “very important” on five of the eight items. However, significantly more young men than young women report that it is very important to earn lots of money and become an expert in their chosen field. Thus, although both young women and young men appear to value labor force participation, young men place more value on money and status generated by jobs. Relative to young men, young women place more emphasis on family relationships. Given the large sample size, these differences are statistically significant, but they are not large in magnitude.

Table 2 shows Pearson correlation coefficients between the work-related and family-related items for young men (top panel) and young women (bottom panel). Responses of “very important” are given a score of 3, “somewhat important” a score of 2, and “not important” a score of 1. For ease of presentation, we do not show correlations within work items and family items. Most correlations are positive but small in magnitude – the largest correlation coefficient is only .17. There appears to be little relationship between work and family values, and no evidence that adolescents who rate work-related items more important are likely to rate family-related items less important, or vice versa. In fact, on average, adolescents who rate family-related items more important are more likely to also rate work-related items important. The positive associations between work and family values are somewhat stronger for young men than for young women for most items. For being successful at work, having steady work, becoming an expert, and getting a good job, correlation coefficients are on the order of twice as large for young men as for young women (though still modest in size).

Latent class analysis

For both young women and young men, a three-class model best described the association among work- and family-related values. We tested models with up to nine classes. Based on fit statistics, either a three-class model or a four-class model would be preferred; although model fit continued to improve when additional classes were added, the improvement was minimal (on the order of 10% or less change in adjusted BIC) with more than four classes. We selected the three-class rather than the four-class model as the most appropriate model based on substantive grounds. The three largest classes in the four-class model were essentially the same as the three-class model, while the fourth class was very small (less than 5% of the sample) and quite similar in characteristics to one of the other classes. Table 3 presents results from the three-class model. The table shows the proportion of the sample classified in each group (based on posterior probabilities) and the proportion of each group that responded “very important” to each item. For ease of presentation, we do not present the proportion responding “somewhat important” or “not important,” but these outcomes are included in the underlying model. Young men are in the top panel of the table and young women are in the bottom panel.

Table 3.

Group membership and item responses based on latent class analysis

	Low work, medium family: “Indifferent”	High work, high family: “Strivers”	High work, low family: “Careerists”
	Young men
Proportion of sample in each group	0.14	0.60	0.26
	Proportion of group responding that each item is “very important”
Being successful in your line of work	0.48	0.97	0.93
Being able to find steady work	0.35	0.95	0.89
Becoming an expert in your field of work	0.33	0.83	0.83
Getting a good job	0.48	0.99	0.98
Having lots of money	0.14	0.46	0.50
Finding right person to marry, happy family life	0.61	0.99	0.45
Having children	0.27	0.66	0.07
Living close to parents and family	0.11	0.27	0.21
	Young women
Proportion of sample in each group	0.12	0.60	0.28
	Proportion of group responding that each item is “very important”
Being successful in your line of work	0.51	0.98	0.97
Being able to find steady work	0.47	0.96	0.91
Becoming an expert in your field of work	0.21	0.77	0.81
Getting a good job	0.56	1.00	0.99
Having lots of money	0.09	0.29	0.34
Finding right person to marry, happy family life	0.71	0.98	0.51
Having children	0.41	0.77	0.07
Living close to parents and family	0.20	0.34	0.21

Latent class analysis of associations between work- and family-related values items. Education Longitudinal Study: 2002; followup 1, 2004, student questionnaire. N=14862 respondents (7408 young men, 7454 young women) with non-missing data on at least one values item.

Statistical tests for measurement invariance find that gender differences in model structure are statistically significant. However, substantively, the models are quite similar in terms of both the characteristics of the classes and the distribution of students across classes. Thus, our results show only minimal differences between young women and young men in work and family values and how those values are combined. We describe the characteristics of groups and the predictors of group membership together for young women and young men, pointing to specific gender differences where relevant.

For both young women and young men, the largest group of students places high importance on both work and family goals. We label this class “strivers.” This group makes up about 60% of the sample for both groups. The next largest class (“careerists”; 28% of young women, 26% of young men) also rates work values as very important, but is less likely to say that marrying and having children are very important. In particular, only 7% of young women and 7% of young men in this group say that having children is very important. Finally, the smallest class (12% of young women and 14% of young men) is in the middle in both work-related items and family-related items. We term this class the “indifferent”: on almost all items, less than half of young women and men in this group responded that the item was very important. In this group, the most important item for both young women and young men was getting married and having a happy family life, but even marriage and family was rated less important than in the “strivers” group. Neither young women nor young men exhibit any evidence of a group that devalues professional success in favor of family goals. That is, there is no group of young people that appears to value family goals exclusively and not value engagement in the paid labor force.

Table 4 shows associations between group membership and individual and family characteristics. The table shows coefficients from multinomial logit models predicting the log-odds of group membership; coefficients are generated as part of the LCA model. Again, young men are shown in the top panel of this table and young women in the bottom panel. The “strivers” (group 2), students who rate both work and family items as very important, are the reference category.

Table 4.

Predictors of group membership (coefficients)

	Reference: strivers	Indifferent			Careerists
		b	SE		b	SE
		Young men
Intercept		−1.15	0.08	***	−1.32	0.09	***
Non Hispanic white
Non Hispanic black		−1.19	0.21	***	0.53	0.10	***
Hispanic		−0.51	0.14	***	0.22	0.10	*
Non Hispanic other		0.00	0.11		0.33	0.10	***
High SES
Low SES		−0.04	0.09		0.31	0.07	***
Highest math quartile
Third math quartile		−0.40	0.11	***	−0.13	0.10
Second math quartile		−0.43	0.12	***	0.08	0.10
Lowest math quartile		−0.20	0.12		0.04	0.10
Intact family
Non-intact family		0.13	0.09		0.36	0.07	***
		Young women
Intercept		−1.23	0.09	***	−1.38	0.09	***
Non Hispanic white
Non Hispanic black		−1.10	0.25	***	1.03	0.10	***
Hispanic		−0.14	0.14		0.64	0.10	***
Non Hispanic other		0.24	0.12	*	0.66	0.10	***
High SES
Low SES		0.06	0.10		0.20	0.07	**
Highest math quartile
Third math quartile		−0.57	0.11	***	−0.13	0.10
Second math quartile		−0.83	0.13	***	0.01	0.10
Lowest math quartile		−0.45	0.13	***	0.00	0.10
Intact family
Non-intact family		0.07	0.10		0.40	0.07	***

Latent class analysis of associations between work- and family-related items. Education Longitudinal Study: 2002; followup 1, 2004, student questionnaire. N=14862 respondents (7408 young men, 7454 young women) with non-missing data on at least one values item.

^*:p<.05;

^**:p<.01;

^***:p<.001.

There are more gender differences in the predictors of group membership than in the structure of classes, although again the overall pattern of results is similar across gender. Relative to non-Hispanic white adolescents, Black adolescents are less likely to be classified in the “indifferent” group and more likely to be in the “careerist” group compared to the “strivers” group. Hispanic young women are less likely than white young women to be classified as “indifferent,” and both Hispanics and respondents of other races (not Black or white) are also more likely to be classified as “careerists.” These latter associations are stronger for young women than young men. Overall, then, respondents from disadvantaged racial-ethnic groups, and especially young women from these groups, appear to rate family relationships as relatively less important than work. Socioeconomic status and family structure are also associated with latent class membership – students in low SES families and non-intact families are relatively more likely to be classified in the careerist group than the striver group. Coefficients are slightly larger for young women than for young men, but differences are minimal.

In terms of academic skills, math test scores are associated with differences between the “indifferent” and “striver” group, but not with differences between “careerists” and “strivers.” Relative to those with the highest test scores, young people with lower achievement scores are less likely to be in the “indifferent” group. Together, these associations suggest that adolescents with lower math achievement scores are likely to be more enthusiastic about work goals. This result is somewhat surprising – we expected that high academic skills would be associated with enthusiasm for work.

Gender differences in general values for work and family achievement are of interest to the extent that they predict differences in more specific educational and occupational goals. In Tables 5 and 6, we present differences between young women and young men and differences across latent classes in educational and occupational expectations, as measured by how far the respondent expects to go in school and the job the respondent expects at age 30. Consistent with previous research, young women are more ambitious than young men, especially in terms of education. Thirty-eight percent of young women expect to continue in school beyond a bachelor’s degree, compared to only 27% of young men, and young women are more likely to expect an advanced professional occupation and less likely to expect a low status occupation than young men.

Table 5.

Educational expectations (proportion in each educational category), by gender and group membership

	Less than BA	BA	More than BA	Don’t know
Full samplea
Young men	0.31	0.32	0.27	0.11
Young women	0.22	0.31	0.38	0.09
Both sexes	0.26	0.31	0.32	0.10
Young men
Indifferent	0.34	0.32	0.22	0.12
Striversa	0.28	0.33	0.30	0.09
Careeristsa	0.35	0.29	0.21	0.15
Young womenb
Indifferent	0.28	0.33	0.27	0.13
Strivers	0.21	0.32	0.40	0.07
Careerists	0.23	0.28	0.37	0.11

N=14,817 respondents (7382 young men, 7435 young women) with non-missing data on at least one work or family values item and non-missing data for educational expectations.

^a:gender differences in the distribution of expectations within latent class are significant, p<.05, based on chi-square tests.

^b:latent class differences in the distribution of expectations within gender are significant, p<.05, based on chi-square tests.

Table 6.

Expected occupation at age 30 (proportion in each occupational category), by gender and group membership

	Low	Medium	Professional	Advanced professional	Don’t know
Full samplea
Young men	0.17	0.17	0.24	0.09	0.33
Young women	0.12	0.16	0.26	0.18	0.28
Both sexes	0.15	0.16	0.25	0.14	0.30
Young menb
Indifferent	0.11	0.15	0.22	0.06	0.46
Striversa	0.18	0.17	0.25	0.10	0.30
Careeristsa	0.19	0.17	0.22	0.09	0.33
Young womenb
Indifferent	0.11	0.14	0.23	0.10	0.43
Strivers	0.13	0.17	0.27	0.18	0.25
Careerists	0.13	0.14	0.27	0.20	0.27

Low: service, laborer, operative. Medium: clerical, sales, technical. N=14,482 respondents (7172 young men, 7310 young women) with non-missing data on at least one work or family values item and non-missing data for career expectations, excluding respondents who expected to be in the military, homemakers, or out of the labor force.

^a:gender differences in the distribution of expectations within latent class are significant, p<.05, based on chi-square tests.

^b:latent class differences in the distribution of expectations within gender are significant, p<.05, based on chi-square tests.

As might be expected, both educational and occupational expectations are higher in the work-focused groups (strivers and careerists) and lower in the indifferent group, which places less importance on work achievement. In addition, “don’t know” responses are most common in the indifferent group, suggesting that this response reflects at least in part lower enthusiasm for work achievement. Differences between the striver and careerist groups are relatively small; for young men, expectations in the striver group are even slightly higher than in the careerist group. That is, the higher importance placed on family by adolescents classified in the “striver” group does not appear to be associated with lower expectations for educational or occupational achievement.

Gender differences in educational and occupational expectations are smaller for the indifferent group than in the striver or careerist groups. For example, 27% of young women and 22% of young men in the indifferent group expected to receive more education than a bachelor’s degree, a difference of about 5%. Among strivers, in contrast, 40% of young women and 30% of young men expected the highest level of education, a difference of about 10%. This difference is greater both proportionally and in absolute terms. For both educational expectations and occupational expectations, gender differences within the indifferent group are not statistically significant. Thus, the biggest gender differences in professional expectations are found among the more ambitious students, and gender differences are smaller among less ambitious students. This finding shows that gender differences in educational and occupational goals are not fully explained by work-family orientations, but persist even within typologies of work-family orientations.

Discussion and conclusions

Despite the dramatic increases in women’s educational attainment and labor force participation over the past four decades, substantial gender differences in work hours, earnings, and occupational prestige persist. Similarly, although gender differences in parenting and relationship dynamics have declined, women continue to be more engaged than men with household and family responsibilities. Multiple processes likely contribute to these differences. First, prior research documents how men and women end up performing different family and work roles due to the constraints and opportunities they face in both domains (Bobbitt-Zeher 2011; Daminger 2019; McLaughlin, Uggen, and Blackstone 2012; Reskin and Roos 1990; Riegle-Crumb et al. 2012; Riegle-Crumb and Humphries 2012; Sassler and Miller 2017; Schneider 2011). However, differences in the kinds of outcomes men and women value may also be a factor; women may hold stronger pro-family values while men may hold stronger pro-work values, and these values may shape the kinds of occupations and family roles they pursue. To evaluate this argument, previous research has demonstrated variation in women’s goals for work and family roles (Eccles 1987, 1994; Hakim 1998, 2000). However, prior work has not examined whether the values adolescents place on work and family outcomes vary systematically by gender, nor whether gender shapes how these values are associated with educational and occupational expectations. In this paper, we fill this gap in the literature by estimating gender differences in the value placed on work and family achievement in adulthood by adolescents as a possible starting point for differences in attitudes and behavior in adulthood.

Basic descriptive statistics do not show overwhelming differences between high school-age young women and men in how they value work and family achievements or in the relationship between these perceptions. Both young women and young men think both domains are important, and values in work and family domains are positively correlated for both young women and young men. Although both young women and young men place high importance on professional goals, young men are more likely to express the importance of money and status as parts of work achievement. Furthermore, positive associations between work and family goals are slightly stronger for young men than for young women, suggesting that young men, more than young women, may see work and family goals as complementary pieces of a “package deal” (Townsend 2002).

Latent class analysis of the associations between the value placed on professional achievement and family relationships found similar patterns for young women and young men. For both young women and young men, LCA identified a group of “strivers” who placed high importance on both work and family; a group of “careerists” who valued professional achievement, but rated marriage and children less important; and a small “indifferent” group who placed relatively low importance on both domains. Associations between latent class membership and educational and occupational expectations revealed few gender differences. Across gender, being in the indifferent group was associated with lower expectations in both domains. We did find small gender differences within the striver and careerist latent classes where young women held, on average, higher expectations for both education and occupation than young men. Based on these measures of expected outcomes, work-oriented young women do not appear more likely than young men with similar values to anticipate facing structural barriers that could interfere with educational or career achievement, such as discrimination or difficulty combining work and parenthood.

Overall, we find relatively little evidence to support the argument that gender differences in work-family preferences and values may explain gender differences in occupational goals. In particular, our results are inconsistent with preference theory (Hakim 1998, 2000), which posits that men and women have different preferences for combining family and that a substantial subset of women prefers to withdraw from the labor force to fulfill family responsibilities. According to preference theory, there are clearly identifiable groups of work-oriented and family-oriented women, along with a larger group of women who would like to combine both domains but who are not usually interested in building careers and may work intermittently or part-time. Among high school aged young women, we do not find these patterns. Rather, most young women are work-oriented, in the sense that they say that work values are very important, and the largest group of young women says that both work and family are very important – a group not anticipated by preference theory. Although Hakim theorizes that a large group of “adaptives” are interested in both having families and working, she anticipates that these women do not place a high value on a career or even working full-time, whereas our group of strivers rate work-related outcomes more highly than the other groups and are particularly enthusiastic about having a career and becoming an expert in their field. They also hold high educational and occupational expectations. Our results that young women and young men have similar combinations of work and family values are also inconsistent with preference theory. Recent research on work and family orientations among adults has begun to identify distinct outlooks that integrate work and family priorities, rather than setting them in opposition. For instance, “dual-centric” individuals place high value on both work and family identities (Galinsky 2014; Kossek et al. 2012), and, particularly in times of economic uncertainty, having a “good” job and providing for the family can be an important element of a family identity (Masterson and Hoobler 2015). Our findings are more consistent with these integrated orientations toward combining work and family than with orientations that oppose the two spheres.

Our findings provide mixed support for expectancy-value theory. We find fewer gender differences in work and family values than anticipated in Eccles’ work. As noted above, young men showed greater interest in status and monetary outcomes than young women. However, we also found support for Eccles’ proposition that values and goals are associated with one another; we showed how young people in the striver and careerist classes were more likely to expect to complete high levels of education and work in relatively high status jobs than those in the indifferent class. Although we are unable to examine these associations longitudinally—and thus cannot say for sure that values affect occupational and educational expectations—our findings do lend support for this theory.

It is possible that our measures do not capture existing gender differences. For instance, although other research finds that most adolescents intend to marry and have children, relatively few students in our sample rate marriage and having children as “very important” in their lives. It may be that marriage and childbearing are understood as more personal decisions than work achievement, or it may be that adolescents are more likely to focus on paid work given the academic and achievement focus of the survey. Furthermore, the survey items available for importance of work were more numerous and nuanced than those available for family life. More detailed measures of family goals might find different associations with work goals or greater gender differences. Similarly, the measures used here for educational and occupational goals are broad measures that may not capture nuances in life plans.

It is also possible that high school students genuinely exhibit fewer gender differences in work-family orientations than do adults. For most adolescents in this sample, college, “good jobs”, marriage, and childbearing are still abstract goals rather than actual experiences. Thus, the importance they place on professional and family roles in adulthood is largely based on ideals rather than reality, and ideals are influenced by the egalitarian and achievement-oriented ideology prevalent in American youth culture. These young people likely observe work-family conflicts that arise among their parents and other adults, but they may not anticipate the same problems in their own lives. As they age, these young people may encounter difficulties in enacting their ideals. A lack of policies that support combining parenthood and paid work can impact career trajectories (Collins 2019; Damaske 2011), and women are not always able to enact gender egalitarianism within romantic relationships (Gerson 2010; Sassler and Miller 2017). Socioeconomic status may also condition the degree to which young people are able to accomplish their professional goals. Orientations toward the future are not static. This study can only speak to gender differences at one point in time, when young people are beginning to select and plan for their future roles. Further analysis of goals and values at different stages of the life course, including longitudinal anlaysis of within-indivdidual change, is needed to understand how egalitarian schemas in adolescence evolve in the face of unequal experiences.

Appendix:

ELS Items for the Question “How Important is Each of the Following in Your Life?”

Work-related items

Being successful in your line of work

Having lots of money

Being able to find steady work

Getting away from this area of the country

Becoming an expert in your field of work

Getting a good education

Getting a good job

Family-related items

Finding the right person to marry and having a happy family life

Having strong friendships

Being able to give your children better opportunities than you’ve had

Living close to parents and relatives

Having children

Having leisure time to enjoy your own interests

Other items

Helping other people in your community

Working to correct social and economic inequalities

Being an active and informed citizen

Supporting environmental causes

Being patriotic

Source: Education Longitudinal Study: 2002; followup 1, 2004, student questionnaire. Items in italics are included in analyses.