Primary and Secondary Effects of Gender on Educational Attainment: Toward a Better Understanding of the Academic Underachievement of Boys

Abstract

Gender segregation in the labor market is closely related to gender-segregated educational choices. Male- and female-typed fields of study are typically associated with different educational requirements. Educational and occupational aspirations may therefore influence how boys and girls invest in education. The author investigates how gender-typed occupational aspirations and gender differences in educational expectations may help explain the gender gaps in academic performance and educational attainment in Norway. Educational attainment among 1,076 youth born in 1992 is examined using a longitudinal survey linked to register data. The findings show that educational expectations and male-typed occupational aspirations explain about one third of the gender gap in grade point average in 10th grade. Moreover, comparing boys and girls with similar grade point averages, expectations and aspirations account for 52 percent of the gender gap in college completion by 28 years of age. Notably, having male-typed occupational aspirations alone explain 37 percent to 43 percent of the gap, depending on whether educational expectations are accounted for.

Keywords

primary and secondary effects gender equality educational attainment academic performance Norway academic achievement

Across the developed world, girls are outperforming boys in most academic subjects and are more likely to enroll in, and graduate from, higher education. There is also pervasive horizontal gender segregation by field of study. Although much research has been devoted to understanding the underperformance of girls in mathematics and the underrepresentation of women in science, technology, engineering, and mathematics subjects, less scholarly attention has been paid to explaining the general academic underachievement of boys.

The main explanations for girls’ lower math achievement and underrepresentation in science, technology, engineering, and mathematics, are heavily influenced by social psychology. According to this literature, gender stereotypes negatively influence girls’ engagement with and interest in math and science, as well as girls’ beliefs about their own math competence (Casad, Hale, and Wachs 2015; Correll 2004; Master 2021; Passolunghi, Ferreira, and Tomasetto 2014). Other studies have argued that lack of interest in technical fields is more intrinsically related to gendered dispositions, where girls are more oriented toward relational and reflective subjects, whereas boys are more oriented toward technical subjects (e.g., Kuhn and Wolter 2022; Stoet and Geary 2022).

Some studies have tried to explain the gender gap in academic performance overall, or in subjects like reading or native language, which has typically been in favor of girls, also historically. An important contribution is that of DiPrete and Buchmann (2013), who provided a comprehensive overview of various explanations in the U.S. context. They underscore that there is no reason to believe that there are inherent ability differences across genders. Differences between boys and girls are typically small or in favor of boys on standardized tests, but at the same time grade differences in school subjects in favor of girls are substantial (Cornwell, Mustard, and Van Parys 2013; Graetz and Karimi 2022; Lekholm and Cliffordson 2008; Saygin 2020).

DiPrete and Buchmann (2013) presented three interrelated explanations for the gender gap in academic achievement: differences in social and behavioral skills, school attachment, and educational expectations. A number of studies have documented that girls score higher than boys on average on school readiness measures such as self-control and social skills (Brandlistuen et al. 2021; Fidjeland et al. 2023). These skills are associated with higher academic achievement. Moreover, on average, girls spend more time on homework and reading outside of school hours than boys and are also more likely to say that it is important to work hard in school (Borgonovi, Ferrara, and Maghnouj 2018; Gershenson and Holt 2015; Smith and Reimer 2024). Educational expectations have been found to be strongly associated with academic performance and educational attainment, and girls typically have higher educational expectations than boys (DiPrete and Buchmann 2013:110). However, an often overlooked reason for the association between college plans and academic achievement may be that school grades matter little for the labor market outcomes of the students who do not plan to go to college but rather plan to enter the labor market after high school (Rosenbaum 2001:9).

In this article, I ask how educational expectations and gender-typed occupational aspirations contribute to explaining gender gaps in academic performance and educational attainment in favor of girls. In line with Rosenbaum’s (2001) observation, I hypothesize that one of the reasons why boys have lower school grades and lower educational attainment than girls, is that the jobs they aspire to, on average, are less likely to require high marks or tertiary degrees. Gender-typical occupational aspirations are likely to spur gender differences in educational decisions, associated with different educational requirements in terms of both grade point average (GPA) and degree attainment.

To my knowledge, no studies have investigated the influence of gender-typed occupational aspirations on school grades and college completion. However, aspirations and plans have been found to significantly influence gender-segregated educational choices. In a study from the United States, Morgan, Gelbgiser, and Weeden (2013) found that occupational plans of high school seniors were significantly more important for explaining gender differences in college major choice than high school coursework and more general work-family preferences. A more recent study from Italy points in the same direction, with occupational aspirations and curricular track choice as central explanatory factors for gender gaps in field of study in higher education (Barone and Assirelli 2020). I extend these important contributions by investigating whether occupational plans as they are formulated early in a student’s educational career, can help explain both gender gaps in academic performance and gender gaps in propensities to get a college degree.

This research is grounded in classical social stratification theory, particularly the concepts of primary and secondary effects in class differentials (Boudon 1974; Breen and Goldthorpe 1997). Following Becker (2014) and Hadjar and Buchmann (2016), I apply this framework to gender, conceptualizing academic performance as a primary effect of gender, and educational choices as a secondary effect of gender, net of academic performance. I argue that cultural stereotypes about appropriate occupations, occupational opportunity structures, as well as costs and benefits of occupational choices vary across genders, influencing both the efforts invested in school (primary effect) and the motivation to complete a college degree (secondary effect).

More specifically, I use a unique dataset that combines longitudinal survey data with population register data from Norway to investigate how gender-typed occupational aspirations and gender gaps in educational expectations in lower secondary school contribute to the gender gaps in academic performance, measured as school grades, and later educational attainment.

Theoretical Framework

Primary and Secondary Effects of Stratification

Classic sociological theory on social stratification in education posits that social class background influences children’s educational attainment in two distinct ways, labeled primary and secondary effects (Boudon 1974). The primary effects of social stratification refer to the impact of social background on academic performance. Children from less privileged backgrounds are less likely to perform well in school. The secondary effects concern educational choices made by students with similar levels of academic performance but different social backgrounds. Even though on average, children from less privileged backgrounds have lower academic performance than children from more privileged backgrounds, there is also considerable overlap in academic performance in the two populations. The secondary effects of stratification imply that children from different social backgrounds with the same level of academic performance, are likely to make different educational choices (Jackson et al. 2007; Morgan 2012).

From a rational choice perspective, Boudon (1974) argued that the benefit is higher and the cost is lower of choosing a higher status option for the higher-status child. In Boudon’s example, the lower class child may expect social promotion from choosing a vocational course, whereas the upper class child will experience social demotion (p. 29). Moreover, although the upper class child is likely to experience a high-status choice as reinforcing family solidarity and as being in line with the choice of their friends, the lower class child may experience the opposite. It has been argued that this theoretical framework can be extended to gender stratification in education (Becker 2014; Hadjar and Buchmann 2016). Boys on average have lower school performance than girls, but there is considerable overlap between the two populations. When a boy and a girl have similar academic performance, the cost and benefit of choosing an academic track over a vocational track differ for boys and for girls, and so does the utility of a higher education degree.

Two key factors contribute to these gender differences in educational decision making. First, the labor market is highly gender segregated. Among vocational occupations, male-dominated occupations are typically of higher status and have higher income prospects than female-dominated ones. Strong gender stereotypes associated with these occupations imply high thresholds for gender-nontraditional choices. Second, there is a gendered asymmetry in educational requirements. Many of the largest female-typed occupations, such as nursing, early childhood education and social work require higher education degrees, whereas common male dominated occupations, such as production, construction and transportation do not. Consequently, a similar rational choice perspective can be applied to explain why girls are more likely to choose academic tracks and enroll in higher education than boys, net of academic performance.

On the basis of the theoretical framework laid out in this section, I expect to find that students who have male-typed occupational aspirations are less likely to complete a tertiary degree (expectation 1) and that horizontal gender differences in occupational aspirations explains a substantial share of the gender gap in college completion, net of academic performance (expectation 2).

Aspirations and Anticipatory Decisions

The relationship between aspirations and academic performance is not straight forward. Scholars critical of the rational choice perspectives have pointed out that educational choices are not only a matter of rational action based on antecedent social structures. In his contemporary critique of Boudon’s (1974) seminal book, Hauser (1976) made the argument that Boudon erroneously views education only as investment, and not also as a consumption good. Hauser gave the example that education also fulfills humanistic goals held by Western elites, which have spread to broader segments of the population along with improved living conditions (p. 927). In this sense, educational choices can be understood as part of a broader cultural repertoire for signaling belonging, as well as self-realization. This implies that engagement with school subjects, occupational and educational aspirations and academic performance are intertwined in identity projects that can be associated not only with class but with gender or other social identities.

In Gambetta’s (1987) classic study of educational inequality in Italy, he distinguished between cultural and economic causation on the one hand, and causation at the level of opportunities and causation at the level of preferences on the other (pp. 71–72). The conundrum regarding social class constraints, is the “consistency between constraints and preferences.” That is, the children who experience the most constraints, are also more likely to hold preferences for less education. He highlights that educational decisions are often not active choices but rather “nondecisions” shaped by social processes that limit individual agency. In his interpretation, the causal arrow between preferences and constraints could go either way, arguing that the best approach to understanding the relationship can be found in the following quote: “It was not always clear which came first: the job ambition or the school performance. Sometimes the desire for the job did seem to be the base for the school motivation, yet sometimes a boy who did well in school became slowly convinced that he was good enough to think of a middle class job” (Kahl 1961:358, cited in Gambetta 1987:119–20). This complexity underscores the difficulty in distinguishing between primary and secondary effects, and the causal relationship between them.

In their study of the relative importance of primary and secondary effects in the reproduction of class differentials, Jackson et al. (2007) discussed “anticipatory decisions,” that is, that decisions about future educational transitions may influence academic performance. If students at some earlier point in time make decisions about their educational aspirations, this may influence their subsequent performance. It may look like their performance independently influences their educational choice, as the educational transition is measured later in time than their academic performance, but in fact the reverse could at least partly be the case, that is, that their anticipated choice influenced their performance. Such an anticipated choice may have positive or negative effects on motivation and effort, inaccurately boosting the apparent effect of academic performance on choice (Jackson et al. 2007:222).

The significance of anticipatory decisions depend in part on how sensitive expectations are to feedback. Andrew and Hauser (2011) discussed whether expectations are mostly stable, influenced by social status and significant others early in life, or whether students adapt their expectations on the basis of new information, such as feedback from school. The discussion draws on descriptive decision making theory, associated with the classic work of psychologists Tversky and Kahneman (1988). On the basis of their analyses of longitudinal U.S. survey data on adolescents from 8th grade to 26 years of age, Andrew and Hauser concluded, “Students’ expectations are highly persistent over time, and if students do revise their expectations, they do so by a modest amount and based on rather large changes in their grade point averages” (p. 498). Karlson (2015, 2019), however, argued in several articles that expectations change with feedback from schools. In the first study, Karlson (2015) used tracking placement rather than school grades as a measure of “feedback” and found significant change in expectations associated with higher track placement. In the second study, he used GPA and investigated whether feedback has weaker consequences for educational expectations among lower socioeconomic status youth. He found the opposite, but his findings also indicate that the majority of students (62 percent in his estimation) do not change their expectations in the expected direction on the basis of changes in GPA (Karlson 2019:724). In this study, I assume that anticipatory decisions about educational aspirations or future occupational destinations may at least partly influence academic performance. On the basis of the theoretical framework laid out in this section, I expect to find that students who have male-typed occupational aspirations have lower GPAs than students with female-typed and gender-neutral occupational aspirations (expectation 3) and that horizontal gender differences in occupational aspirations explains a substantial share of the gender gap in academic performance, net of general educational expectations (expectation 4).

Using longitudinal survey data linked to register data in the Norwegian context, I have detailed subjective measures of aspirations and expectations measured in lower secondary school, as well as objective measures of exit GPA and higher education completion obtained at a later point in time. Although the analyses cannot identify the direct causal link between aspirations and outcomes, the unique data allow a qualified assessment of the theoretical expectations. The Norwegian context is interesting because of the country’s relatively egalitarian gender ideology and social structure, and the relatively open and universal education system, where the first transition occurs late, at 16 years of age. Moreover, gender gaps in GPA at the end of lower secondary school are substantial, in favor of girls.

Gender Gaps and Gender Segregation in the Norwegian Education System

Norwegian primary and lower secondary education is uniform and compulsory up to 10th grade, at 15 to 16 years of age (Reisel, Hermansen, and Kindt 2019). At the end of 10th grade, pupils receive an academic scorecard, with teacher graded subject marks from the end of lower secondary school, as well as marks from centrally administered and anonymously graded exams. A GPA is calculated from these marks. The transition to upper secondary school introduces a choice between vocational and academic school tracks. In European terms, this first track choice transition is considered to occur relatively late, which has been associated with a female advantage in education (Scheeren, van de Werfhorst, and Bol 2018). The Norwegian education system also has a strong vocational orientation, which is associated with more gender-typical occupational expectations among boys and higher concentrations of boys in male-dominated fields of study (Hillmert 2015; Reisel, Hegna, and Imdorf 2015).

The pupils apply to their school track of choice according to their own preferences, ranking their top three choices both in terms of study programs and, where possible, in terms of schools. In case of competition for placement in upper secondary school tracks, the pupils with the highest GPAs are admitted. At the end of 10th grade, girls on average receive higher marks than boys in all subjects except physical education. The gender gap has been relatively constant over the past decades, with girls scoring on average about 0.4 marks higher on a scale ranging from 1 to 6 (Statistics Norway 2023a). The gender gap in Oslo is somewhat smaller than in the country as a whole. In Norway, academic performance is more gender balanced in low stakes national tests than in school marks. As in many other countries, literacy and numeracy skills when evaluated in international assessments such as the Programme for International Student Assessment, the Trends in International Mathematics and Science Study, and the Programme for the International Assessment of Adult Competencies show a consistent gender mirroring between literacy and numeracy and no gender gaps in English (foreign language) and science (Borgonovi et al. 2018).

About half of the pupils in upper secondary school are in vocational programs, including the pupils who are in apprenticeships, and this distribution has been stable in recent decade (Statistics Norway 2023b). In Oslo, where the data used in this study were collected, the share of pupils who choose vocational programs is lower. Moreover, this share has gone down from 35 percent in 2010 to 28 percent in 2023 (Statistics Norway 2023b). Most of the vocational education and training (VET) programs are highly gender segregated, with about 5 percent of the pupils choosing study programs that are numerically dominated by the opposite sex (Bufdir 2024).

As in most developed nations, the gender gap in higher education has been reversed over the past 40 years. Whereas 40 percent of students in higher education were female in the early 1970s, about 40 percent of students are now male (Reisel and Seehuus 2022). Nationally, 37 percent of those 19 to 24 years of age were enrolled in higher education, about 3 percentage points up from a decade earlier. In 2023, 45.1 percent of women and 28.8 percent of men in this age range were registered as students in tertiary education. In 2013, when the sample used in the analyses in this article were in their early 20s, the comparable percentages were 40.1 and 27.5, respectively (Statistics Norway 2024). The share of students in higher education is higher in Oslo than the national numbers.

Norwegian higher education is composed mainly of public institutions, and there are no tuition fees for enrollment in university. Some private university colleges charge tuition, but most university colleges are also public, with no tuition fees. Modest amounts of student loans and grants are available for living expenses. In the Norwegian higher education system, there is substantial variation in the number of enrolled students in different fields of study. The three biggest fields of study are business and administration, nursing and care, and information and computer technology, enrolling almost 30 percent of all students combined. Among those who were enrolled in higher education in 2020, one third of all female students were enrolled in the five largest female dominated fields of study (early childhood education, health care, teacher education, psychology and social work) (Reisel and Seehuus 2022).

Data and Variables

This study is based on a longitudinal survey of youth in public schools in Oslo who were born in 1992 (LUNO). There were three waves of data collection from 2006 to 2010. The first wave of data collection (T1) was during the autumn of 2006 (first term, 9th grade, the second year of lower secondary school). The second wave of data collection (T2) occurred 16 to 20 months later, in the spring of 2008 (second term, 10th grade, the last year of lower secondary school), starting the day after the application deadline for upper secondary education. The third wave of data collection (T3) was during the winter of 2009 to 2010 (Hegna 2014). The survey is linked to register data from Norwegian public registries, which makes it possible to monitor educational choices made after the last wave of data collection, up to fall of 2020, when the respondents were 28 years old.

The dependent variables are (1) GPA in 10th grade and (2) completing a higher education degree by 28 years of age. Both variables are taken from population register data. GPA is the average grades in all subjects in lower secondary school, and also includes grades from any (externally graded) final exams taken. The variable is standardized within the analysis sample, such that the mean is 0 and the standard deviation is 1. The GPA scorecards were finalized and communicated to the pupils some months after T2. Higher education completion is measured by highest degree achieved by October 2020, the year the respondents turned 28.

The main independent variables of interest are gender and having male-typed occupation aspirations. Gender is coded as a binary variable, with male coded as 1 and female as 0. Male dominated occupational aspirations (T2) is based on the gender composition of the occupations that the respondents aspired to at T2. The respondents were asked to write in an open rubric which occupation they imagine that they will have as an adult. The answers were hand coded to International Standard Classification of Occupations, 1988 edition (ISCO-88), and coded as male dominated, female dominated, or gender balanced according to labor market shares of younger employees (14–35 years of age) derived from the Norwegian Labor Force Survey and matched to ISCO-88.

Other independent variables are educational expectations (T1): whether students expect to enroll in higher education and, if so, in short (1–4 years) or long (≥5 years) study programs. The likelihood of enrolling in a vocational education program (T1) was based on the question “How likely is it that you will enroll in a vocational program or an academic preparatory program in upper secondary school?” Very likely or quite likely to enroll in vocational education program is coded as 1 and other responses as 0. Given that more boys than girls enroll in VET and more girls than boys enroll in higher education, these variables are included to pick up the general influence of having vocational or higher education aspirations on GPA and college completion, which could be conflated with having male-typed occupational aspirations.

I control for social class and minority status, as well as pubertal maturation and planfulness. In recent years, an emerging literature has investigated how pubertal maturation and personality traits, which tend to differ on average across genders, influence adolescent academic achievement. It is well established that girls on average enter puberty earlier than boys (Marceau et al. 2011). This has inspired hypotheses linking the timing of maturation to gender gaps in academic achievement favoring girls. In a recent study of British twins, Torvik et al. (2021) found that gender differences in the timing of maturation had consequences for the gender gap in school achievement at 16 years of age. The authors speculate that maturation may be influencing school achievement through its association with conscientiousness, such that immature students are less motivated and less capable of systematically engaging with their schoolwork than pupils that are more mature. Conscientiousness is generally associated with higher academic achievement (Mammadov 2022). Pubertal maturation (T1, age 14) is based on the question “Looking at yourself now, did you reach puberty earlier or later than your peers?” The question has five answer categories from 1 (“much earlier”) to 5 (“much later”). In the analyses, the variable is used as a continuous measure and centered on the mean. Planfulness (T2) was created by an index based on five questions, with an internal consistency measured with a Cronbach’s α coefficient of 0.77. The index includes answers to questions about goal setting, planning, reaching goals, and organization of time. The index is standardized with a mean of 0 and a standard deviation of 1 in the analyses.

Social class is based on fathers’ occupation (mothers’ occupation if missing), which has been converted to ISCO-88 codes and coded according to the Goldthorpe class schema (Goldthorpe and McKnight 2006). I distinguish between 1 upper class (upper service class), 2 middle class (lower service class, lower officials, and self-employed; reference group), 3 working class (skilled and unskilled workers), and 4 (not working or don’t know). I also control for minority status, which is defined as having two foreign born parents (20 percent of the sample).

Analytic Strategy

Academic performance, operationalized as 10th grade GPA, is recorded as a continuous variable. To investigate how the independent variables of interest – educational expectations and occupational aspirations mediate the association between gender and GPA, I fit a series of nested models and compare the coefficient for gender (male) across the models, using ordinary least squares (OLS) regression. Comparing the nested regressions, I test the reduction of the coefficient for male across models, using bootstrap standard errors.

I first fit models with the control variables, before adding the main independent variables of interest, which are the likelihood of entering a VET program, the respondents’ expectations toward enrolling in higher education, and finally whether they aspire to a male-typed occupation. Among the control variables, pubertal maturation and planfulness were only significantly associated with GPA for girls. I therefore include interaction terms in the models, so that I do not overestimate the influence of the other explanatory variables through underestimating the influence of the control variables. Including interaction terms for pubertal maturation and planfulness produced models with better model fit than models with only main effects. I tested interactions between gender and the other control variables, but these were not significant and produced poorer model fit and were thus omitted.

The second dependent variable, completion of higher education, is binary, and therefore require a logit or probit analysis rather than OLS regression. In such analyses, coefficients cannot be directly compared across models, as some of the change in coefficients may be due to the residual variance of the models, in addition to any mediation of the variables added to the model (Breen, Bernt Karlson, and Holm 2021). Karlson, Holm, and Breen (2012) (KHB) developed a method for comparing logit or probit coefficients where the mediating variables are residualized with respect to the independent variable of interest, making it possible to compare logit and probit models with the same residual variance. This KHB method separates the true impact of the mediating variables on the coefficient of interest from the change caused by differences in residual variance across the models. This method has recently been revised and simplified, consisting of three steps. First the full model is fitted using a nonlinear probability model such as logit. Second, the predicted values are saved as a latent index of the dependent variable. Third, this new dependent variable, which is now continuous, is used to estimate the mediation of the variables of interest in the model, using nested OLS models and bootstrap standard errors (Breen et al. 2021).

Using the revised KHB method, I consecutively add variables measuring maturation, educational expectations and occupational aspirations to the models, and compare the change across models in the size of the coefficient for male. The analyses indicate the percentage change in the coefficient across models and estimates the significance level of the change using bootstrap standard errors. To substantially interpret the coefficients from the KHB models, I calculate the inverse of the logit coefficients, exponentiating the constant (and coefficient for male) and dividing by 1 plus the exponentiated value. This gives us the predicted margins for female (constant) and male (constant + coefficient for male), which I then subtract and report as the conditional percentage point difference in probabilities of graduating by 28 years of age.

Because the size of change in coefficients across models is partly dependent on the order by which I enter the various independent variables to the models, I run the models several times, varying the order in which I enter the variables. I present the minimum and maximum change produced by entering a variable or set of variables in the different orders. Supplementary analyses are presented in the appendix.

Results

The sample consists of 1,076 adolescents born in 1992, residing in Oslo, with nonmissing values on all the relevant variables. This constitutes about 20 percent of the total population of interest. The descriptive statistics for the sample are presented in Table 1. Comparing the sample with available data from Statistics Norway for the same or adjacent years, the sample is slightly more female than the adolescent population in Oslo as a whole and the GPA scores in the sample are somewhat higher, although the gender gap in GPA is representative. Comparing completion rates for higher education in the sample with the population of interest (by 28 years of age for those born in 1992 in Oslo), the sample has a somewhat higher completion rate than the population of interest (about 5 percentage points higher for both genders).

Table 1.

Descriptive Statistics (n = 1,076).

Variable	Male	Female	Total Sample
Gender	46	54	100
Upper class (I)	28.8	30.0	29.5
Middle class (II, III, IVa, IVb)	40.6	42.4	41.5
Working class (V, VI, VII)	21.1	21.1	21.1
Not working/DKN	9.5	6.2	7.9
Minority	21.7	19.0	20.3
Pubertal development much earlier	10.3	5.7	7.8
A bit earlier	23.9	20.6	22.1
Just the same	47.1	51.3	49.4
A bit later	15.2	17.8	16.6
Much later	3.5	4.6	4.09
Planfulness (1–5)	4.0 (.709)	3.92 (.708)	3.97 (.710)
Male-dominated occupational aspirations	37.7	5.7	20.4
Expect no higher education	14.0	7.2	10.3
Expect short higher education	32.1	34.8	33.6
Expect long higher education	54.0	58.0	56.1
VET not likely	43.4	51.6	47.9
VET likely	56.6	48.4	52.1
GPA	41.3 (7.52)	44.4 (7.00)	43.0 (7.40)
Completed HE	61.9	76.3	69.7

Note: DKN = don’t know; GPA = grade point average; HE = higher education; VET = vocational education and training.

Other observations from the descriptive statistics are that in the sample, girls report later pubertal maturation than boys relative to their peers, and boys score slightly higher than girls on the index for planfulness. It is likely that pubertal maturation picks up perceived maturation within gender rather than across genders, as it is natural to compare gender-specific secondary sex characteristics when asked about pubertal development. Given the lack of alternative sources of information for the planfulness measure, I do not know if the distribution across genders represents the true variation, or whether the sample restrictions skew the sample in favor of somewhat more planful boys.

Chronologically, students first graduate lower secondary school and receive their exit GPA, and later enroll in and potentially graduate from higher education. In the following, I therefore analyze the gender gaps in GPA first, before I turn to the hypotheses regarding graduation from higher education.

Gender-Segregated Occupational Aspirations and 10th Grade Exit GPA

In this section, I focus on the association between occupational aspirations and GPA and to what extent the total association between respondents’ gender and their GPAs is mediated by respondents’ aspirations and expectations, net of social background, maturation and general inclinations to engage in planful behavior. Figure 1 shows the bivariate relationships between the explanatory variables and GPA, by gender. In line with the overall gender gap in GPA, girls in all categories score higher than boys, but the confidence intervals overlap in all the measures, except planfulness and the likelihood of enrolling in VET. As expected, coming from an upper class background is positively associated with GPA, relative to middle class background, while working class social background or having fathers that do not work is associated with lower GPA. Minority students have lower GPA on average than nonminority youth. In line with the most consistent research on pubertal development and academic performance, pubertal development does not influence boys’ GPAs, while later maturation is somewhat positive for girls. Likewise, planfulness seem to have little influence on boys’ GPA but matters for girls. Having expectations of enrolling in higher education is positively associated with GPA for both boys and girls, especially longer study programs (MA or higher). Expecting to enroll in VET is associated with a lower GPA for both boys and girls but more so for boys. Aspiring for a male dominated occupation is also associated with lower GPA, for both boys and girls, but because there is more uncertainty related to this measure for girls, the association is only significantly different from zero for boys.

Figure 1.

Bivariate relationships between GPA and independent variables.

In Table 2, I present six nested OLS regression models. In each model, I consecutively add the explanatory variables, and test how much the coefficient for male is reduced across the models (Table 3). The baseline model contains only the coefficient for male. In the sample, male students on average receive GPAs that are 35 percent of a standard deviation lower than girls. When I control for social class and minority status, the difference is reduced by 8 percent to 32 percent of a standard deviation, but this reduction is not statistically significant. Model 3 introduces the variables for pubertal maturation and planfulness and their interaction terms with gender (both centered on their means). These variables contribute to increasing the gender gap in GPA slightly, but this change is also not statistically significant. In model 4, I introduce the expectation from T1 about enrollment in VET. This contributes to a small but significant reduction in the gender gap in GPA. In model 5, I add higher education expectations, measured at T1, that is, about 1.5 years before final GPA is set. When comparing boys and girls with similar expectations of enrolling in higher education, the gender gap in GPA is reduced to 28.5 percent of a standard deviation (not statistically significant). Holding class background, minority status, maturation and vocational and higher education expectations constant, model 6 adds a dummy variable for having occupational aspirations for a male dominated occupation. Controlling for all other variables, students with male-typed occupational aspirations have GPAs that on average are 18 percent of a standard deviation lower than students with female-typed or gender-neutral occupational aspirations, confirming expectation 3. This variable alone explains 19.7 percent of the gender gap in GPA and reduces the gap to 23 percent of a standard deviation. This confirms expectation 4, that having male-typed occupational aspirations explains a substantial part of the gender gap in GPA. Taken together, educational expectations and occupational aspirations reduce the gender gap in GPA by 32 percent (Table 3). In alternative specifications, in which I enter male-typed occupational aspirations in model 4 (before VET plans and higher education expectations) or in model 3, before the maturation variables, male-typed occupational aspirations account for 21.6 percent and 22.6 percent reductions in the gender gap (see Tables A1 and A2, respectively, in the Appendix).

Table 2.

Multivariate Nested Ordinary Least Squares Regression Models Predicting Exit Grade Point Average in 10th Grade (Standardized).

	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Male	−.347***	−.32***	−.337***	−.314***	−.285***	−.229***
Male	(.05)	(.047)	(.048)	(.047)	(.045)	(.049)
Upper class (I)		.222***	.21***	.187***	.14**	.132*
Upper class (I)		(.057)	(.056)	(.055)	(.053)	(.053)
Working class (V, VI, VII)		−.300***	−.303***	−.278***	−.257***	−.253***
Working class (V, VI, VII)		(.064)	(.064)	(.062)	(.060)	(.060)
Not working		−.630***	−.646***	−.621***	−.587***	−.584***
Not working		(.093)	(.092)	(.091)	(.087)	(.087)
Minority		−.222***	−.243***	−.256***	−.270***	−.273***
Minority		(.061)	(.061)	(.06)	(.057)	(.057)
Pubertal development (c)			.064	.0698*	.069*	.069*
Pubertal development (c)			(.036)	(.035)	(.034)	(.034)
Male × pub dev			−.064^a	−.076^a	−.079^a	−.075^a
Male × pub dev			(.051)	(.05)	(.048)	(.048)
Planfulness (standardized)			.167***	.168***	.133***	.132***
Planfulness (standardized)			(.032)	(.032)	(.031)	(.03)
Male × planfulness			−.158***	−.160***	−.143**	−.138**
Male × planfulness			(.047)	(.047)	(.045)	(.045)
VET likely				−.307***	−.237***	−.228***
VET likely				(.046)	(.045)	(.045)
Expect short HE					.265***	.265***
Expect short HE					(.079)	(.079)
Expect long HE					.624***	.620***
Expect long HE					(.076)	(.076)
Male-typed occupational aspirations						−.176**
Male-typed occupational aspirations						(.06)
Constant	.299***	.379***	.406***	.557***	.076	.086
Constant	(.034)	(.043)	(.043)	(.048)	(.082)	(.082)
n	1,076	1,076	1,076	1,076	1,076	1,076
R ²	0.042	0.158	0.181	0.213	0.278	0.284
Adjusted R²	0.041	0.154	0.174	0.206	0.270	0.275
AIC	2,641	2,510	2,488	2,447	2,358	2,352

Note: Values in parentheses are standard errors. AIC = Akaike information criterion; HE = higher education; VET = vocational education and training.

Not statistically significant.

p < .05. **p < .01. ***p < .001.

Table 3.

Testing Change in Gender Coefficient across Models for Grade Point Average.

	Observed Coefficient	Bootstrap SE	z	P > z	Percentage Change
M1 → M2 (background)	−.028	.0194	−1.42	.154	8.0
M2 → M3 (maturation)	.017	.0129	1.32	.188	−5.3
M3 → M4 (VET likely)	−.023	.0105	−2.16	.030	6.8
M4 → M5 (HE expectation)	−.029	.0148	−1.95	.052	9.2
M5 → M6 (occupational aspiration)	−.056	.0201	−2.80	.005	19.7
M3 → M6 (all motivation variables)	−.108	.0266	−4.05	.000	32.0

Note: Statistically significant reductions in the observed coefficient are highlighted in boldface type. HE = higher education; M = model; VET = vocational education and training.

Tenth Grade GPA, Male-Typed Occupational Aspirations, and Completion of Higher Education

I now turn to gender gaps in the probability of higher education completion by 28 years of age. Using the revised KHB method for comparing variables across models in analyses with binary outcome variables, I present nested models in Table 4 and test the reduction in the coefficient for male using bootstrap standard errors (see Table 5).

Table 4.

Predicting Educational Attainment Using the Revised Karlson, Holm, and Breen Method.

Variable	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Male	−.926*** (.193)	−.868*** (.191)	−.371* (.167)	−.367* (.167)	−.317^a (.167)	−.18^a (.177)
Upper class (I)		.713** (.227)	.432* (.208)	.427* (.207)	.387^a (.208)	.369^a (.208)
Working class (V, VI, VII)		−.572* (.245)	−.030^a (.209)	−.024^a (.210)	−.081^a (.211)	−.076^a (.211)
Not working		−1.44*** (.409)	−.189^a (.309)	−.187^a (.309)	−.232^a (.302)	−.237^a (.302)
Minority		−.241^a (.252)	.154^a (.223)	.146^a (.225)	.113^a (.219)	.100^a (.220)
Exit GPA (standardized) lower secondary school			1.56*** (.133)	1.55*** (.133)	1.5*** (.131)	1.48*** (.130)
Exit GPA²			−.473** (.163)	−.475** (.163)	−.451** (.162)	−.449** (.163)
VET likely				−.117^a (.167)	−.116^a (.167)	−.097^a (.167)
Expect short HE					1.08*** (.266)	1.08*** (.266)
Expect long HE					.877** (.256)	.881** (.253)
Male-typed occupational aspirations						−.453* (.199)
Constant	1.42	1.47	1.16	1.22	.37	.39

Note: Values in parentheses are bootstrap standard errors (500 replications). GPA = grade point average; HE = higher education; VET = vocational education and training.

Not statistically significant.

p < .05. **p < .01. ***p < .001.

Table 5.

Reduction in Male Coefficient across Models Predicting Completion of Higher Education.

Model Comparison	Observed Coefficient	Bootstrap SE	z	P > z	Percentage Change
M1 → M2 (background)	−.058	.0479	−1.22	.223	6.3
M2 → M3 (GPA)	−.496	.0928	−5.35	.000	57.2
M3 → M4 (VET likely)	−.005	.0091	−0.51	.610	1.2
M4 → M5 (HE expectations)	−.050	.0238	−2.10	.036	13.6
M5 → M6 (occupational aspirations)	−.137	.0622	−2.20	.028	43.3
M3 → M6 (all motivation variables)	−.192	.0669	−2.87	.004	51.6

Note: Statistically significant reductions in the observed coefficient are highlighted in boldface type. GPA = grade point average; HE = higher education; M = model; VET = vocational education and training.

Model 1 includes only the coefficient for male. In the sample of youth from Oslo born in 1992, boys had an estimated 18.4 percentage points lower probabilities of graduating from higher education by 28 years of age than girls. This estimate is somewhat higher than the observed difference reported in Table 1. Controlling for class background and minority status does not significantly reduce the male coefficient, indicating that social background, although important for educational attainment, does not explain much of the gender gap in educational attainment. Model 3 adds a control for exit GPA in lower secondary school (standardized) and its squared term to account for any nonlinearities. This reduces the gender gap in the probability of completing higher education by 57 percent (Table 5). The following two models introduce educational expectations from T1, that is, whether respondents thought it likely that they would enroll in VET, and whether they expect to enroll in higher education, in early 9th grade. Controlling for GPA, VET expectations make little difference for the gender gap. Expecting to enroll in higher education reduces the gender gap in the probability of completing higher education by 13.6 percent, indicating that gender differences in educational expectations early on have consequences for educational attainment, over and above academic performance. When adding this variable to the model, the male coefficient also becomes nonsignificant. In other words, when I compare boys and girls with similar social backgrounds, similar GPAs and similar educational expectations, their probability of completing higher education is likely also similar. Finally, I add aspirations toward a male-typed occupation, measured in spring of 10th grade. In line with expectation 1, students who aspired to a male-typed occupation when they were in lower secondary school have 6.8 percent lower probability of completing a higher education degree by 28 years of age than students who have female-typed or gender-neutral occupational aspirations, all else equal. This variable alone, controlling for all the others, explain 43 percent of the remaining gender gap in probability of completing higher education by 28 years of age and reduces the gender difference in probability of graduating to 2.6 percentage points. This is in line with expectation 2, that horizontal gender differences in occupational aspirations explain a substantial share of the gender gap in college completion, net of academic performance.

In alternative specifications of the model (see Table A3), introducing male-typed occupational aspirations in model 4, before VET plans and higher education expectations are included in the model, I find that having male-typed occupational aspirations explain 37 percent of the gender gap in graduation probability, When entering male-typed occupational aspirations without including GPA in model 3 (Table A4), aspirations explain 30 percent of the gap in graduation probability. The gender gap in graduation probability is reduced to 11 percentage points when only controlling for social background and male-typed occupational aspirations, while the gap is 3.4 percentage points when controlling for social background and male-typed occupational aspirations, as well as lower secondary school GPA.

Discussion

In this study, I have addressed the research question of how educational expectations and gender-typed occupational aspirations contribute to explaining gender gaps in academic performance and educational attainment in favor of girls. Male- and female-typed fields of study are typically associated with different educational requirements, providing boys with weaker incentives to invest in education. This may be especially the case in Norway, where track choices are made relatively late, and where the education system has a strong vocational orientation. I expected to find that having male-typed occupational aspirations in lower secondary school was associated with both lower exit GPA and lower probability of completing a higher education degree. I also expected occupational aspirations to explain a substantial portion of the gender gap in academic performance and educational attainment. On the basis of analyses of a longitudinal survey of Norwegian youth linked to administrative register data, the findings provide several important insights into the mechanisms underlying the gender disparities in educational attainment in favor of girls.

First, the results confirm that there is a strong association between educational expectations and male-typed occupational aspirations on the one hand, and academic performance on the other. Theoretically, academic performance or beliefs about academic abilities prior to the distribution of school marks, may be influencing adolescents’ expectations and aspirations. However, in line with the debate on the stability of educational expectations and the role of anticipatory decisions (Andrew and Hauser 2011; Jackson et al. 2007; Karlson 2015, 2019) it could also be that educational and occupational expectations formed early in secondary education have consequences for school motivation and academic performance.

Second, when comparing boys and girls with similar social backgrounds and lower secondary school exit GPAs, I found that educational expectations and occupational aspirations explain 14 percent to 23 percent and 37 percent to 43 percent, respectively, of the gender gap in the probability of completing higher education by 28 years of age, depending on the order in which they are entered into the models. Taken together, educational expectations and occupational aspirations account for 52 percent of the gender gap in the probability of completing higher education, when comparing boys and girls with similar social backgrounds and GPAs. This substantial explanatory power of occupational aspirations, in particular, supports the arguments put forth by Morgan et al. (2013) and Barone and Assirelli (2020) regarding the importance of occupational plans in shaping educational trajectories.

Interpreting these findings within the theoretical framework of primary and secondary effects of stratification (Boudon 1974; Breen and Goldthorpe 1997), I suggest that the primary effect of gender in education would be the influence of gender related traits, expectations and experiences on academic performance, including anticipatory decisions related to gender-typed occupational aspirations and educational expectations. Gender differences in educational transitions, such as gender gaps in the completion of higher education among male and female students with the same levels of academic performance, would be the secondary effects of gender. The findings in this study indicate that gender-typed occupational aspirations significantly contribute to these secondary effects.

Although there may be a rational element to the relationship between aspirations, performance, and attainment, rooted in the different role educational attainment plays in the labor market opportunities of men and women, the strong explanatory power of gender-typed occupational aspirations also aligns with culturalist explanations of gender differences in educational choices (Charles and Bradley 2009). It suggests that gendered notions of appropriate occupations and self-realization play a significant role in shaping both academic performance and educational attainment. This finding highlights the need to consider how cultural beliefs about gender-appropriate careers may be reinforcing educational inequalities.

It should be noted that the findings regarding pubertal maturation did not significantly explain the gender gap in academic performance. This contrasts with some previous research (e.g., Koerselman and Pekkarinen 2017) but aligns with others (e.g., Suutela et al. 2022) in suggesting that maturation differences may not be a primary driver of gender gaps in education.

Limitations

Although this study provides valuable insights, it is not without limitations. The timing of the measures, particularly the collection of occupational aspirations data at T2, makes it challenging to fully disentangle the causal relationship between aspirations and academic performance. Future research should aim to measure occupational aspirations at multiple time points, including earlier time points, to better assess their role in shaping academic outcomes.

Additionally, the measure of pubertal maturation likely reflects relative maturation within gender rather than absolute differences between genders. More precise measures of physical and cognitive maturation could provide clearer insights into how developmental differences might contribute to gender gaps in education.

Conclusion

This study contributes to our understanding of gender gaps in education by highlighting the role of gender-typed occupational aspirations and educational expectations. The findings suggest that these factors explain a substantial portion of gender differences in both academic performance and educational attainment.

These results have important implications for educational policy and practice. Efforts to reduce gender gaps in education should not focus solely on addressing disparities in academic performance but should also consider how to broaden students’ occupational horizons and challenge gendered notions of appropriate careers. Interventions that expose students to a wide range of career possibilities and challenge gender stereotypes in occupations could potentially help reduce educational gender gaps.

In conclusion, the study demonstrates that understanding and addressing gender gaps in education requires looking beyond academic performance to consider the broader social and cultural contexts that shape students’ aspirations and expectations. Future studies could benefit from a more detailed examination of how occupational aspirations form and change over time, and how they interact with feedback from academic performance. Longitudinal designs with more frequent data collection points for occupational aspirations as well as GPA and various measures of maturation and motivation could help unpack these dynamics.

Footnotes

Appendix

Table A4.

Reduction in Male Coefficient across Models Predicting Completion of Higher Education: Alternative Specification with Occupational Aspirations Entered in Model 3.

Coefficient Reduction	Observed Coefficient	Bootstrap SE	z	P > z	Percentage Change
M1 → M2 (background)	−.0583548	.0479016	−1.22	.223	6.3
M2 → M3 (occupational aspirations)	−.2615566	.0785249	−3.33	.001	30.2
M3 → M4 (GPA)	−.3716703	.0970294	−3.83	.000	61.3
M4 → M5 (VET likely)	−.0016071	.0067064	−0.24	.811	.69
M5 → M6 (HE expectations)	−.0531812	.0254218	−2.09	.036	22.9

Note: Statistically significant reductions in the observed coefficient are highlighted in bold. GPA = grade point average; HE = higher education; M = model; VET = vocational education and training.

Acknowledgements

I am indebted to Kristinn Hegna and Christian Imdorf for initial survey data collection and coding work. A special thanks to the LUNO team for hand coding the open answers on occupational aspirations to ISCO-88 occupational codes. I would also like to thank Sara Seehuus for useful feedback on multiple drafts. An earlier version of this article was presented at the RC28 stream at the XX ISA World Congress of Sociology in 2023. I am grateful for the helpful comments provided by the conference participants and for the constructive feedback from the two anonymous reviewers.

Correction (October 2025):

This article has been updated with the missing Acknowledgments section.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Research Council of Norway (grants 283603 and 212293).

ORCID iD

Liza Reisel

Author Biography

Liza Reisel is a research professor at the Institute for Social Research in Oslo, Norway. She earned her PhD in sociology from City University of New York in 2010. Her research address inequalities in education, gender segregation in the labor market, equality and diversity legislation, and education and labor market outcomes among children of immigrants. Her recent publications include “Unpacking the Logics of Gendered Educational Choices: 10th Graders’ Evaluation of Appropriate Educational Tracks” (2023) in Educational Review and “The Equality Hurdle: Resolving the Welfare State Paradox” (2024) in Work, Employment and Society.

References

Andrew

Megan

Hauser

Robert M.

2011. “Adoption? Adaptation? Evaluating the Formation of Educational Expectations.” Social Forces 90(2):497–520.

Barone

Carlo

Assirelli

Giulia

. 2020. “Gender Segregation in Higher Education: An Empirical Test of Seven Explanations.” Higher Education 79(1):55–78.

Becker

Rolf

. 2014. “Reversal of Gender Differences in Educational Attainment: An Historical Analysis of the West German Case.” Educational Research 56(2):184–201.

Borgonovi

Francesca

Ferrara

Alessandro

Maghnouj

Soumaya

. 2018. “The Gender Gap in Educational Outcomes in Norway.” OECD Education Working Papers No. 183. Paris: OECD Publishing.

Boudon

Raymond

. 1974. Education, Opportunity, and Social Inequality. New York: John Wiley.

Brandlistuen

Ragnhild E.

Flatø

Martin

Stoltenberg

Camilla

Helland

Siri S.

Wang

Mari V.

2021. “Gender Gaps in Preschool Age: A Study of Behavior, Neurodevelopment and Pre-academic Skills.” Scandinavian Journal of Public Health 49(5):503–10.

Breen

Richard

Goldthorpe

John H.

1997. “Explaining Educational Differentials: Towards a Formal Rational Action Theory.” Rationality and Society 9(3):275–305.

Breen

Richard

Karlson

Kristian Bernt

Holm

Anders

. 2021. “A Note on a Reformulation of the KHB Method.” Sociological Methods & Research 50(2):901–12.

Bufdir (Barne-, Ungdoms- og Familiedirektoratet). 2024. “Utdanning i et Kjønnsperspektiv” [“Education from a Gender Perspective”]. Barne-, Ungdoms- og Familiedirektoratet. Retrieved July 5, 2024. https://www.bufdir.no/statistikk-og-analyse/kjonnslikestilling/utdanning-kjonn.

10.

Casad

Bettina J.

Hale

Patricia

Wachs

Faye L.

2015. “Parent-Child Math Anxiety and Math-Gender Stereotypes Predict Adolescents’ Math Education Outcomes.” Frontiers in Psychology 6:1597.

11.

Charles

Maria

Bradley

Karen

. 2009. “Indulging our gendered selves? Sex segregation by field of study in 44 countries”. American Journal of Sociology 114(4):924–976.

12.

Cornwell

Christopher

Mustard

David B.

Van Parys

Jessica

. 2013. “Noncognitive Skills and the Gender Disparities in Test Scores and Teacher Assessments: Evidence from Primary School.” Journal of Human Resources 48(1):236–64.

13.

Correll

Shelley J.

2004. “Constraints into Preferences: Gender, Status, and Emerging Career Aspirations.” American Sociological Review 69(1):93–113.

14.

DiPrete

Thomas A.

Buchmann

Claudia

. 2013. The Rise of Women: The Growing Gender Gap in Education and What It Means for American Schools. New York: Russell Sage.

15.

Fidjeland

Andreas

Rege

Mari

Solli

Ingeborg F.

Størksen

Ingunn

. 2023. “Reducing the Gender Gap in Early Learning: Evidence from a Field Experiment in Norwegian Preschools.” European Economic Review 154:104413.

16.

Gambetta

Diego

. 1987. Were They Pushed or Did They Jump? Individual Decision Mechanisms in Education. Cambridge: Cambridge University Press.

17.

Gershenson

Seth

Holt

Stephen B.

2015. “Gender Gaps in High School Students’ Homework Time.” Educational Researcher 44(8):432–41.

18.

Goldthorpe

John H.

McKnight

Abigail

. 2006. “The Economic Basis of Social Class.” Pp. 109–36 in Mobility and Inequality: Frontiers of Research in Sociology and Economics, edited by Morgan

S. L.

Grusky

D. B.

Fields

G. S.

Stanford, CA: Stanford University Press.

19.

Graetz

Georg

Karimi

Arizo

. 2022. “Gender Gap Variation across Assessment Types: Explanations and Implications.” Economics of Education Review 91:102313.

20.

Hadjar

Andreas

Buchmann

Claudia

. 2016. “Education Systems and Gender Inequalities in Educational Attainment.” Pp. 159–84 in Education Systems and Inequalities, edited by Hadjar

Gross

Bristol, UK: Policy Press.

21.

Hauser

Robert M.

1976. “On Boudon’s Model of Social Mobility.” American Journal of Sociology 81(4):911–28.

22.

Hegna

Kristinn

. 2014. “Changing Educational Aspirations in the Choice of and Transition to Post-compulsory Schooling—A Three-Wave Longitudinal Study of Oslo Youth.” Journal of Youth Studies 17(5):592–613.

23.

Hillmert

Steffen

. 2015. “Gender Segregation in Occupational Expectations and in the Labour Market: International Variation and the Role of Education and Training Systems.” Pp. 123–48 in Gender Segregation in Vocational Education. Bingley, UK: Emerald Group.

24.

Jackson

Michelle

Erikson

Robert

Goldthorpe

John H.

Yaish

Meir

. 2007. “Primary and Secondary Effects in Class Differentials in Educational Attainment: The Transition to a-Level Courses in England and Wales.” Acta Sociologica 50(3):211–29.

25.

Kahl

Joseph A.

1961. “Common Man Boys.” Pp. 348–66 in Education, Economy, and Society: A Reader in the Sociology of Education, edited by Halsey

A. H.

Floud

Anderson

C. A.

New York: Free Press of Glencoe.

26.

Karlson

Kristian Bernt

. 2015. “Expectations on Track? High School Tracking and Adolescent Educational Expectations.” Social Forces 94(1):115–41.

27.

Karlson

Kristian Bernt

. 2019. “Expectation Formation for All? Group Differences in Student Response to Signals About Academic Performance.” Sociological Quarterly 60(4):716–37.

28.

Karlson

Kristian Bernt

Holm

Anders

Breen

Richard

. 2012. “Comparing Regression Coefficients between Same-Sample Nested Models Using Logit and Probit: A New Method.” Sociological Methodology 42(1):286–313.

29.

Koerselman

Kristian

Pekkarinen

Toumas

. 2017. The Timing of Puberty and Gender Differences in Educational Achievement. IZA Discussion Papers (No. 10889).

30.

Kuhn

Andreas

Wolter

Stefan C.

2022. “Things versus People: Gender Differences in Vocational Interests and in Occupational Preferences.” Journal of Economic Behavior & Organization 203(C):210–34.

31.

Lekholm

Alli Klapp

Cliffordson

Christina

. 2008. “Discrepancies between School Grades and Test Scores at Individual and School Level: Effects of Gender and Family Background.” Educational Research and Evaluation 14(2):181–99.

32.

Mammadov

Sakhavat

. 2022. “Big Five Personality Traits and Academic Performance: A Meta-analysis.” Journal of Personality 90(2):222–55.

33.

Marceau

Kristine

Ram

Nilam

Houts

Renate M.

Grimm

Kevin J.

Susman

Elizabeth J.

2011. “Individual Differences in Boys’ and Girls’ Timing and Tempo of Puberty: Modeling Development with Nonlinear Growth Models.” Developmental Psychology 47(5):1389.

34.

Master

Allison

. 2021. “Gender Stereotypes Influence Children’s Stem Motivation.” Child Development Perspectives 15(3):203–10.

35.

Morgan

Stephen L.

2012. “Models of College Entry in the United States and the Challenges of Estimating Primary and Secondary Effects.” Sociological Methods & Research 41(1):17–56.

36.

Morgan

Stephen L.

Gelbgiser

Dafna

Weeden

Kim A.

2013. “Feeding the Pipeline: Gender, Occupational Plans, and College Major Selection.” Social Science Research 42(4):989–1005.

37.

Passolunghi

Maria Chiara

Ferreira

Tania Irene Rueda

Tomasetto

Carlo

. 2014. “Math–Gender Stereotypes and Math-Related Beliefs in Childhood and Early Adolescence.” Learning and Individual Differences 34:70–76.

38.

Reisel

Liza

Hegna

Kristinn

Imdorf

Christian

. 2015. “Gender Segregation in Vocational Education: Introduction.” Pp. 1–22 in Gender Segregation in Vocational Education, Comparative Social Research, edited by Imdorf

Reisel

Hegna

Bingley, UK: Emerald Group.

39.

Reisel

Liza

Hermansen

Are Skeie

Kindt

Marianne Takvam

. 2019. “Norway: Ethnic (In)Equality in a Social-Democratic Welfare State.” Pp. 843–84 The Palgrave Handbook of Race and Ethnic Inequalities in Education, edited by Stevens

P.A.J.

Dworkin

A. G.

Cham, Switzerland: Springer International.

40.

Reisel

Liza

Seehuus

Sara

. 2022. “Flere Kvinner Enn Menn i Høyere Utdanning” [“More Women Than Men in Higher Education”]. Oslo, Norway: Institute for Social Research.

41.

Rosenbaum

James E.

2001. Beyond College for All: Career Paths for the Forgotten Half. New York: Russell Sage.

42.

Saygin

Perihan O.

2020. “Gender Bias in Standardized Tests: Evidence from a Centralized College Admissions System.” Empirical Economics 59(2):1037–65.

43.

Scheeren

Lotte

van de Werfhorst

Herman G.

Bol

Thijs

. 2018. “The Gender Revolution in Context: How Later Tracking in Education Benefits Girls.” Social Forces 97(1):193–220.

44.

Smith

Emil

Reimer

David

. 2024. “Understanding Gender Inequality in Children’s Reading Behavior: New Insights from Digital Behavioral Data.” Child Development 95(2):625–35.

45.

Statistics Norway. 2023a. “Tabell 07495: Grunnskolepoeng, etter Statistikkvariabel, Kjønn, Foreldrenes Utdanningsnivå og År” [“Table 07495: Primary School Credits, by Statistical Variable, Gender, Parents’ Educational Level and Year”]. Retrieved August 12, 2024. https://www.ssb.no/statbank/table/07495.

46.

Statistics Norway. 2023b. “Tabell 09378: Elever, Lærlinger og Lærekandidater i Videregående Opplæring, etter Studieretning/Utdanningsprogram (Todelt) (F) 2010–2023” [“Students, Apprentices and Teaching Candidates in Upper Secondary Education, by Field of Study/Educational Program (Two-Part)”]. Retrieved July 5, 2024. https://www.ssb.no/statbank/table/09378.

47.

Statistics Norway. 2024. “Studenter i Universitets- og Høgskoleutdanning” [“Students in University and College Education”]. Retrieved November 29, 2024. https://www.ssb.no/utdanning/hoyere-utdanning/statistikk/studenter-i-universitets-og-hogskoleutdanning.

48.

Stoet

Gijsbert

Geary

David C.

2022. “Sex Differences in Adolescents’ Occupational Aspirations: Variations across Time and Place.” PLoS One 17(1):e0261438.

49.

Suutela

Maria

Miettinen

Päivi J.

Kosola

Silja

Rahkonen

Ossi

Varimo

Tero

Tarkkanen

Annika

Hero

Matti

Raivio

Taneli

. 2022. “Timing of Puberty and School Performance: A Population-Based Study”. Frontiers in Endocrinology 13:936005.

50.

Torvik

Fartein Ask

Flatø

Martin

McAdams

Tom A.

Colman

Ian

Silventoinen

Karri

Stoltenberg

Camilla

. 2021. “Early Puberty Is Associated with Higher Academic Achievement in Boys and Girls and Partially Explains Academic Sex Differences.” Journal of Adolescent Health 69(3):503–10.

51.

Tversky

Amos

Kahneman

Daniel

. 1988. “Rational Choice and the Framing of Decisions.” Pp. 167–92 in Decision Making: Descriptive, Normative, and Prescriptive Interactions, edited by Bell

D. E.

Raiffa

Tversky

New York: Cambridge University Press.