First-Generation Students, College Majors, and Gendered Pathways

Abstract

Emerging literatures have highlighted the social- and resource-related inequalities among first-generation college students. Less attention has been devoted to the curricular pathways (i.e., college majors) these students follow and their potentially gendered character. We build on educational inequality and gender literatures in this article, and arguments surrounding habitus and class-based dispositions to address this gap. Our analyses draw on several waves of the Education Longitudinal Survey (ELS-2002) merged with national data on sex composition of fields of study. Our results suggest unique pathways in college for first-generation compared to continuing-generation students. Specifically, first-generation students are more likely to choose occupationally specific “applied” majors than their continuing-generation counterparts. Modeling by gender reveals little to moderate variation between first- and continuing-generation students’ representation in female-dominated majors. These patterns generally hold for 2- and 4-year college going samples. We conclude by discussing the relevance of these findings for educational inequality, eventual job returns, and occupational mobility.

Americans have long pointed to college degree attainment as a main driver of social mobility (Hout 2012; Torche 2011). Despite potential risks, the benefits of a college degree certainly outweigh the costs, and with the most disadvantaged students benefitting the most (Abel and Deitz 2014; Hout 2012). It is also the case, however, that differences in curricular pathways (i.e., fields of study) can still perpetuate inequalities among degree holders (Roksa and Levey 2010). This may be especially pertinent to our population of interest in this article, namely first-generation college students (i.e., those whose parents do not hold a bachelor’s degree or higher).

We know from prior literatures that first-generation students experience disadvantage when it comes to the type of post-secondary education received (Charles and Bradley 2002; Roksa and Levey 2010). This might include underrepresentation in fields of study that lead to higher career earnings and job security (Cataldi, Bennett, and Chen 2018; Van Noy and Ruder 2017). Given that unique gendered processes are often seen as the root of segregation in college major selection (Vaisey 2010), there may also be important intersections of gender and first-generation status that warrant consideration. Our central focus in this article is the possibility of such dual-pronged status variation in college major choices.

We begin by building on influential and important literatures surrounding inequality, education, and family that are informative for our analyses. We know from these literatures, for instance, that parents’ work values and occupational experiences tend to be transmitted to children and have implications for children’s valuations of work and education. One consequence, we suspect, is that first-generation students who are mostly of working-class background are more likely to approach higher education with a “utilitarian” view – a view that lends itself to job-specific credentials and practical skills that are assumed to be important for securing more immediate employment (Kohn 1969; Lehmann 2009; Parcel and Menaghan 1994a, 1994b; Weis 1990; Wilkins 2014). Gender, however, should also be considered given that gendered socialization, norms, and expectations shape higher education orientations and occupational pathways (Bobbitt-Zeher 2007; Kim, Tamborini, and Sakamoto 2015; Mickelson 2003). How precisely do first-generation background and gender intersect in these regards? Are there observable variations in applied versus academic majors and in the gendered character of major selection?

Our analyses, which draw on several waves of the Education Longitudinal Study (ELS: 2002), are focused on: (1) whether first-generation women and men are more likely than their continuing-generation counterparts to choose more applied (versus academic) majors – i.e., majors characterized by occupationally specific training with clearer links to the labor market; and (2) whether there are observable variations and a simultaneous filtering into gender-typed majors. Our attention to the patterning of college majors by first-generation status and gender is pertinent to inequality, mobility, and gender literatures; makes explicit status intersections that may be meaningful in the course of educational mobility; and highlights educational pathways that likely have implications for intergenerational labor market inequalities and ongoing patterns of occupational sex segregation. We elaborate on these contributions in our conclusion. We also discuss the broader implications of our findings for research on higher education generally and first-generation college students specifically.

First-generation Students, Family Background, and Higher Education

First-generation students are generally more likely to be female, older, Black or Hispanic, have dependent children, and come from low-income families compared to their continuing-generation counterparts (Engle 2007). They are also more likely to enter higher education with limited resources, skills, and knowledge of the college environment compared to those whose parents have college degrees (Hurst 2010; Lee 2016; Wilbur and Roscigno 2016).

Especially pertinent to background inequalities is cultural capital and, specifically, the taken-for-granted assumption that one will pursue higher education in the first place; something that is especially strong among students from more affluent backgrounds (Grodsky and Riegle-Crumb 2010). Without such cultural capital and its connection to expectations and knowledge of the college experience, first-generation students often feel isolated when traversing the college experience (Cataldi, Bennett, and Chen 2018; Jack 2014; Stuber 2009; Wilbur 2021; Wilbur and Roscigno 2016) while also experiencing pressures to conform to middle-class norms and behaviors (Hurst 2010; Lee 2016). Feelings of confusion and isolation while trying to be educationally mobile lead to difficulties in balancing connections and responsibilities to one’s family (London 1989; Terenzini et al. 1996; Wilbur 2021).

Such patterns are borne out by research on the college years themselves. For instance, first-generation college students are less likely participate in academic and social activities – activities often associated with college success – and rarely take advantage of institutional resources to facilitate their transition to college (Roksa and Silver 2019; Stuber 2009; Stuber, Klugman, and Daniel 2011). Further, they are less likely to participate in academic or social clubs or to form close ties to peers and faculty outside of the classroom (Nuñez and Cuccaro-Alamin 1998; Pascarella et al. 2004, 2003; Terenzini et al. 1996). Limited academic and social integration are partly the result of working longer hours per week for the purpose of paying tuition, fees, and living expenses (Pike and Kuh 2005; Wilbur and Roscigno 2016). Such disparities along with work obligations and depressed levels of interaction with peers and faculty ultimately impact academic and extracurricular engagement (Lee 2016; Silver 2020; Stuber 2011).

First-generation status along with many of the potential inequalities that it may entail, of course, are also associated with other dimensions of social class background such as parental occupation and income. There is indeed some evidence suggesting that first-generation, low-income, and working-class students share many of the same barriers to higher education, particularly when it comes to information about college, different academic majors, and career paths (Horn and Nuñez 2000; Hoxby and Turner 2015; Ishitani, 2006). We believe that the first-generation focus is worthy of attention, however, given: its adoption as a core marker of status vulnerability by colleges and universities across the country; literatures pointing to first-generation inequalities above and beyond traditional indicators of social class (e.g., Wilbur and Roscigno 2016); and the possible connection between parental education, occupations and the college majors that their children end up choosing.

First-Generation Students and College Major Pathways

First-generation background has clear implications for educational and occupational aspirations or, according to Bourdieu (1973, 1986), one’s habitus – i.e., a conception of one’s place in the social structure and sense of what is possible to achieve in one’s lifetime. In this regard, those of first-generation status, who also tend to be of working-class and low-income origins, come to appreciate which adult statuses and ambitions are possible to reach as well as those that are more improbable (Mickelson 2003). It is likely for this reason that, along with gatekeeping by school personnel, such students generally have a higher likelihood of being filtered into vocational education during the high school years, have less cultural capital and receive lower educational returns than their more privileged counterparts (Roscigno and Ainsworth-Darnell 1999). But what of those who beat the odds and make it to college? Are their curricular pathways shaped uniquely and in ways consistent with more applied applications to the labor market? Although little contemporary work has addressed this question, classic literature on family background and expectations is informative.

Kohn’s (1969) seminal work, Class and Conformity, illustrates how employment experiences across the social class hierarchy help to structure socialization in the household, including what is expected from children. Parents’ occupations, according to Kohn, reinforce values of either self-direction (e.g., freedom from supervision, intellectual and moral flexibility) or conformity (e.g., adherence to supervision, intellectual and moral rigidity; see also Kohn 1963). In a somewhat parallel vein, Lamont (2000:4) points to ways in which working-class values of hard work, responsibility, and integrity function as an alternative to economic definitions of success and serve as “moral” boundary work in the maintenance of worth and dignity. Such orientations “provide a model of the kinds of behaviors they encourage in their children” and, especially when coupled with limited educational resources in the home, have consequences for children’s self-direction and valuation surrounding educational attainment (Parcel and Menaghan 1994a:110, 1994b).

These perspectives – perspectives emphasizing the transmission of values – have been applied to the sociological understanding of class differences in college enrollment and completion. Students who feel academically marginalized owing to background inequalities develop contextually protective identity strategies that allow for the maintenance of dignity but also help reify certain social inequalities (Bettie 2003; Hurst 2010). Growing bodies of work concur in this regard, pointing to a “utilitarian” view of education and a related acquisition of job-specific skills and credentials to secure employment rather than personal and intellectual development, a prevailing narrative of college-going in the United States (Goyette and Mullen 2006; Quadlin 2017; Weis 1990; Wilkins 2014).

It is important to note that such utilitarian views, while probably perceived as practical in the immediate sense, most likely come at a cost when considering overall and long-term mobility. Indeed, occupationally specific degrees tend to be associated with somewhat higher rewards upon initial labor market entry, yet have the lowest growth in occupational status, fewer prospects for promotion, and lower wage growth over time (Roksa and Levey 2010; Spilerman and Lunde 1991; Thomas and Zhang 2005). In contrast, those with academically-oriented degrees – i.e., degrees that prioritize general skills – typically enter the labor market with lower occupational status but eventually experience higher returns and more opportunities for mobility (Roksa and Levey 2010).¹

Addressing such variations in college major selection with specific attention to first-generation students is especially timely given college demographic shifts over time and the rising costs and efficacy of a college degree (Abel and Deitz 2014). The inequalities that we are discussing also relevant to more general sociological concerns regarding mobility, inequalities within higher education and the intergenerational patterning of stratification. Our most general expectation, building on the literatures above, is that:

First-generation college students will be more highly represented in applied college majors compared to their continuing-generation peers—majors that have clearer and more explicit links to specific jobs in the labor market. Such group differences will hold even when accounting for other key background controls (e.g., prior achievement, socioeconomic status, etc).

The Potential Gendered Character of First-Generation Pathways

In addition to vertical stratification by generational status, it is essential to acknowledge the possibility of persistent horizontal stratification by gender (e.g., Alon and Gelbgiser 2011; Mann and DiPrete 2013) and how it may be particularly consequential for first-generation students. Recent analyses interrogating gaps in college majors as they relate to women’s college advantage, men’s lower degree attainment, and differential returns to education by gender are especially informative in these regards (e.g., Bobbitt-Zeher 2007; Mann and DiPrete 2013; McFarland 2017).

Much prior work on women’s college going and degree attainment focuses on lack of integration into science, technology, engineering, and mathematics (STEM) majors and, to a lesser extent, the limited integration of men into female-dominated fields of study (DiPrete and Buchmann 2013; England and Su 2006; Mann and DiPrete 2013). Underlying gender segregation in college majors, it is suggested, are cultural beliefs rooted in gender essentialism (i.e., beliefs that men and women are inherently different and suitable for different kinds of work), differential family-work conflicts, and differences in job preferences (Konrad et al. 2000; DiPrete and Buchmann 2013; Duxbury and Higgins 1991; Eccles 2007; England and Su 2006; Quadlin 2020). There is also emerging evidence that university programming itself tends to facilitate women’s investment in “gendered dispositions” and, thus, their movement toward female-dominated majors (Armstrong and Hamilton 2013; Charles and Bradley 2009; Hamilton 2014).

The infrastructure and peer cultures within colleges and universities likewise seem to imbue women with gendered cultural models for achieving economic security. In this regard, Hamilton (2014), in her longitudinal ethnographic study of women in higher education, finds that moderately selective, flagship universities encourage women – especially white, high-SES women – toward “gender complementarity.” Under this gender-traditional model of economic security, financially dependent women are matched with career-oriented men (Hamilton 2014). This logic privileges the enactment of “affluent, white, heterosexual femininity” and, in turn, encourages advantaged women to pursue majors that facilitate their investment in gendered dispositions (Hamilton 2014:5).

While men from more advantaged backgrounds tend to choose business and life/health majors, advantaged female students are more inclined toward the social sciences, humanities, and “easy majors” – i.e., business (from less selective institutions), communications, tourism, recreation studies, education, human development, fitness, and fashion (Hamilton 2014). Easy majors are characterized by “the ease of obtaining a high GPA, little evidence of general skills improvement during college; and a heavy focus on appearance, personality, and charm” (Armstrong and Hamilton 2013:70; Arum and Roksa 2011). The significant social and cultural capital held by advantaged female students facilitates entry into the labor market regardless of their academic performance or credentials (Armstrong and Hamilton 2013).

Unlike middle and upper-class female students, less privileged women lack the time and resources to invest in femininity rewarded in postsecondary institutions (Armstrong and Hamilton 2013; Hamilton 2014). Instead, disadvantaged but upwardly mobile women may rely on gender logics characterized by the pursuit of pragmatic careers and emphasize women’s need to be self-reliant (England 2010). And although the educational decisions of less privileged college women may be practical, they also tend to be highly gendered. Despite achieving educational mobility, many remain in female-typed work and pursue applied, pink-collar majors that ensure immediate returns in the labor market (Armstrong and Hamilton 2013; England 2010). Applied, female-dominated careers may be a step-up from those of their parents, but they pay less than occupations with a lower percentage of women even after controlling for differences in education requirements, skill levels, and working conditions (Cohen and Huffman 2003a, 2003b; England, Budig, and Folbre 2002).

Recent national estimates indicate that men are over-represented in fields such as engineering, computer science, and physical sciences. Women, in contrast, are over-represented in fields such as nursing, elementary education, biological sciences, veterinary medicine, and most arts and humanities fields (England and Su 2006; Quadlin 2018). Significant progress toward desegregation in college majors since 1971 can be attributed to women’s integration in male-dominated fields (England and Su 2006). Women within male-dominated STEM fields, however, have yet to reap the same returns as men when they enter the labor market, and face persistent stereotypes surrounding competency, job commitment, and likability (Musto 2019; Quadlin 2018). Men, for their part, have not moved into numerically female-dominated fields of study in in any significant degree owing to devaluation both socially and in the labor market (England and Su 2006; Yavorsky and Dill 2020).

Less clear are the implications of the gender-specific patterns, noted above, for first-generation students. It is plausible that intersections between gender and first-generation status exacerbate inequalities in the pathways (i.e., majors) that students pursue. There is evidence, for instance, that students from lower-SES families tend to choose fields that have higher pay upon graduation and clear career trajectories, such as technical, life/health science, and business majors (Kim, Tamborini, and Sakamoto 2015; Ma 2009). And, relative to gender specifically, Ma (2009) finds that lower-SES women tend to favor applied college majors with relatively high economic returns in the early career, such as nursing majors and education. These pathways are notable because, as was suggested previously, such applied majors tend to have relatively low wage growth over time (Roksa and Levey 2010).

Gaps in STEM major enrollment and completion may also be particularly pertinent for understanding how first-generation and gender status intersect (Chen and Soldner 2013; Dika and D’Amico 2016). We know full well from growing bodies of work that women are especially underrepresented in STEM majors, and students who are more reliant on student loans and/or who have little in the way of a parental financial safety net tend to cluster in majors that are both outside of STEM and more “applied” in character (Quadlin 2017). This, along with prior literatures noted above, suggests the possibility of confounding effects of first-generation status and gender, and more broadly the expectation that:

First-generation men and women, along with selecting majors with a more applied focus, will also be more likely than their continuing-generation peers to do so in normatively gendered ways.

Data and Measurement

Our data are drawn from the Education Longitudinal Study of 2002 (ELS-2002), a nationally representative study conducted by the National Center for Education Statistics (NCES). The ELS focuses on student trajectories in high school and patterns of college access and persistence beyond high school (U.S. Department of Education 2015). ELS also offers rich information on students’ class backgrounds, including detail surrounding parents’ educational attainment, income, and occupation, as well as pre-college information such as academic preparation during the secondary school years. We draw explicitly from the first and second waves for this information. The baseline tenth-grade cohort was followed at two-year intervals as students completed high school and some entered post-secondary education. We restrict the sample to the 6, 240 respondents who entered a post-secondary institution² in 2006 and declared a major,³ and use multiple imputation with chained equations to handle missing data in our analyses.⁴ Our main models reported in the text make use of the entire 2- and 4-year college-going sample.⁵

Applied versus Academic Major, and Gendered Composition of Major

We focus analytically on two distinct classifications, applied majors and academic majors. This distinction, utilized in recent research on unique breakdowns among college majors (see Charles and Bradley 2002; Quadlin 2017; Roksa and Levey 2010), captures horizontal dynamics within higher education based on the variation in and the consequences of the type of major students choose and graduate with. Detailed classifications based on the respondents’ self-reported field of study (see Table 1) were collapsed into one of the two categories based on the Classification of Instructional Programs (CIP) developed by the National Center for Education Statistics (NCES 2012).⁶

Table 1.

Coding of applied and academic majors, N = 6,240.

Major (as listed in ELS dataset)	Proportion of sample
Applied majors
Agriculture/natural resources/related fields	.01
Architecture and related services	.01
Business/management/marketing/related fields	.18
Communication/journalism/comm tech	.05
Computer/informatic sciences/support tech	.03
Education	.09
Engineering technologies/technicians	.06
Family/consumer sciences, human sciences	.01
Health professions/clinical sciences	.15
Legal professions and studies	.01
Parks/recreation/leisure/fitness studies	.01
Personal and culinary services	.01
Public administration/social services	.01
Security and protective services	.03
Total applied majors	.67
Academic majors
Arts - visual and performing	.06
Biological and biomedical sciences	.07
English language and literature/letters	.02
Foreign languages/literature/linguistic	.01
Liberal arts/science, general studies/humanities	.02
Mathematics and statistics	.01
Philosophy, religion and theology	.01
Physical sciences	.02
Psychology	.05
Social sciences (except psychology)	.07
Total academic majors	.33

Sources: Education Longitudinal Study of 2002 (ELS-02); National Center for Education Statistics.

Proportions may not add to 1 due to rounding.

Consistent with Quadlin (2017) and Roksa and Levey (2010), applied majors are fields that entail specific occupational training and clear corollaries in the labor market (for example, education majors train to become teachers, nursing majors train to become nurses, etc.). Applied fields, listed in the top section of Table 1, include architecture and related services; business management, marketing, and related fields; communication, journalism, and communication technologies; computer, informatics sciences, and support technologies; education; engineering technologies and technicians; architecture; health professions and clinical sciences; mathematics and statistics; physical sciences; public administration and social services; and security and protective services.

Academic fields have no obvious corollaries in the labor market and entail less occupationally specific training. These majors, listed in the bottom section of Table 1, and include: agricultural and natural resources; visual and performing arts; biology and biomedical sciences; English language and literature; family sciences, consumer sciences, and human sciences; foreign languages, literature, and linguistics; legal professions and legal studies; liberal arts and sciences, general studies, and humanities; parks and recreation, leisure, and fitness studies; personal and culinary services; philosophy, religion, and theology; psychology; and social sciences (except psychology). Our categorizations are generally in line with recent research assessing linkages between education and the labor market (Bol et al. 2019).

Our second analytic focus centers on the gendered character of college majors, which we capture with gender (female) composition of field of study. Such gendered segregation in fields of study over the past 20 years is reflected in national data on choice of major and awarded bachelor’s degrees (Alon and Gelbgiser 2011; Buchmann and DiPrete 2006; Eccles 2007; Mann and DiPrete 2013). In this regard, we use a measure of the national-level sex composition of bachelor’s-level fields of study at the time students in this dataset would have been choosing majors (i.e., for most, this would have been around 2006; see NCES 2008: Table 285).⁷ We report means and standard deviations for this outcome, our applied versus academic distinction, as well as other background attributes and controls in Table 2.

Table 2.

Descriptive statistics, N = 6,240.

	Mean or proportion (SD)
First-generation college student	.44
Female	.58
Father’s occupation
Low-wage service	.12
Manual	.28
Skilled or semi-skilled	.15
High wage	.43
Outside labor market	.02
Mother’s occupation
Low-wage service	.33
Manual	.07
Skilled or semi-skilled	.08
High wage	.45
Outside labor market	.06
Race/ethnicity
White	.63
Black	.11
Latinx	.10
Asian	.11
Other race	.04
Family income (log; categories)	10.97 (.90)
High School GPA (categories)	3.11 (.60)
Sector of undergraduate institution
Public	.71
Private, not for profit	.26
Private, for profit	.03
Level of undergraduate institution
Two-year or community college	.19
Four-year or bachelor’s-level	.81
Year of birth
1983	<.01
1984	.01
1985	.34
1986	.64
1987	.01
Lived with both parents in high school	.69

Source: Education Longitudinal Study of 2002 (ELS-02).

First-Generation Status and Gender

Our two central explanatory variables are first-generation status and gender. First-generation status is derived from parents’ educational attainment as reported in the student and parental survey components during the high school years. The most conventional definition of a first-generation college student, used by most colleges and universities across the U.S., is a person for whom neither parent holds a bachelor’s degree or higher (McConnell 2000; Petty 2014; Próspero and Vohra-Gupta 2007). Continuing-generation college students are those for whom at least one parent has a bachelor’s degree or higher. Gender is derived from self-reports. In our main analyses, we disaggregate models by gender to explicitly capture any potential gender-specific variations associated with first- versus continuing-generation status.

Other Background Attributes and Controls

We consider race/ethnicity in our modeling with a series of binary indicators for four nonwhite racial/ethnic categories (i.e., Black, Latinx, Asian, and “Other Race”) with White as the referent.⁸ Family income is recoded as a continuous variable, taking the midpoint of each income interval in the original income scale measured in the base year of the survey. We use the natural log transformation of this indicator to minimize skewness.

To capture the extent to which parental occupation may help explain any first-generation and/or gender effects observed, and in concert with literature suggesting that parents’ work experiences and values shape children’s educational and work pathways (Kohn 1963, 1969), we include a set of binary indicators regarding both mother’s and father’s occupations. Parental occupation is derived from the parent survey, with responses initially coded into categories of: No job for pay; Clerical; Craftsperson; Farm/farm manager; Homemaker; Laborer; Manager/administrator; Operative; Professional A (doctor, lawyer, dentist, etc.); Professional B (actor, writer, etc.); Proprietor /owner; Protective services; Sales; School teacher; Service; and Technical.⁹ We standardize these measures into a broader set of classifications that represent important social class distinctions. These classifications are informed by Goldthorpe John’s (1980) occupation-based social classification, which assumes that individuals in similar occupations share similar economic conditions, work conditions, lifestyle, and life chances (see also Connelly, Gayle, and Lambert 2016; Evans 1992).

The resulting occupational categories include: (1) low-wage service, (2) manual, (3) skilled or semi-skilled, (4) high-wage, and (5) outside the labor market. Low-wage service occupations are comprised of clerical, service, and sales jobs; they require no specialized skills or training and are associated with low pay. Manual occupations, or blue-collar jobs, include manual laborers, craftspeople, farmers/farm managers, laborers, and operatives. Skilled or semi-skilled occupations are comprised of those in protective services, proprietors/owners, and technical jobs. These are jobs that are associated with higher levels of expertise, training and experience, and higher wages, compared to those in manual occupations or low-wage occupations. High-wage occupations consist of careers that require job-specific training or credentialing, including all types of professionals and teachers. Finally, those outside the labor market include parents who are homemakers and those who are not employed.

Along with key explanatory indicators, we include several controls that highlight individual attributes prior to college entry. For instance, we include high school GPA as a proxy for college readiness and/or academic ability. High school GPA is measured on a continuous 0.0– 4.0 scale. Additionally, we include a control for the sector of post-secondary institution attended – i.e., whether it is public, private for-profit, or private not-for-profit. Lastly, we include a family composition variable. The ELS-2002 includes a multi-categorical variable to represent students’ family composition at the first survey wave, and we use this variable to measure whether students lived with both of their parents.

Analytic Strategy and Results

Our analyses proceed in two steps consistent with our earlier expectations. First, and using logistic regression, we analyze the patterning of college majors among first- and continuing-generation students, focusing on whether students chose majors in applied versus academic fields. These models are disaggregated by gender. We include baseline controls in the initial models, including parental occupational indicators and a measure of family income. Doing so helps distinguish unique first-generation effects relative to more traditional indicators of social class. The second set of models are even more rigorous and include indicators of family structure, high school achievement, and the type of college the student attended (i.e., public versus private nonprofit versus private for-profit).

The second portion of our analysis, which makes use of OLS regression, tests the extent to which first-generation college student status is associated with the gender composition of students’ majors. Like the prior analyses, we disaggregate these models by gender. General sociodemographic controls are included in the first model, like the previous analyses, and family structure, high school achievement and type of college are introduced in the second model.

As a robustness check, we replicate our analyses of college major selection and its potentially gendered character on 2- and 4-year college-going samples and report these results in our appendices. Our rationale for doing so is two-fold. First, two-year colleges likely place a greater emphasis on applied majors with more proximate connections to employment. Secondly, first-generation college students are much more highly represented at 2- versus 4-year institutions. Despite such differences however, our findings in these regards (noted in our results discussion) are relatively parallel to the main analyses.

Generational Status and Applied versus Academic Majors

Table 3 reports the effects of first-generation status and sociodemographic characteristics on enrollment in an applied college major, and among women and men. Note that the omitted category for the outcome is an “academic” major. As such, these coefficients effectively indicate students’ chances of selecting an applied major compared to an academic major. As the reader will see in Model 1, and consistent with our first hypothesis, both first-generation women and men are more likely to be concentrated in applied fields compared to academic fields relative to their continuing-generation counterparts (both p < .001). These patterns persist in Model 2 for both men and women, thus demonstrating the centrality of first-generation student status (net of SES) for major choice (both p < .001).

Table 3.

Effects of first-generation status and sociodemographic characteristics on enrollment in applied major - by gender.

	Women		Men
	(1)	(2)	(1)	(2)
First-generation college student	0.37***	0.27**	0.34**	0.27*
	(0.09)	(0.09)	(0.11)	(0.11)
Black	0.02	−0.16	0.53**	0.41*
	(0.14)	(0.14)	(0.17)	(0.18)
Latinx	−0.40**	−0.53***	0.00	−0.13
	(0.13)	(0.13)	(0.16)	(0.16)
Asian	−0.63***	−0.60***	−0.19	−0.17
	(0.12)	(0.12)	(0.14)	(0.14)
Other race	−0.56**	−0.65***	−0.14	−0.15
	(0.18)	(0.19)	(0.20)	(0.20)
Father’s occupation
Manual	0.22	0.21	0.08	0.03
	(0.13)	(0.13)	(0.16)	(0.17)
Skilled or semi-skilled	−0.06	−0.01	−0.10	−0.12
	(0.14)	(0.14)	(0.17)	(0.17)
High wage	−0.11	−0.06	−0.12	−0.12
	(0.12)	(0.13)	(0.15)	(0.15)
Outside labor market	0.62*	0.52	0.03	−0.16
	(0.31)	(0.31)	(0.32)	(0.33)
Mother’s occupation
Manual	0.27	0.21	0.13	0.07
	(0.16)	(0.17)	(0.21)	(0.21)
Skilled or semi-skilled	−0.07	−0.06	−0.16	−0.16
	(0.15)	(0.15)	(0.17)	(0.17)
High wage	−0.08	−0.05	−0.24*	−0.24*
	(0.09)	(0.09)	(0.10)	(0.10)
Outside labor market	−0.26	−0.27	0.06	−0.02
	(0.16)	(0.16)	(0.21)	(0.21)
Family income	−0.09	−0.07	−0.00	0.01
	(0.05)	(0.05)	(0.06)	(0.07)
Lived with both parents		0.13		0.09
		(0.09)		(0.11)
High school GPA		−0.45***		−0.26**
		(0.08)		(0.08)
Attended private, not for profit		−0.35***		−0.39***
		(0.09)		(0.11)
Attended private, for profit		0.47		0.15
		(0.30)		(0.32)
Attended four-year college		−0.42***		−0.32*
		(0.11)		(0.14)
Age		−0.19*		−0.06
		(0.08)		(0.08)
n	3,612		2,628

Source: Education Longitudinal Study (ELS-02). Note: Logistic regressions; logs odds reported.

Omitted categories are not first-generation college student; White; low-wage service occupation; did not live with both parents; attended public institution; attended two-year college.

Although a greater proportion of students in our sample are enrolled in applied majors, rather than academic majors (see Table 1), first-generation college students are especially likely to major in applied fields. Importantly, such group differences are observed net of family income and parents’ occupational background. They also persist net of key controls included in Model 2 (e.g., prior achievement and college sector).

The overall first-generation differences reported in Table 3 are likewise observed, for the most part and particularly among those attending 4-year institutions, when we disaggregate the sample by college type (Table A1). Directionality is similar for those attending 2-year institutions, although effects do not reach significance owing to sample size reductions. It is also the case, according to more detailed analyses undertaken, that there is little variation overall in who attends two-year colleges. About half of two-year college attenders in our sample are first-generation and majored in an applied subject. There is simply much more variation in both background inequalities and majoring in the context of 4-year institutions.

The results thus far reported suggest that first-generation college students tend to choose majors that they perceive as benefiting them at the point of labor market entry, in the sense that many of these majors are perceived as “practical.” Recall, however, that such majors tend to be associated with low occupational status growth over time (Roksa and Levey 2010). Thus, although applied majors may be beneficial in the early career (i.e., when many students will be paying back loans), they are generally less beneficial as the employment life course unfolds.

Notably, we find no significant effects in Model 1 of Table 3 for parental occupation on women’s choice of an applied major. For men on the other hand, we find moderate effects of mother’s occupation, such that men whose mothers work in high-wage occupations are less likely to select an applied major versus an academic major (p < .05 in both Model 1 and Model 2 for men). These results mirror those of Leppel, Williams, and Waldauer (2001), to some degree, who find that male students whose mothers are in professional or executive occupations are more likely to choose education and humanities and social sciences majors and are less likely to choose physical science or engineering majors. Supplementary analyses suggest that this effect does not differ for first- versus continuing-generation men.

Among other status attributes and controls highlighted in Table 3, we find important effects as well as some key variations across gender groups. Prior high school achievement (i.e., high school GPA) and private (and arguably more selective) college enrollment have generally strong negative effects on the choice of an applied college major (both p < .001) – a pattern that suggests the importance of college preparedness and variations in curricular emphases. Such effects are consistent across gender groups, while those regarding race/ethnicity point to important gender and race/ethnic intersectional pathways. Specifically, Latinx women, Asian women, and women in other racial groups are less likely than their white counterparts to choose an applied major. We do not find such effects for men, but instead find that Black men are somewhat more likely than their white counterparts to choose an applied major, even when controlling for prior achievement.¹⁰

Generational Status and Gendered Pathways

We also consider first-generation and gender gaps in college majors relative to the gendered composition of majors (i.e., the percentage of students in each major who are female) and report our findings in this regard in Table 4. The potentially gendered character of majors is important owing to the general devaluation of fields of study that are seen as feminized (England and Su 2006) as well as the tangible inequalities in job returns for fields that are gendered and/or treated under the more general umbrella of “care work” (Cohen and Huffman 2003a, 2003b; Dwyer 2013).

Table 4.

Effects of first-generation status and sociodemographic characteristics on gender composition of major (% female) - by gender.

	Women		Men
	(1)	(2)	(1)	(2)
First-generation college student	1.48*	1.04	0.87	0.44
	(0.61)	(0.61)	(0.79)	(0.79)
Black	−0.80	−1.80*	−1.51	−3.23**
	(0.83)	(0.87)	(1.09)	(1.14)
Latinx	−1.53	−2.06*	1.14	−0.03
	(0.84)	(0.84)	(1.17)	(1.17)
Asian	−3.90***	−3.73***	−1.74	−1.40
	(0.85)	(0.86)	(1.06)	(1.06)
Other race	−1.44	−1.83	1.27	1.22
	(1.27)	(1.27)	(1.52)	(1.51)
Father’s occupation
Manual	0.55	0.43	−0.80	−0.96
	(0.86)	(0.85)	(1.22)	(1.21)
Skilled or semi-skilled	−1.64	−1.49	−1.68	−1.71
	(0.96)	(0.95)	(1.26)	(1.25)
High wage	−1.64*	−1.46	−1.04	−0.90
	(0.83)	(0.82)	(1.11)	(1.10)
Outside labor market	1.46	0.96	3.15	1.80
	(1.94)	(1.94)	(2.32)	(2.33)
Mother’s occupation
Manual	0.42	0.18	−1.63	−1.86
	(0.99)	(0.99)	(1.42)	(1.41)
Skilled or semi-skilled	0.38	0.42	0.80	0.74
	(0.96)	(0.96)	(1.23)	(1.22)
High wage	0.30	0.44	0.29	0.30
	(0.59)	(0.59)	(0.79)	(0.79)
Outside labor market	−0.15	−0.08	−0.28	−0.72
	(1.07)	(1.06)	(1.54)	(1.54)
Family income	−0.42	−0.34	−0.99*	−0.85
	(0.30)	(0.31)	(0.46)	(0.48)
Lived with both parents		0.24		−0.08
		(0.56)		(0.79)
High school GPA		−2.11***		−3.17***
		(0.48)		(0.60)
Private, not for profit		−0.87		−0.31
		(0.59)		(0.76)
Private, for profit		0.47		−3.17
		(1.55)		(2.07)
Attended four-year college		−1.69*		−2.18*
		(0.71)		(1.02)
Age		−0.20		−0.98
		(0.49)		(0.59)
n	3,612		2,628

Source: Education Longitudinal Study (ELS-02). Note: OLS regressions; coefficients reported.

Like the prior analyses, these include the same sets of controls. Our results, however, suggest very little in the way of first-generation gender effects. Although first-generation women are generally more likely than either continuing-generation women or men to choose majors that are more heavily female in concentration, such differences become statistically non-significant in model 2 once prior achievement is accounted for. This seems to suggest that any first-generation female filtering into traditionally female majors (model 1) is mostly due to possible curricular variations and achievement prior to college entry. First-generation effects are likewise not observed in our modeling for men. Re-analyses and disaggregation of effects for 2- and 4-year college attendees, reported in Table B1, point to largely parallel patterns overall, and little in the way of an interplay of first-generation and gender status when it comes to gendered majors.

We also briefly consider the role of parental occupation in these analyses. In this regard, we find a small effect of father’s occupation on the gender composition of majors for women students. Having a father in a high-wage occupation (versus a low-wage occupation) is associated with major choices that are less female-dominated for women (i.e., more male-dominated; p < .05). This effect dissipates when we incorporate prior achievement in Model 2. This pattern is notable, especially recalling how mother’s occupation tended to be associated with men’s major choices in Table 4. Taken together, these results suggest some moderate significance of parental occupation in major choice generally, and potential impacts on children’s self-direction and educational attainment (Kohn 1963, 1969; Lamont 2000; Parcel and Menaghan 1994a, 1994b).

Discussion of Applied/Academic and Gendered Dimensions of Major Choice

We conclude the analyses by providing a synthesis of our results and by tying together the applied versus academic and gendered dimensions of major choice that we have considered in this article. We refer the reader to Table 5, which shows the most and least prevalent majors by generational status and gender.

Table 5.

Most and least prevalent majors - by generational status and gender.

Rank	First generation Women		Continuing generation Women		First generation Men		Continuing generation Men
Most common majors, by group
1 (Top)	Health professions	.27	Health professions	.17	Business	.23	Business	.23
2	Business	.17	Business	.14	Engineering	.11	Engineering	.14
3	Education	.12	Education	.11	Health professions	.08	Social sciences	.11
4	Psychology	.06	Arts	.08	Security/protective	.08	Biology	.08
5	Biology	.06	Social sciences	.08	Computer science	.07	Arts	.06
Least common majors, by group
1	Agriculture	.01	Mathematics	.01	English	.01	Agriculture	.01
2	Architecture	.01	Parks/recreation	.01	Foreign languages	<.01	Legal professions	.01
3	Foreign languages	.01	Philosophy	.01	Family sciences	<.01	Personal/culinary	<.01
4	Mathematics	<.01	Agriculture	.01	Legal professions	<.01	Public admin	<.01
5 (Bottom)	Philosophy	<.01	Legal professions	.01	Public admin	<.01	Family sciences	<.01

As noted in our analyses, first-generation women tend to select majors that are applied but not overwhelmingly or explicitly female dominated. This is especially clear in the breakdowns we report in Table 5. First-generation women are highly concentrated in health professions, business, education, psychology, and biology – majors that are largely applied but, if anything, have only a slight gendered character to them. Continuing-generation women share the same top three majors as their first-generation counterparts but are also concentrated in the arts and the social sciences. These majors are more academic than applied in character. Notable as well, first-generation women are significantly underrepresented in mathematics and philosophy, both of which are highly academic and lean somewhat toward male-dominated. This pattern generally aligns with recent work centering on gender gaps in the most male-dominated fields (Chen and Soldner 2013; Dika and D’Amico 2016; Quadlin 2017).

Alignment and divergence are likewise observed for first- versus continuing-generation men. Men generally share business and engineering as their top two most prevalent majors, regardless of their generational status. However, the remaining top three most prevalent majors differ. First-generation men are more highly represented in applied majors, including health professions, security and protective services, and computer science. Continuing-generation men, in contrast, are more prominently enrolled in academic majors, including the social sciences, biology, and the arts. This pattern comports with our finding that first-generation men are more likely to select applied majors than their continuing-generation peers – a choice that is likely to pay off in the early career but that may lead to diminishing returns in the long run as individuals accrue more job experience and credentials.

Conclusions

Although a college degree remains a key pathway to social mobility, fields of study yield differential returns (Hout 2012; Roksa and Levey 2010). Even when they beat the odds and attain a degree, first-generation college students experience disadvantages surrounding the type of education they receive (Charles and Bradley 2002; Roksa and Levey 2010), including underrepresentation in fields of study that lead to enhanced earnings over the course of their careers (Cataldi, Bennett, and Chen 2018; Van Noy and Ruder 2017). For these reasons, we analyzed first-generation students’ choice in college major; how their major choices varied from continuing-generation students; and whether their major selection varied in gendered ways.

We hypothesized that first-generation college students would be more likely than their continuing-generation counterparts to be represented in applied college majors – majors that have more explicit links to specific jobs in the labor market. This expectation was supported rather clearly in our first set of analyses. First-generation college students indeed tend to choose majors that are considered more “applied” rather than “academic” in nature. This pattern exists for both first-generation men and women, persists net of SES and prior achievement and is especially pronounced at four-year colleges and universities – universities where there is both more variation in family background inequalities and a wider plethora of college major choices.

Applied majors are undoubtedly perceived as “practical,” in the sense that they have clear corollaries in the labor market. This may be particularly appealing to first-generation college students who often lack family resources, rely on self-directed research and planning for guidance, and typically desire some assurance that their degree will pay off once they graduate. Such assurance may also be appealing to their parents, who may have reservations regarding of the value, costs, and necessity of a college degree (Armstrong and Hamilton 2013; Dumais and Nichols 2016; Hamilton 2016; Roksa and Silver 2019).

Although applied majors may make it easier for students to get their “foot in the door” relative to labor market entry and certain types of jobs, prior research suggests that such jobs are also associated with low occupational status growth over time (Roksa and Levey 2010). Initial labor market outcomes and variation in occupational trajectories over time, of course, have implications for patterns of inequality and life chances over the life course. They may reliably and more quickly lead to jobs, which can serve as a motivating force in college completion,¹¹ but they likely do not lead to the highest long-term payoffs to a college degree.

We further hypothesized that first-generation men and women would select and pursue majors that are more deeply gendered than their continuing-generation counterparts. Our findings in this regard were a bit less clear. Although first-generation women choose more heavily female-dominated majors than continuing-generation women, such effects were explained away once our modeling accounted for prior achievement in high school. This hardly means, of course, that gendered pathways in majoring do not impact first-generation women. They surely do, but the effect seems to be explained more so by prior high school achievement than decision-making while in college. Moreover, to the extent there is a gendering of major choices during the college years, it appears to hold for all women, not just first-generation women. Such gendered majors tend to lead to gendered employment, characterized by segregation, devaluation, and lower wages. First-generation women, like all college going women, are vulnerable in this regard but (like their first-generational male peers) may also experience eventual labor market inequalities owing to the applied majors they disproportionately come out of college with.

First-generation college students typically do not receive the same support or possess the same level of social and cultural capital as middle-class students whose parents have a college degree (Hurst 2010; Lee 2016; Pascarella et al. 2004; Roscigno and Ainsworth-Darnell 1999). As such, they are likely to make what they view as practical decisions that lead to a job rather than risk choosing a major with unclear ties to the labor market. Class background and gender may also impact the career trajectories that first-generation men and women see as available to them and, consequently, contribute to the concentration of women in female-dominated fields of study (DiPrete and Buchmann 2006; Duxbury and Higgins 1991; Eccles 2007; England and Su 2006). Decision making in these regards can be conceived of as strategic and beneficial to a population with limited resources and substantial background challenges, or as a form of inequality structured by constraints – constraints that restrict the real or perceived options that individuals see. Regardless, educational pathways have consequences – consequences for labor market returns, stratification, and mobility prospects throughout the life course.

Considering that first-generation college students stand to become more numerous in the coming years, attending to this population is imperative (Grawe 2018). Research like ours can inform policy around enrollment gaps and patterns of attrition by gender and among first-generation college students while simultaneously speaking to broader sociological questions and issues of inequality, mobility, and status attainment trajectories. We hope future work will proceed in this fashion, extend upon on our findings and address current limitations. One key weakness, for instance, surrounds the fact that major choice and shifts across time tend not to be captured in large datasets like the ELS, thus limiting analyses to students’ initial choice in major. Transitions across time, we suspect, may be more common among first-generation college students given that they tend to experience more barriers and disruptions in their educational careers (Benson and Lee 2020). Future data collection efforts in these regards – i.e., longitudinal, quantitative as well as qualitative – would be useful and provide further important insights into first-generation college student experiences and general and gendered trajectories.

Table A1. Effects of first-generation status and sociodemographic characteristics on enrollment in applied major - by gender and type of institution attended.

	Women		Men
	Two-year College	Four-year College	Two-year College	Four-year College
First-generation college student	0.32	0.27**	0.12	0.29*
	(0.24)	(0.10)	(0.31)	(0.12)
Black	−0.32	−0.11	0.11	0.45*
	(0.37)	(0.16)	(0.46)	(0.19)
Latinx	−0.48	−0.55***	−0.08	−0.14
	(0.28)	(0.15)	(0.41)	(0.19)
Asian	−0.44	−0.63***	−0.72	−0.10
	(0.35)	(0.14)	(0.41)	(0.15)
Other race	−0.98*	−0.60**	0.21	−0.19
	(0.48)	(0.20)	(0.71)	(0.22)
Father’s occupation
Manual	0.49	0.13	0.47	−0.03
	(0.32)	(0.14)	(0.43)	(0.18)
Skilled or semi-skilled	0.13	−0.04	0.08	−0.15
	(0.40)	(0.15)	(0.47)	(0.19)
High wage	−0.14	−0.06	0.39	−0.18
	(0.34)	(0.13)	(0.43)	(0.16)
Outside labor market	−0.51	0.90*	0.35	−0.20
	(0.59)	(0.39)	(0.68)	(0.41)
Mother’s occupation
Manual	0.00	0.28	0.31	0.03
	(0.33)	(0.20)	(0.48)	(0.24)
Skilled or semi-skilled	0.11	−0.10	−0.05	−0.18
	(0.39)	(0.16)	(0.51)	(0.19)
High wage	−0.01	−0.05	−0.30	−0.23
	(0.26)	(0.10)	(0.29)	(0.12)
Outside labor market	−0.30	−0.27	−0.40	0.06
	(0.41)	(0.18)	(0.50)	(0.24)
Family income	−0.02	−0.07	0.14	−0.02
	(0.12)	(0.05)	(0.17)	(0.07)
Lived with both parents	−0.15	0.19	−0.18	0.13
	(0.22)	(0.10)	(0.31)	(0.11)
High school GPA	−0.31	−0.48***	−0.01	−0.29**
	(0.19)	(0.09)	(0.21)	(0.09)
Attended private, not for profit	−0.11	−0.37***	−0.78*	−0.36**
	(0.36)	(0.09)	(0.38)	(0.11)
Attended private, for profit	1.00	0.34	0.13	0.21
	(0.71)	(0.37)	(0.61)	(0.38)
Age	−0.40*	−0.16	−0.30	−0.03
	(0.20)	(0.08)	(0.25)	(0.09)
n	949	2,663	698	1,930

Source: Education Longitudinal Study (ELS-02). Note: Logistic regressions; logs odds reported.

Omitted categories are not first-generation college student; White; low-wage service occupation; did not live with both parents; attended public institution.

Table B1. Effects of first-generation status and sociodemographic characteristics on gender composition of major (% female) - by gender and type of institution attended.

	Women		Men
	Two-year College	Four-year College	Two-year College	Four-year College
First-generation college student	2.35	0.69	−0.20	0.65
	(1.48)	(0.68)	(1.91)	(0.91)
Black	0.17	−2.48*	−4.06	−3.12*
	(1.93)	(0.99)	(2.96)	(1.31)
Latinx	−1.44	−2.20*	−2.02	0.57
	(1.66)	(0.99)	(2.69)	(1.37)
Asian	0.77	−4.57***	−2.20	−1.45
	(2.32)	(0.96)	(3.20)	(1.15)
Other race	−4.45	−1.31	0.67	1.33
	(3.24)	(1.38)	(4.12)	(1.64)
Father’s occupation
Manual	0.20	0.53	−1.11	−0.96
	(1.80)	(0.99)	(3.18)	(1.34)
Skilled or semi-skilled	−0.77	−1.58	−3.32	−1.36
	(2.31)	(1.07)	(3.66)	(1.40)
High wage	−1.29	−1.53	−1.87	−0.65
	(1.99)	(0.92)	(3.47)	(1.22)
Outside labor market	−2.74	2.20	1.68	1.46
	(4.16)	(2.31)	(4.56)	(2.91)
Mother’s occupation
Manual	0.15	0.16	−2.35	−1.72
	(1.88)	(1.22)	(3.11)	(1.69)
Skilled or semi-skilled	2.34	−0.10	−1.67	1.20
	(2.22)	(1.12)	(3.34)	(1.35)
High wage	0.51	0.45	1.03	0.15
	(1.55)	(0.68)	(2.10)	(0.87)
Outside labor market	−2.12	0.47	−1.46	−0.45
	(2.49)	(1.22)	(3.28)	(1.79)
Family income	−0.30	−0.36	−2.11	−0.57
	(0.74)	(0.35)	(1.21)	(0.53)
Lived with both parents	1.01	−0.07	−0.73	0.09
	(1.26)	(0.65)	(1.97)	(0.90)
High school GPA	−1.07	−2.42***	−2.33	−3.47***
	(1.01)	(0.56)	(1.38)	(0.67)
Attended private, not for profit	−0.48	−0.93	−0.15	−0.40
	(2.08)	(0.62)	(2.68)	(0.80)
Attended private, for profit	3.53	−1.01	1.27	−5.38*
	(2.59)	(2.00)	(3.87)	(2.41)
Age	0.16	−0.36	−1.15	−0.93
	(1.11)	(0.56)	(1.47)	(0.66)
n	949	2,663	698	1,930

Source: Education Longitudinal Study (ELS-02). Note: Logistic regressions; logs odds reported.

Omitted categories are not first-generation college student; White; low-wage service occupation; did not live with both parents; attended public institution.