Because of the importance of gender role attitudes (GRA) for both academic and social outcomes, it is crucial to understand how GRA is constructed and changes over time. A systematic literature review was conducted to look into the relationship between youngsters’ GRA and individual, home, and school characteristics. Thirty-five international studies were identified through searches in different databases. The review reveals that the studies mostly apply a deterministic view to studying the construction of GRA, focusing predominantly on parent–child transmission. Effects of the school environment and individuals’ own life experiences are under-studied. Also, data are mostly cross-sectional and leave little room for investigating evolutions of GRA over time. Suggestions for future research are formulated focusing on (a) a life-course approach that considers GRA as situated, experience-related, and therefore changing over time and (b) an intersectionality-informed approach investigating GRA at the intersects of multiple diversity dimensions.
- gender role attitudes
- systematic literature review
Gender role attitudes (GRA) have been identified as important predictors of scholastic outcomes. Traditional GRA is found to negatively affect academic participation (Vella, 1994) and academic achievement through its influence on students’ sense of school belonging (Huyge, Van Maele, & Van Houtte, 2015) and students’ expectations for academic success (Davis & Greenstein, 2009; Davis & Pearce, 2007; McDaniel, 2010).
GRA have also been identified as predictors for decision processes and behavior (Davis & Greenstein, 2009; Vespa, 2009) such as beliefs about violent and criminal behavior (Ben-David & Schneider, 2005; Flood & Pease, 2009) and expectations for, and participation in, future life activities (Corrigall & Konrad, 2007; Davis & Pearce, 2007; Kaufman, 2005) like post-secondary education (Corrigall & Konrad, 2007; Crocket & Beal, 2012; Davis & Pearce, 2007), labor-market participation, and career choices (Corrigall & Konrad, 2007; Cunningham, 2008b; Kaufman, 2005; Tiklin, Croxford, Ducklin, & Frame, 2005; Vella, 1994). Because of the importance of GRA to both academic and social outcomes, it is important to understand how GRA is constructed and changes over time. GRA construction is interesting especially in adolescence, when career exploration and initial career decision making are central (e.g., Rogers & Creed, 2011; Swanson & D’Archiardi, 2005). Gaining insights into the construction process of GRA can provide us with leads for (educational) interventions to alter GRA, which might positively influence youngsters’ scholastic outcomes and future labor-market participation.
The Sex Versus Gender Debate
Two paradigms with regard to differences between men and women evolved in the 20th century (Kessler & McKenna, 1978). A biological deterministic view considers differences between boys and girls as innate biological or evolutionary. A sociological approach considers these differences as socially and culturally shaped (Archer & Lloyd, 2002; Carrigan, Connell, & Lee, 1985; Eurydice Network, 2010; Kessler & McKenna, 1978). This debate is also visible in the use of terminology (Unger, 1979), with sex mostly referring to biological components and gender referring to social, psychological, and cultural components (Archer & Lloyd, 2002; Beere, 1990; Kessler & McKenna, 1978; Lindsey, 2011; Unger, 1979).
Many scholars have critiqued the sex–gender distinction and argue that both gender and sex are culturally constructed and multi-faceted (see, for example, Butler, 1990). The terminology used by scholars depends on theoretical roots and disciplinary differences, and up until now, no consensus has been established as to which term is preferably used.
In this study, we use gender-related terminology (e.g., GRA) that is congruent with recent literature and in line with Ungers’s (1979) reasoning, “Biologically determinist models of sex differences make it less likely that environmental sources of such differences will be explored” (p. 1085).
Roles can be defined as the behavior that is expected from a person holding a particular status or position within a specific context (Kessler & McKenna, 1978; Lindsey, 2011). A gender role, therefore, can be described as the expected behavior toward an individual, based on the individual’s sex (Boehnke, 2011; Kessler & McKenna, 1978; Lindsey, 2011) and includes different components such as interests, skills, activities, and clothing (Kessler & McKenna, 1978). Individuals’ attitudes or views on the appropriate roles for men and women are labeled with different terms such as gender role attitudes, sex role attitudes, and gender ideology (Davis & Greenstein, 2009; Legge & Misra, 1998; McHugh & Frieze, 1997). We use the term GRA because this is recently the most dominant discourse.
GRA are mostly set out on an axis with two extremes: egalitarian and traditional. Individuals with a traditional view support a gendered division of family labor, regarding women as homemakers and responsible for parenting and men as the wage earner. Egalitarian GRA include a more equal view on participation in both paid labor and domesticities (Davis & Pearce, 2007; Katz-Wise, Priess, & Hyde, 2010; Legge & Misra, 1998).
Because GRA is a latent concept, it is not easily measurable (McHugh & Frieze, 1997; Vespa, 2009) and a variety of operationalizations is used. Caution is needed when comparing studies (Beere, 1990; McHugh & Frieze, 1997). Davis and Greenstein (2009) identified six recurring dimensions in GRA conceptualizations: (a) belief in separate spheres based on gender, (b) primacy of the breadwinner role, (c) motherhood and the feminine self, (d) working women and relationship quality, (e) household utility, and (f) male-privilege acceptance.
Construction of GRA
Diverse developmental theories have revealed different mechanisms influencing the development of gender-related concepts, including the acquisition of GRA (Maccoby, 2000). Three perspectives have been especially important: (a) cognitive-developmental theory (Kohlberg, 1966; Kohlberg & Ullian, 1974), (b) social-learning theory (Bandura, 1969, 1971), and (c) life-course perspective (Elder, 1998; Elder, Johnson, & Crosnoe, 2008).
An attempt to reconcile these three perspectives was postulated by Bussey and Bandura’s (1999) social-cognitive theory on gender construction (Lindsey, 2011). Bussey and Bandura (1999) stressed the triadic and reciprocal relationship between personal (e.g., cognition and biological), behavioral, and environmental aspects. This triadic relationship is subject to the situation, time, and activities in which the development of GRA takes place.
This theoretical framework integrates (a) intrapsychic cognitive mechanisms that are influential in the acquisition of GRA such as cognitive capabilities, structures, and organization processes (Bussey & Bandura, 1999; Kohlberg, 1966; Kohlberg & Ullian, 1974) and (b) social components of GRA development like modeling, observational, and vicarious learning processes and reinforcement (Bandura, 1969; Bussey & Bandura, 1999). The social environment is defined by several actors such as parents, peers, non-familial adults, and teachers (Bandura, 1969, Bussey & Bandura, 1999; Maccoby, 2000). Different actors can serve simultaneously as a role model (Bandura, 1969).
Bussey and Bandura (1999) noted as follows:
Through cognitive processing of direct and vicarious experiences, children come to categorize themselves as girls or boys, gain substantial knowledge of gender attributes and roles, and extract rules as to what types of behavior are considered appropriate for their gender. (p. 696)
In line with (c) life-course perspective, human development over its life span and within its social, cultural, and historical context is central. Life experiences and transitions that influence individuals and groups of people (e.g., cohorts) through time are also looked into (Bussey & Bandura, 1999; Elder, 1998; Elder et al., 2008; Vespa, 2009).
Research on GRA development in adults has demonstrated that adults’ GRA shift as a result of changes in life situations, such as the transition to parenthood (Bolzendahl & Myers, 2004; Katz-Wise et al., 2010; Vespa, 2009) and work experiences (Bolzendahl & Myers, 2004; Vespa, 2009). Similarly, changes in youngsters’ GRA could be triggered by changes in life situations such as pubertal development, increasing importance of peers, and exposure to new school environments. During childhood and emerging adulthood exploration of different life situations (e.g., college and first working experience; Arnett, 2000, 2007), career exploration and initial career decision making (e.g., Rogers & Creed, 2011; Swanson & D’Archiardi, 2005) are central.
The gender-intensification hypothesis states that in adolescence, an acceleration of socialization in traditional GRA takes place (Clemans, DeRose, Graber, & Brooks-Gunn, 2010; Galambos, 2009; Hill & Lynch, 1983; Pettitt, 2004). Peers (Galambos, 2009; Hill & Lynch, 1983; Pettitt, 2004; Priess & Lindberg, 2011), parents (Galambos, 2009; Hill & Lynch, 1983; Priess & Lindberg, 2011), and schools (Galambos, 2009; Hill & Lynch, 1983) are mostly named as critical social agents in this process. Empirical evidence for the gender-intensification hypothesis is scarce and at times conflicting (e.g., Galambos, 2009; Priess & Lindberg, 2011; Priess, Lindberg, & Hyde, 2009).
Previous research on youngsters’ GRA has, however, focused primarily on parent–child transmission of GRA. Less attention was paid to the role of secondary social agents such as school, teachers, and peers. As youngsters grow older, their exposure—socialization, schooling, and personal experiences—becomes more important (Bolzendahl & Myers, 2004; Davis, 2007; Galambos, 2009; Vespa, 2009). The few studies investigating the relationship with schooling show that schooling has an important role in gender socialization (e.g., Blumberg, 2015; Bolzendahl & Myers, 2004; Council of Europe, 2014; Cunningham, Beutel, Barber, & Thornton, 2005; Delamont, 2012; Shu, 2004; Stromquist, 2007). Because empirical evidence concerning predictors of GRA is scattered, it is needed to synthesize the results of multiple studies to get an insight into the possible simultaneous effects of multiple spheres of influence.
Bearing this in mind, a systematic literature review was performed involving articles that focus on the construction and evolution of GRA of school-aged youngsters up to emerging adulthood (late teens up to mid- to late 20s; Arnett, 2000, 2007), their family, and their school (including peers and teachers).
The following research question is put forward:
Research Question: What is the relationship between school-aged youngsters’ GRA and individual, home, and school characteristics?
Preliminary Search of Review Studies
A systematic review approach was applied, which implies that the authors searched literature in a transparent and systematic way (Petticrew & Roberts, 2006). The research started with an exploration of already existing reviews that focus on the influences on GRA. The following databases were used to detect existing review articles: The Evidence for Policy and Practice Information and Co-ordinating Centre (EPPI-center), education database, Web of Knowledge, Campbell Systematic Reviews, Campbell Collaboration Library of Systematic Reviews, Cochrane Library, Centre for Reviews and Dissemination, European Sociological Review, ERIC (via EBSCOhost [Elton Bryson Stephens Company host]; see https://www.ebscohost.com), Education Full Text (H. W. Wilson, via ESBCO), and Sociological abstracts. In this search, the selected type of document was “review,” and the following keywords were used: GRA, sex role attitudes, gender ideology, gender attitudes, gender role ideology, sex role ideology, and sex attitudes. We were able to identify eight reviews (Azuma & Suzuki, 1991; Banks, 1995; Davis & Greenstein, 2009; Gibbons, Hamby, & Dennis, 1997; Glover, 1992; Jugovic, 2010; Kane, 2000; McHugh & Frieze, 1997). Most reviews were conducted in the 1990s and do not use a systematic approach (Petticrew & Roberts, 2006). They focus mainly on the conceptualization and measurement of GRA (Jugovic, 2010; McHugh & Frieze, 1997) or cross-cultural differences (e.g., Gibbons et al., 1997; Kane, 2000). Two out of eight reviews were not fully available in English (Azuma & Suzuki, 1991; Jugovic, 2010). Although it should be acknowledged that the review of Davis and Greenstein (2009) is valuable when focusing on different socializing levels, no review was found that includes predictors of GRA on multiple levels (individual, family, school) using a systematic review approach on school-aged youngsters’ taking into consideration articles from different countries.
Search Strategy Systematic Review
The following databases were used in detecting articles: Web of Knowledge, PsycInfo, Sociological Abstracts, and ERIC (latter two via ProQuest). Research was limited to English peer-reviewed articles: peer-reviewing ensures the appropriate academic level. All publication years were included.
The first data search was performed in July 2014. Table 1 gives an overview of the search strings that were used. Within-database duplicates were detected, and the following numbers of unique records per database were identified: Web of Knowledge: 112, PsycInfo: 257, ERIC: 103, and Sociological Abstracts: 95. In total, this resulted in 567 records. After the elimination of between-database duplicates, 185 unique records remained. Figure 1 gives a visual overview of the search phases (Liberati et al., 2009).
The second search was performed in November 2014 and focused on detecting more articles related to the school environment, including peers and teachers within school. The search strings are also included in Table 1. A flowchart of this second search can be found in Figure 2. In total, 274 records were identified through searching Web of Knowledge (85), PsycInfo (96), ERIC (51), and Sociological abstracts (42). This resulted in 185 unique records.
Furthermore, the researchers applied snowballing techniques on the collected literature.
Inclusion and Exclusion Criteria
The articles were selected in three separate selection phases. First, in the database research phase, the following inclusion criteria were applied: (a) peer-reviewed articles and (b) fully available in English. In selection phase 2, both title and abstract of the unique records were screened for relevance. Articles were included if the focus was on the construction of school-going youngsters’ GRA, including variables related to individual, family, and school level. Articles that discussed only the measurement of GRA or the influence of GRA on other variables (e.g., behavior) were eliminated. In the third phase, full articles were read and, in total, 28 studies met the criteria for inclusion in the systematic research. The snowballing technique offered another eight articles, which resulted in 36 articles that were quality-appraised.
The quality appraisal tools for quantitative research articles designed by the National Institute for Health and Care Excellence (NICE) were used. The quality appraisal checklist for (a) quantitative intervention studies and (b) quantitative studies reporting correlations and association were used. Although there were quality differences between the articles, only one article was excluded in this research phase because it provided insufficient information to interpret and discuss in depth the results at the same level as the other articles. The reader should bear in mind that several studies did not meet external validity requirements, which implies that the results are not generalizable to populations other than the studied target group. After this final selection step, 35 articles were included for this review and analyzed in depth.
Describing the Sample
Table 2 gives an overview of the 35 included articles in the review. The studies cover different study areas in 10, mostly western, countries. Furthermore, it should be noted that Kulik’s (2000a, 2000b, 2002a, 2002b) contributions determine the high presence of Israeli studies. Data were collected between 1960 and 2013. All the included articles were published after 1980: 1980s (12), 1990s (7), 2000s (12), and 2010s (4).
Table 3 provides an overview of the studied variables. Twenty-four studies focused on both boys (B) and girls (G), with 11 studies performing separate analysis for boys and girls. Eleven studies focused solely on girls, none on boys only. All but one of the studies including data on parents (30) contained information about the mother (M, n = 29) and 19 studies included data of the respondents’ father (F). Only five studies included school-related variables (other than the years of schooling or school grade/year), with only one related to peers (Chatterjee & McCarrey, 1989) and another one focusing on teachers (Zulich, 1986).
Although there was no restriction on the research methodology, all included articles use a quantitative approach. The majority of studies collected own data and 11 used secondary data. Most of the research is cross-sectional and nine studies describe longitudinal data. Only one study reports an experimental design (Erden, 2009).
The majority of the studies (22) researched differences between respondents. Furthermore, different forms of regression analysis (16) and correlational analysis (11) are strongly represented. The use of other statistical techniques such as loglinear analysis (2) and structural equation modeling (SEM) analysis (5) is limited.
Although all included studies focus on GRA, different terminology was used (see Table 2). Moreover, only eight studies explicitly define this concept. A detailed analysis of the measurement scales shows that the implicit definitions match the definition presented in “Theorizing GRA” section.
The measurement of GRA is diverse: some authors used a self-developed tool (12), while others relied on available measurement tools used in national surveys or other accessible data sources (11 studies). Furthermore, established scales were used integrally or in combination with other scales, whereof the short version of the Attitudes Toward Woman Scale (ATW) developed by Spence and colleagues (Spence & Helmreich, 1972; Spence, Helmreich, & Stapp, 1973) was cited the most. Measurement tools were diverse in amount of included items and the roles being questioned.
Due to heterogeneity in research contexts, respondents, and methods of analysis, we opt for a narrative analysis. This narrative approach to systematic review focuses on synthesizing and summarizing articles narratively as opposed to meta-analysis (Petticrew & Roberts, 2006). In synthesizing the results of the studies, we use the terminology originally used by the authors as much as possible. We do so to illustrate the theoretical traditions in which the studies have been performed, which adds to understanding the results.
An overview of all variables examined within the included studies (N = 72) can be found in Table 3. We structure this presentation of results in a framework built on three topics: the relationship between youngsters’ GRA and characteristics related to (a) the individual level, (b) the home environment, and (c) the school environment.
The Individual Level
The majority of predictors that have been studied at the individual level relate to socio-demographic identity markers.
Traditionalism and egalitarianism
Evolutions in GRA in Western society can be noted (Cotter, Hermsen, & Vanneman, 2011; Cunningham, 2008a; Pew Research Center, 2010; Thornton & Freedman, 1979; Valentova, 2012) in two ways. First, four American longitudinal articles show that respondents have become more egalitarian over time (Davis & Wills, 2010; Fan & Marini, 2000; Rollins & White, 1982; Tallichet & Willits, 1986). Second, several studies report youngsters having more egalitarian views compared with their parents (Burt & Scott, 2002; Carlson & Knoester, 2011; Ex & Janssens, 1998; Thornton, Alwin, & Camburn, 1983). Kulik (2000b, 2002a, 2002b) in Israel questioned this cohort-effect. The author found Israeli mothers to be the most egalitarian, followed by their children and the fathers, respectively, who hold the most traditional gender role ideology. Kulik (2002a) argued that this is probably due to the fact that children adopt the attitudes of both mother (more egalitarian) and father (less egalitarian) and are therefore, situated in between both parents GRA.
Several U.S. studies and one Dutch and one Israeli study argue that respondents are generally positioned on the egalitarian-side of the axe (Crouter, Whiteman, McHale, & Osgood, 2007; Ex & Janssens, 1998; Kulik, 2002b; Nelson & Keith, 1990; Richmond-Abbott, 1984; Rollins & White, 1982). This is in contrast with some studies performed in Egypt, Saudi Arabia, Japan, and Korea that indicate that respondents are traditional minded (Kucinskas, 2010; Mensch, Ibrahim, Lee, & El-Gibaly, 2003; Sagara & Kang, 1998).
One finding that is recurrent in western and non-western countries is that male respondents are generally more traditional minded than their female counterparts (Burt & Scott, 2002; Carlson & Knoester, 2011; Crouter et al., 2007; Davis & Wills, 2010; Fan & Marini, 2000; Kucinskas, 2010; Kulik, 2000a; Marks, Bun, & McHale, 2009; Mensch et al., 2003; Nelson & Keith, 1990; Richmond-Abbott, 1984; Sagara & Kang, 1998; Stephan & Corder, 1985; Streitmatter, Santa-Cruz, & Ellis-Schwabe, 1984). Three longitudinal U.S. studies conclude that the difference between boys and girls narrows over time (Davis, 2007; Fan & Marini, 2000; Tallichet & Willits, 1986), although this is contested by another U.S. study of Galambos, Almeida, and Petersen (1990), who found boys becoming less egalitarian and girls more egalitarian over time. We should note that the studies have been performed in different time frames and comparing results of studies performed up to 30 years apart could be problematic (see Brooks & Bolzendahl, 2004). Also, we compared methodologies and study designs of the different longitudinal studies and found no clear-cut explanation for Galambos et al.’s divergent findings except for the lower age of the sampled youngsters (middle school vs. high school and older).
Race, nationality, and religion
Kiecolt and Acock (1988) found small effects of race on GRA in the 1970s and 1980s U.S. population, indicating that Black women hold more traditional views toward gender stereotypes compared with White women. Black men demonstrate less egalitarian views than White men, but only with regard to the support of women in politics. Fan and Marini (2000) found Black American respondents in the 1970s and 1980s to demonstrate more egalitarian GRA compared with White American and Hispanic American respondents, with the latter being most traditional. This divergence might be linked to the specific roles that have been inquired into. Although Fan and Marini (2000) focused solely on vocational roles (female employment), Kiecolt and Acock (1988) also questioned political roles and the support for gender stereotypes. Fan and Marini’s findings are in line with a review on racial and ethnic variations in GRA in adults’ GRA (Kane, 2000). Kane (2000) found African Americans to be particularly more egalitarian toward female employment. More recent U.S. studies found no relationship between race and GRA in youngsters (e.g., Carlson & Knoester, 2011; Davis, 2007; Davis & Wills, 2010). This supports the convergence-hypothesis of Kane (2000), arguing that racial and cultural differences in GRA in the United States might be declining over time.
Race was only included in U.S. studies. Sagara and Kang (1998) and Kucinskas (2010), in Japanese/Korean and Saudi Arabian/Egyptian adolescents, respectively, focused on nationality differences in GRA. These studies indicate that Korean youngsters express more traditional views than Japanese, and Saudi Arabian youngsters express more traditional views than Egyptian.
Religiosity was examined in five studies, and the concept was operationalized in an equally diverse way, ranging from religious service attendance (e.g., Davis, 2007; Davis & Wills, 2010; Fan & Marini, 2000; Kucinskas, 2010) to praying behavior (Kucinskas, 2010), religious affiliation (Davis, 2007; Davis & Wills, 2010; Fan & Marini, 2000), and self-reported religiosity (Kucinskas, 2010; Kulik, 2000b). Three U.S. studies (Davis, 2007; Davis & Wills, 2010; Fan & Marini, 2000) focus mainly on Christian respondents (both Protestant and Catholic) although they also include “other religions” and “not being religious.” These studies argue that religiosity is not of importance in the construction of youngsters’ GRA. Kucinskas (2010), however, demonstrated that the Islamic religion has a clear relationship with GRA in Saudi Arabia and Egypt. Praying 5 times a day was not significant in predicting gender egalitarianism, but self-identified religiosity was a significant positive predictor for female Egyptian youth’s egalitarian GRA. Religious orthodoxy and weekly mosque attendance, in contrast, are negative predictors for gender egalitarianism but only for male respondents, both in the Egyptian and Saudi Arabian sample. Kulik (2000b) also found, in Israel, that moderate or orthodox Jews express a more traditional gender ideology toward roles inside and outside the home. These results seem to indicate that self-identifying as religious is not per se related to traditional GRA, rather the religion, degree of religiosity, and religious activities performed seem of importance.
Situated in Home Environment
Thirty out of 35 studies included parental variables. Most of these studies found congruence in GRA between children and parents, indicating that offspring of parents with traditional views hold more traditional views themselves (Antill, Cunningham, & Cotton, 2003; Blee & Tickamyer, 1987; Carlson & Knoester, 2011; Crouter et al., 2007; Davis, 2007; Ex & Janssens, 1998; Meyer, 1980; Sagara & Kang, 1998). A few studies contest this parent–child congruence. Marks and colleagues (2009) found three family patterns: traditional congruence, egalitarian congruence, and “divergence” with more traditional parents and less traditional sibling (68 out of 358 families). The authors state, “The emergence of a divergent pattern suggests that, whereas social-learning processes explain children’s GRA development in some families, different mechanisms may be at work in other families” (Marks and colleagues, 2009, p. 12). Parent–child congruence is not definite, but depends on different aspects. Thus, a more complex process of intergenerational transmission of GRA is suggested. We will discuss the intervening aspects at play in intergenerational transmission in the following paragraphs.
Mother versus father
Different researchers show mixed results on the importance of maternal versus paternal influences. Antill and colleagues (2003) found that fathers’ GRA did not predict either boys’ or girls’ attitudes. Carlson and Knoester (2011), Rollins and White (1982), and Thornton and colleagues (1983) stated that youngsters show a greater similarity to their mother. Davis and Wills (2010), however, nuanced these findings and argued that parental gender interacts with socialization of GRA in the family. Davis and Wills (2010) explained their findings: “The most egalitarian adolescents are those with egalitarian fathers, and among those adolescents mothers’ ideology has no influence. In contrast, when fathers hold a more traditional gender ideology, mothers’ ideology has a strong, positive effect on adolescents’ gender ideology” (p. 507).
Same sex hypothesis
Social learning theory would suggest greater same-sex congruence due to same-sex modeling, but several studies do not support this same-sex hypothesis (e.g., Antill et al., 2003; Burt & Scott, 2002; Thornton et al., 1983) or found mixed results. Kulik (2002a), for example, found a same-sex effect for Israeli fathers and sons but not for mothers and their daughters. Sagara and Kang (1998) supported the same-sex hypothesis regarding Japanese students, but no support was found for Korean students.
It is suggested that same-sex congruence between parent and child interacts with the quality of parent–child relationships (Carlson & Knoester, 2011; Nelson & Keith, 1990). Carlson and Knoester (2011) identified two factors mediating parental socialization effects: relationship quality and age difference between parent and child. Carlson and Knoester (2011) argued that the smaller the age difference and the better the relationship quality, the higher the similarity in gender ideology between parent and child.
Family structure is another factor indicated as mediating the parent–child congruence. Carlson and Knoester (2011) found father–child congruence to differ according to family structure: children and their fathers show more similarity toward GRA in a two-parent biological or a single-parent family structure than in a stepfamily family structure. These results were not found for mother–child congruence.
Two studies that adopt very fine-grained operationalization of family structure are the studies of Carlson and Knoester (2011) and Kiecolt and Acock (1988). Carlson and Knoester (2011) distinguished between living in a two-parent biological family, biological single-parent family, or stepfamily. A double influence was found: Family structure influences parent’s gender ideology and also influences the attitudinal similarity between parent and child (as discussed above). Kiecolt and Acock (1988) identified four categories: (a) living with the mother only because of the death of the father, (b) living with the mother only because of divorce, (c) living the with mother and stepfather because of the death of the father, and (d) living with the mother and stepfather because of divorced parents. The results show family structure to be a direct and significant predictor for GRA, but dependent on the roles and family type that are examined. In researching the rejection of traditional gender roles, for example, no type of family structure was a significant predictor for GRA. However, boys living with a widowed mother have more egalitarian views toward working women. Likewise, boys’ and girls’ egalitarian views toward women in politics could be partly explained by the fact that they were living with their divorced mother and, for girls only, by living in a family with a mother and stepfather after parental divorce. Thus, family structure seems to be of importance for youngsters’ GRA through the type of roles being questioned.
Parents’ socioeconomic status (SES)
Most information included in the reviews on parental level were related to the SES, including variables such as education of mother (15 studies), father (9), or both parents (1); family income (4); maternal employment (16); paternal employment (3); and a few other variables associated with employment such as occupational prestige (1), duration of employment (1), and satisfaction with employment status (1).
First, most studies found a positive relationship between the offspring’s egalitarian views and both mothers’ education (Antill et al., 2003; Davis, 2007; Ex & Janssens, 1998; Fan & Marini, 2000; Kiecolt & Acock, 1988; Kulik, 2002a; Marks et al., 2009; Richmond-Abbott, 1984; Tallichet & Willits, 1986) and fathers’ education (Antill et al., 2003; Fan & Marini, 2000; Kulik, 2002a; Marks et al., 2009; Mensch et al., 2003; Richmond-Abbott, 1984; Thornton et al., 1983): Parents with higher educational levels have more egalitarian offspring. Ex and Janssens (1998) stated that the influence of maternal education on her child’s GRA acts through its influence on other variables such as parental GRA and employment. Not all studies concur that parental education influences youngsters’ GRA (Ahrens & O’Brien, 1996; Blee & Tickamyer, 1987; Mensch et al., 2003; Zulich, 1986). Antill and colleagues (2003) found that only opposite-sex parents’ education is a significant predictor for the offspring’s views.
Second, studies have investigated the effects of parental employment departing from the division of responsibilities inside and outside the home, which have been changing over time (Hansen, 1991; Pew Research Center, 2010). It is hypothesized that balancing paid work and household work became central (Hansen, 1991), resulting in predominantly a de-traditionalization of GRA (Valentova, 2012). Not surprisingly, these changes are reflected in empirical evidence: No less than 16 studies included maternal employment, demonstrating a positive relation with the offspring’s egalitarian GRA (Davis, 2007; Ex & Janssens, 1998; Fan & Marini, 2000; Kiecolt & Acock, 1988; Kulik, 2002a; Nelson & Keith, 1990). Different authors add that the influence of maternal employment differs depending on the timing of employment (more effect during childhood or adolescence; Blee & Tickamyer, 1987; Kiecolt & Acock, 1988; Thornton et al., 1983). An important nuance to these findings is made by Rollins and White (1982) who found that the mother’s employment effects differ according to the reason why mothers work. They found differences between traditional families (mother unemployed), dual-work families (mother employed for economic reasons), and dual-career families (mother choosing for a professional career). Children in dual-career families report more egalitarian sex role attitudes than traditional and dual-work families. The latter two do not significantly differ from each other. Thus, not the employment of the mother as such seems to be of importance, but the timing of the employment and the reasons behind the employment are. A few earlier studies did not find any positive relationship between maternal employment and the offspring’s egalitarian GRA (Meyer, 1980; Smith & Self, 1980; Tallichet & Willits, 1986; Zulich, 1986). These differences in findings between studies performed in the 1980s and the years after might be related to cohort replacement processes (Brooks & Bolzendahl, 2004) and different operationalizations of the concept (dual-work vs. dual-career families), as discussed by Rollins and White (1982).
Third, the results regarding family income influencing GRA were rather mixed. Blee and Tickamyer (1987) and Kucinskas (2010) found family income to be insignificant in predicting sex role attitudes (Blee & Tickamyer, 1987) and gender egalitarianism (Kucinskas, 2010). Erarslan and Rankin (2013), however, found family income to be a positive significant predictor for GRA toward work life and social life, but not for gender roles related to family life. Marks and colleagues (2009) found “traditional families” (parent and child traditional) and “divergent families” (parents traditional, child egalitarian) to be characterized by a lower family income. “Although the attitudes of parents in this group may have been grounded in socioeconomic factors, it appears that those of their children were not” (Marks et al., 2009, p. 12).
Situated in School Environment
Burt and Scott (2002) stated that although there is evidence for parent–child socialization of GRA, there is a greater similarity within generations (boys–girls) than between the two generations (parent–child). Kulik (2000b) confirmed this. The environment in which children interact on a daily basis with peers is the school environment. Therefore, we will discuss evidence on the relationship between GRA and the school environment.
Schooling: Liberalizing or traditionalizing?
Schooling was taken into account in 14 studies by means of including educational grade or number of completed school years. Twelve studies used age as a proxy for school progression. Only seven out of 12 studies included both youngsters’ age and schooling and thus were able to differentiate between the effect of both interacting predictors (Carlson & Knoester, 2011; Davis, 2007; Davis & Wills, 2010; Ex & Janssens, 1998; Kiecolt & Acock, 1988; Kucinskas, 2010; Streitmatter et al., 1984). Studies that only discuss age or school progression cannot exclude that the age effect is in fact a schooling effect and not the other way round. In the following paragraphs, we discuss the results of studies including both variables.
Three groups of research outcomes can be identified with regard to the effect of schooling.
The first group shows that pupils in lower school grades (fourth-fifth grade) in North Carolina (USA) display more egalitarian views when compared with higher grades (ninth grade; Streitmatter et al., 1984). The mean scores by grade show a consistent and gradual increase to more traditional perceptions on sex roles.
The second group indicates more egalitarian views in students who have received more schooling. Carlson and Knoester (2011), for example, found in a national U.S. study on 18- to 23-year-olds that compared with students who droped-out, students attending high school, having a high school degree, or a degree higher than high school have more egalitarian views. In researching this, they controlled, for example, parental gender ideology and students’ sex and age. Davis (2007) found greater levels of education to be correlated with more egalitarian ideologies but only from college-age on (age 18 and older). Ex and Janssen’s (1998) study in Nijmegen (the Netherlands) found a main effect of education on girls’ attitudes about motherhood: Those students in lower education (secondary education) were more traditional than those in university. Moreover, an interaction effect between education and age was found, indicating younger respondents (15-17) were more traditional than older respondents (20-22), and this is especially true for those younger respondents with lower education. In a study of Kiecolt and Acock (1988, USA), the years of education significantly predict non-traditional GRA for both male and female respondents aged 18 and older. This was controlled for age and confirmed for all three subscales (Women in Politics, Married Women Working, and Rejecting Traditional Gender Roles). Kucinskas (respondents’ age = 18-25) found that having at least some university education was predicting gender egalitarianism for Egyptian men, but for neither Egyptian women nor Saudi Arabian respondents.
The third group found educational attainment to be no significant predictor for GRA in 14- to 15-year-old students within the United States (Davis & Wills, 2010).
Thus, we can conclude that studies do not fully agree on the effect of education. The results seem to indicate that schooling generally is positively related with egalitarian GRA (e.g., Carlson & Knoester, 2011), especially higher educational studies (e.g., Davis, 2007; Ex & Janssens, 1998). The findings of Kucinskas (2010), however, suggested the relationship between schooling and gender egalitarianism to vary among groups. Mensch and colleagues (2003) added to this that education as such does not enhance egalitarian views but the effects are conditional and depend on the present views within schools.
With the exception of the continuation of schooling, only few school-related variables were detected within the 35 included articles (Chatterjee & McCarrey, 1989; Erarslan & Rankin, 2013; Erden, 2009; Streitmatter et al., 1984; Zulich, 1986). Furthermore, most studies including school-related variables focused on a limited number of variables. Therefore, most variables are only discussed in one study, which makes it impossible to compare results of the different studies.
The work of Erarslan and Rankin (2013) in Istanbul showed that going to a school located in a high-SES neighborhood is a positive predictor for female students’ egalitarian attitudes toward social life and family life. Attending a single-sex school is only predictive for girls’ GRA toward family life: Girls going to a single-sex school have more egalitarian views toward family life compared with girls in co-education. Both predictors were not significant for the students’ attitudes toward work life and balancing work and home responsibilities.
Erden (2009), the only experimental study, found that a course on gender equity within further education (Ankara, Turkey) had a significant influence on GRA. Although the attitudes of the students in the same program who did not follow the gender equity course remain the same over time, the students of the experimental group report more egalitarian GRA after following the semester-long gender equity course compared with the beginning of the course. However, it should be noted that students were not randomly assigned to experiment or control groups and, therefore, selection bias is probably present. Also, only short-term effects are examined. Nonetheless, this result is promising as it demonstrates the possibility to alter students’ GRA.
Relational school aspects
Regarding teachers’ influence on students’ GRA, Zulich (1986) found that ninth- to 12th-grade students’ current class teacher’s sex role attitudes significantly and positively predict students’ sex role attitudes, indicating students’ and current teacher attitudes to be in line. This was not the case for the GRA of other teachers identified as role models by students themselves.
The GRA of peers as perceived by their fellow students were found to be of importance. Chatterjee and McCarrey (1989) argued that female undergraduate students in more traditional female training programs (cf. nursing) have more traditional inferred sex role attitudes than their peers—that is, they perceive their peers as having traditional views—and are more traditional themselves compared with students in non-traditional programs (cf. trades, technology). The authors stress the congruence between the perception of peers’ attitudes and the individuals’ own attitudes. Although not focusing on the school environment, Richmond-Abbott (1984) stated that the respondents reported peers to be of most influence on their views on sex roles.
Most variables related to the individual and home environment were structural aspects such as SES and family structure. Relational aspects like time spent with a child, parenting style, family conflict, parental closeness with a child, relationship quality between parent and child, and father’s involvement in child rearing were studied far less.
This focus on structural aspects also applies to the school-related variables: most studies study educational program, school structure, co- versus single-sex education, and SES of the school’s neighborhood. Only a few relational aspects were taken into account such as perceived peer attitudes (Chatterjee & McCarrey, 1989) and the teacher–student relationship with the current teacher and role model teacher (Zulich, 1986), but these studies are limited and out-dated.
The limited studies including relational factors confirm that relational aspects are important in studying GRA. Relationship quality with parents, for example, was identified as an important determinant in researching parent–child congruence. Future research needs to focus on relational aspects and its interaction with structural predictors.
As society has been changing, traditional family patterns have given way to different family arrangements (Eurostat, 2010; Pew Research Center, 2010). Changing family compositions are furthermore associated with changing gender roles and GRA (Pew Research Center, 2010). Nonetheless, most research focuses on the so-called nuclear, two-parent families: only five studies focus on diverse family compositions. Furthermore, only three studies included family structure in their research and family structure has been operationalized in different ways. We can conclude that although little attention was given to the diversity of family arrangements, family structure has been identified as an important variable, both directly and indirectly influencing the youngsters’ GRA.
Boys and Fathers
While 11 studies focused solely on girls, none of the reviewed studies focused solely on boys. Furthermore, it should be noted that only 11 out of 24 studies performed separate analysis for boys and girls. The other 13 studies performed analysis on all youngsters together, which limits comparison between boys and girls. The same can be said for parents: While data concerning mothers were widely available (29), data on fathers were less forthcoming (19). The lack of male respondents was also discussed by different researchers as a limitation and recommendation for future research (e.g., Erden, 2009; Ex & Janssens, 1998). Our review shows that we cannot simply assume that the same explanatory mechanisms are in play for boys and girls alike. Differentiating between both can provide detailed information on the differential construction of GRA in boys and girls.
Boys, in this regard, have been under-addressed in previous research. Boys express—in general—more traditional GRA than girls, and traditional GRA are, in turn, related to negative social and scholastic outcomes (Huyge et al., 2015). This places boys at greater risk for academic underachievement and understanding GRA construction in boys could help increase boys’ sense of school belonging (Huyge et al., 2015) and reduce their engagement in violent and criminal behavior (e.g., Ben-David & Schneider, 2005; Flood & Pease, 2009).
With most studies focusing on parent to child transmission of GRA, a rather deterministic view has been preserved, neglecting the dynamic nature of GRA. We identify two dimensions in this deterministic view: (a) a content dimension (rigid focus on parent–child transmission of GRA not taking into account other socializing agents) and (b) a methodological dimension (mostly applying cross-sectional methods not taking into account changes over time). The deterministic view contrasts strongly with life-course perspective and social-cognitive theory, both stressing the continuous development of GRA and taking into account one’s own life experiences and situational aspects in the construction of GRA. Other socializing agents to which boys and girls are exposed to (e.g., Galambos, 2009)—both in the home environment (e.g., siblings) and other environments like school (e.g., peers and teachers)—were under-studied. However, the need to take into account these agents has been acknowledged by different researchers in discussing their results (Burt & Scott, 2002; Carlson & Knoester, 2011; Galambos et al., 1990; Kiecolt & Acock, 1988; Nelson & Keith, 1990; Smith & Self, 1980). Davis (2007) concluded clearly with the following statement:
The findings offer little support that adolescent social and background characteristics are deterministic of young adult gender ideology. After initial effect in adolescence, background characteristics bear little relationship with ideology trajectories. . . . What is more important over time is the lived experience of young adults in the construction and reconstruction of gender ideology. (p. 1035)
This is supported by research of Crouter and colleagues (2007), Fan and Marini (2000), and Tallichet and Willits (1986), who all found that individual variation in GRA can be attributed to variation in respondents’ own life experiences. The importance of exposure to life experiences and situational aspects was already established in research on adult GRA (e.g., Bolzendahl & Myers, 2004; Vespa, 2009). Likewise, further research into the dynamics of youngsters’ GRA needs to step away from a deterministic point of view and take into account multiple socializing agents and the individuals’ own life experiences.
Another way in which the life-course perspective is not fully met is the largely cross-sectional nature of most articles (25 out of 35 articles). The studies are not able to do causality statements and are limited to description and prediction of GRA at one time point rather than providing us with robust insights into evolutions in GRA. Studies are mostly limited to correlational analysis, comparing groups and predicting GRA. Different researchers identify this as a weakness and advise longitudinal future research (Ahrens & O’Brien, 1996; Burt & Scott, 2002; Davis & Wills, 2010; Erarslan & Rankin, 2013; Marks et al., 2009; Nelson & Keith, 1990; Zulich, 1986). The results of the few longitudinal studies focusing on these developmental aspects and change of GRA over time support the above described importance of the individuals’ own life experiences (Crouter et al., 2007; Fan & Marini, 2000; Tallichet & Willits, 1986) and schooling (Davis, 2007; Fan & Marini, 2000; Tallichet & Willits, 1986; Williams, Radin, & Allegro, 1992). Therefore, the use of longitudinal data in analyzing GRA is recommended.
An Intersectional-Informed Approach
When considering the researched samples, we notice rather homogeneous groups. As argued earlier, most of the studies focus on female respondents and rather traditional family compositions. Besides this, different researchers also note that their sample is rather homogeneous with regard to ethnicity and SES focusing mostly on White, middle-class families (Ahrens & O’Brien, 1996; Crouter et al., 2007; Galambos et al., 1990; Marks et al., 2009; Meyer, 1980; Williams et al., 1992). These findings are in line with the USA-review study performed by Kane (2000) on GRA in adults, indicating racial and ethnic variation in GRA received poor attention.
Different studies make a plea for moving away from a universal study approach to the construction of GRA, underlining the diversity in attitudes and variables explaining GRA and promoting research into this diversity (Crouter et al., 2007; Fan & Marini, 2000; Kucinskas, 2010; Nelson & Keith, 1990; Sagara & Kang, 1998).
Also, when diversity is taken into account, a rather one-dimensional approach is used, placing one group opposed to another on one diversity axe such as sex (women/men; Antill et al., 2003; Burt & Scott, 2002) or race (Blee & Tickamyer, 1987). Due to the situatedness of GRA in social, cultural, and historical contexts (Bussey & Bandura, 1999; Elder, 1998), it seems valuable to apply an intersectional-informed approach (Collins, 1998; Crenshaw, 1991) when studying GRA. The intersectional-informed approach acknowledges the human being as irreducible to a single characteristic, but as a nest of multiple, interacting characteristics (Hankivsky, 2012, 2014). Therefore, the intersectional-informed approach opts for research at the intersects of multiple diversity dimensions such as sex, ethnicity, culture, age, geography, and class (Hankivsky, 2014) as opposed to examining GRA on each distinctive axe. Some research has already focused on intersections of diversity axes (e.g., Sagara & Kang, 1998), comparing GRA predictors of four different groups based on both nationality and sex: Japanese boys, Japanese girls, Korean boys, and Korean girls. In line with Vespa (2009) and Kane (2000), we argue for an intersectionality-informed approach in researching GRA, accounting for heterogeneity both between and within groups.
Inconsistency in Measuring GRA
We have noticed that although all studies measure attitudes toward appropriate roles for men and women, many different scales are used to grasp this concept. We noticed that many researchers develop their own scale (e.g., Erarslan & Rankin, 2013; Erden, 2009; Kulik, 2000a; Tallichet & Willits, 1986). These scales differ, for example, in the amount of items, the target group of the scale (e.g., whether the scale was developed specifically for youngsters), and the subroles being interrogated (e.g., marital and familial roles, vocational roles, societal roles). Beere (1990) already advised in the early 90s that greater consistency in measuring GRA should be achieved to improve study comparison. We would like to put this back on the agenda, even though greater consistency should not be at the cost of instrument development and modification to target specific groups and contexts.
This study focused solely on youngsters’ GRA. During the search, we used a somewhat artificial demarcation of child and emerging adulthood to distinguish youngsters from adults. We acknowledge that the transition from emerging adulthood into adulthood does not take place on a fixed age and therefore can differ between individuals. Furthermore, we would like to stress that also in adulthood, GRA is not fixed and changes in GRA occur (e.g., Vespa, 2009). This dynamic nature of GRA throughout the entire life span is emphasized in the life-course approach, and it is confirmed for youngsters in the results of our systematic review.
Second, the search was limited to texts fully available in English. This explains—at least in part—the underrepresentation of studies in non-Western countries. To better investigate the effects of cultural differences in the construction and predictors of GRA, a review of literature in other languages may be needed.
This article applied a systematic approach in bringing together articles that focus on the construction of youngsters’ GRA within different countries. This review confirms the complex relationship between individual, home, and school characteristics and school-aged youngsters’ GRA. Moreover, this article provides an overview of the current state of the art and denotes currently under-researched topics such as the quality of parent–child relationships, non-nuclear family structures, boys and fathers, and changes in GRA over time. Also, it identifies promising approaches to researching youngsters’ GRA, such as an intersectionality-informed approach and a life-course perspective. This review will hopefully be able to guide future research on youngsters’ GRA.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Research Foundation Flanders (FWO; Grant FWOTM773).
- © The Author(s) 2016
This article is distributed under the terms of the Creative Commons Attribution 3.0 License (http://www.creativecommons.org/licenses/by/3.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access page (https://us.sagepub.com/en-us/nam/open-access-at-sage).
Myriam Halimi is a PhD fellow of the Research Foundation Flanders (FWO) employed at the educational sciences department, Vrije Universiteit Brussel. Myriam is a member of the Centre of Expertise on Gender, Diversity, and Intersectionality (RHEA).
Els Consuegra is a post-doctoral researcher in the field of gender and educational sciences at the Vrije Universiteit Brussel and a member of the RHEA. Els is a guest professor at the Department of Educational Studies at the Universiteit Gent.
Katrien Struyven is assistant professor at the Vrije Universiteit Brussel. Working in Brussels, the multicultural capital of Europe, diversity as a theme has gained importance in her research on teaching and assessment.
Nadine Engels is vice-dean of the Faculty of Psychology and Education at the Vrije Universiteit Brussel and program director of teacher education. Her main research expertise is in the fields of professional development and diversity in education.
References marked with an asterisk indicate studies included in the systematic review.