SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 1
Support for Gender Equality among Early Adolescents – the Role of School Characteristics for Gender Attitudes
Benthe van Wanrooij 11355875 Master Educational Sciences University of Amsterdam Supervisor: Hester Mennes First reviewer: dr. Anke Munniksma Second reviewer: dr. Remmert Daas Date: 7 July 2021 Word count: 10419
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 2
Abstract
The aim of this study was to analyse which student and school characteristics have an
influence on the support of gender equality among early adolescents. Research on the factors that play a role for support for gender equality has thus far focussed on sociodemographic factors, whereas school characteristics have been neglected. The research question therefore is: To what extent are student and school characteristics related to students’ support for gender equality? This study considers both student demographic factors as well as school characteristics: gender, parental education level, maternal employment status, perception of open classroom climate and teacher-student relationships, as well as classroom averages of these factors to look for classroom compositional effects. Furthermore, I studied whether an open classroom climate and teacher-student relationships can play a positive role in the differences in support for gender equality between boys and girls. Based on data from the International Civic and Citizenship Education from 2016 in five countries (Belgium, Denmark, Finland, the Netherlands and Sweden), I performed linear multilevel regression analysis to see which of these factors had a relation with support for gender equality. Girls show higher levels of support for gender equality. Furthermore, I found a positive significant relationship between support for gender equality and parental education level, maternal employment status, perception of open classroom climate and teacher-student relationships.
Classroom compositional effects are limited to classroom average parental educational level and maternal employment status, but there is no significant relationship between the
classroom average open classroom climate perception or teacher-student relationships. The perception of a positive teacher-student relationship plays a role in reducing differences between boys and girls. An open classroom climate does not take on that role.
Keywords: support for gender equality, gender, adolescents, sociodemographic factors, school characteristics, classroom composition, ICCS
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 3
Contents
Abstract ... 2
Introduction ... 4
Theoretical Framework ... 8
Methods ... 18
Results ... 26
Conclusions ... 34
References ... 41
Appendix A ... 53
Appendix B ... 54
Appendix C ... 61
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 4
Support for gender equality among early adolescents – the role of school characteristics for gender attitudes
Introduction
Gender inequality is visible and dominant in society. Worldwide, women are
underrepresented in politics (Inter-Parliamentary Union, 2020) and upper level management (Lewellyn & Muller-Kahle, 2020). In Western countries, women in heterosexual relations take on the largest share of work in the household, despite working the same amount of hours outside of the house as men do (Sullivan, 2018), which can have severe negative impact on women’s health (Erarslan & Rankin, 2013). In most European countries, the options for paternity leave are fairly limited and unequal compared to maternity leave policy (Kane, 2018). These inequalities between men and women are persistent, and research on the causes of these inequalities attributes a significant part to traditional attitudes that people and
societies hold about gender roles (Farré & Vella, 2013; Inglehart & Norris, 2003).
Traditional attitudes towards gender reflect negative and unequal perceptions of the role of women in society, whereas egalitarian attitudes perceive women and men to be equal and to have equal rights in society (Inglehart & Norris, 2003). The support for gender equality develops from an early age through socialisation, in which especially the early adolescence is influential in the formation of gender attitudes (Leaper & Farkas, 2015) and after which the support for gender equality for people solidifies (Kågesten et al., 2016). The development of gender attitudes has a great impact on the future of gender equality: the more egalitarian these attitudes are, the more gender equal society could become (Dotti Sani &
Quaranta, 2017). Given that gender equality persists in society, and early adolescence is a crucial period for the development of gender attitudes, we need insight in which factors affect support for gender equality during adolescence.
Research about factors that influence the support of gender equality among early
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 5
adolescents thus far focusses on individual background characteristics (Kågesten et al., 2016).
Individual characteristics such as biological sex (Crouter et al., 2007), resources at home and interpersonal interactions with parents and peers are influential. Support for gender equality is related to parents’ educational level, level of religiosity and immigration background (Antill et al., 2003; Crouter et al., 2007; Dotti Sani & Quaranta, 2017; Sánchez Guerrero &
Schober, 2020), as well as through parents’ household and parenting practices (Antill et al., 2003; Burt & Scott, 2002). Next to parents, peers are seen as influential (Crouter et al., 2007;
Halimi et al., 2020; Hibbard & Buhrmester, 1998), and especially adolescents are heavily influenced by peers and group dynamics (Harris, 1995). Overall, individual characteristics have shown to be of impact on whether and to what extent adolescents support gender equality and have previously been heavily researched.
However, recent meta-analyses point out (Halimi et al., 2016; Kågesten et al., 2016) that research on one critical agent in gender socialisation is scarce: the school. Thus far, a limited numbers of school-related aspects have been investigated, such as co- versus single- sex education (Halimi et al., 2016), teacher behaviour and attitudes (Treviño et al., 2017) or school gender culture (Halimi et al., 2020). An open classroom climate (Barber, Sweetwood, et al., 2015; Barber, Torney-Purta, et al., 2015; Claes et al., 2017; Treviño et al., 2017) and teacher-student relationships (Sampermans & Claes, 2018; Wanders et al., 2020) as part of educational experiences for adolescents have been found to impact civic attitudes in general, but research on the relation between educational experiences and support of gender equality is limited (with the exception of Sandoval-Hernández et al., 2018). Furthermore, studies considering the classroom composition in relation to support for gender equality are rare.
Schulz et al. (2018) found classroom composition to matter in relation to civic attitudes generally.
Overall, empirical quantitative studies on the composition of the school and
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 6
educational experiences, and its relation with support for gender equality, have not yet been the focus of research. Yet, as the school is considered one of the important institutions of socialisation among students (Persson, 2015), it does urges us to look into these factors in the school. This thesis therefore adds to the existing literature by looking at the diverse factors that influence gender attitudes among early adolescents, with a particular focus on school characteristics.
The goal of this thesis is to answer which student and school characteristics are related to students’ support for gender equality, and whether there is an interrelation between both student and school characteristics with regards to gender and educational experiences. This question will be answered by looking at gender attitudes of early adolescents in five
European countries (Belgium, Denmark, Finland, the Netherlands and Sweden), measured in the International Civic and Citizenship Education Study 2016 (Schulz et al., 2018). The research question that follows from this:
To what extent are student and school characteristics related to students’ support for gender equality?
Lately, there has been growing attention on the differences among boys and girls in educational experiences, such differences in academic culture between boys and girls (Van Houtte, 2004), political knowledge (Pereira et al., 2015) or experiences regarding safety in schools (IEA, 2019). Considering that girls in general express more positive attitudes towards gender equality than boys (Halimi, Consuegra, et al., 2020; Schulz et al., 2018), I wonder whether educational experiences have a different effect on the support for gender equality among boys and girls. That leads me to the following sub questions:
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 7
To what extent do boys and girls differ in their support for gender equality?
To what extent does an open classroom climate compensate for the differences in the support for gender equality among boys and girls?
In the following section (Chapter 2), I start by introducing theories and previous research on factors that explain the support for gender equality among adolescents. Following from these previous studies, I formulate several hypotheses which are tested in this thesis.
Chapter 3 provides insights into the methods used to perform the analysis. Chapter 4 shows the results of the multilevel analysis. In Chapter 5, I reach a conclusion, discuss limitations to this study, and give recommendations for future research.
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 8
Theoretical Framework
Gender equality centres around the idea that the position of both men and women in society is equal. Support for gender equality is thus the individual’s perception on whether this equality is to be present in society. Support for gender equality is embedded in
individuals’ attitudes towards gender and gender norms, which can be either more traditional, reflecting more negative perceptions towards gender equality, or more egalitarian, reflecting support for gender equality (Inglehart & Norris, 2003). Various definitions are used with regards to gender attitudes and support for equality (Constantin & Voicu, 2015); however, most authors tend to focus on the role of men and women in society at large (i.e. Bryant, 2003; Halimi, Consuegra, et al., 2020; Lundgren et al., 2013). Are both men and women able to perform similar tasks or jobs? Should men and women divide household tasks evenly?
What is the role of men and women in politics? These are some central questions in definitions of support for gender equality.
Gender socialisation1 is the ‘process through which individuals learn the gender norms of society’, develop their own identity and values with regards to gender (Ryle, 2011, p. 120).
In this theoretical framework, I introduce how support for gender equality is developed among early adolescents and which factors have previously been found to be influential for this socialisation process. This theoretical framework has been built from an ecological perspective (Bronfenbrenner, 1992): we cannot study the support for gender equality without taking into account the context or environment of the individual, such as someone’s family or school. There are a variety of actors and processes that have an influence on the formation of gender attitudes – socialisation is ‘multidimensional and complex’ (Stockard, 2006, p. 223).
1 Gender is a social construct. Ever since Bem's (1993) The Lenses of Gender, sociologists of gender have moved away from thinking of gender as constructed by biological differences between sexes (Risman et al., 2018).
Rather, gender is seen as something that is constructed by society, individuals and among peers – one ‘does gender’ through interactions (Risman et al., 2018).
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 9
Boys and girls differ in their support for gender equality
The level of support for gender equality differs between boys and girls: girls are generally more supportive of gender equality than boys (i.e. Antill et al., 2003; Burt & Scott, 2002; Crouter et al., 2007; Farré & Vella, 2013; Sampermans & Claes, 2018; Sandoval- Hernández et al., 2018). There are different explanations as to why gender equality support is higher among girls than boys. From an early age on, children are exposed to gender-
stereotyped factors that might differ according to biological sex2, such as the toys they play with (Boe & Woods, 2018) or the friends they have (Hibbard & Buhrmester, 1998). These socially created differences might influence the development of gender attitudes among children (Crouter et al., 2007). Interest-based explanations suggest that the benefits of support for gender equality are higher for girls than boys, for instance as girls are more likely to get higher pay or decrease their household labour tasks when society becomes more gender equal. Therefore, girls might show more support for gender equality than boys (Bolzendahl &
Myers, 2004). Meta-analyses (Halimi, Consuegra, et al., 2020; Kågesten et al., 2016) show that the differences for boys and girls in support for gender equality are consistent across studies. Therefore, I expect that the support for gender equality is higher for girls than for boys (Hypothesis 1).
Primary influencers in the support for gender equality: parents
In the development of support for gender equality, family members and especially parents3 are considered primary influencers (Glass et al., 1986; Marks et al., 2009).
According to socialisation theories (Leaper & Farkas, 2015) such as social learning theory (Bandura, 1971) and gender schema theory (Bem, 1981), the parents are an influential actor
2 Previous research looked at biological sex, which is why I chose this term. Nevertheless, I want to acknowledge and point out that gender is something one does, and that therefore the biological sex does not always comply with the gender someone aligns with, nor do I agree with the binarity of the biological sex as portrayed here.
3 There are more family structures than heterosexual “father-mother” couples. Considering the scope of this thesis, and the traditional focus of research on traditional family structures, I have decided to only pay attention to heterosexual, two-parent family structure. I elaborate on this limitation in the discussion section of this thesis.
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 10
in the development of support for gender equality among children through behaviour.
Children see other people perform gendered behaviours and will respond to this by modelling this behaviour (Stockard, 2006). Empirical research has found a plethora of support for these theories. First of all, gender attitudes of children and parents are often alike, suggesting intergenerational transmission of support for gender equality (i.e. Carlson & Knoester, 2011;
Crouter et al., 2007; Cunningham, 2001; Kretschmer, 2018; Marks et al., 2009; Tenenbaum
& Leaper, 2002; Updegraff et al., 2014). Studies looking at parents’ behaviour, such as their gender-stereotyped household and parenting practices (Antill et al., 2003; Dawson et al., 2016; Marks et al., 2009) portray that these factors influence their support for gender
equality. Parents’ background with regards to educational level and/or employment is another indicator. Parents’ ideas regarding support for gender equality become ‘enlightened’ through education, as they are exposed to more egalitarian ideas with regards to gender (Bolzendahl
& Myers, 2004). Not all studies agree on whether it is the mothers’ (Bolzendahl & Myers, 2004; Davis & Greenstein, 2009; Dotti Sani & Quaranta, 2017) or fathers’ educational level (Judge & Livingston, 2008; Marks et al., 2009) that matters – or both (Antill et al., 2003), yet all studies indicate a positive relation – the higher the educational level, the more support for gender equality.
In addition to the educational attainment level of parents, employment status of parents is considered as well. Over the past decades, the worldwide labour market
participation of women has steadily increased (Parcheta et al., 2013), as a result of which more mothers are ‘working mothers’. Women who are part of the workforce tend to be more supportive of gender equality, and tend to move away from traditional views on household labour or other ‘domestic’ issues (Bolzendahl & Myers, 2004). Nowadays, more children grow up in households where the mother is employed, and are modelled a more equal household organisation (Davis & Greenstein, 2009). Longitudinal research into gender
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 11
attitudes shows that adolescents who grow up with a working mother show more support for gender equality over time compared to peers whose mother did not have a paid job
(Gesthuizen et al., 2013). Recent research by Dotti Sani and Quaranta (2017) found an association with maternal employment and gender attitudes for girls only, yet Sánchez Guerrero & Schober (2020) found support for both boys and girls. Overall, the maternal employment status seems to be related to the development of support of gender equality.
However, the effect of maternal employment could be related to their educational level (Farré
& Vella, 2013), as there is quite some overlap between employment status of women and educational attainment (Gehringer & Klasen, 2015). For example, Thijs et al. (2019) found that employment in itself did not account for more support for gender equality in the Netherlands, as it was related to a higher educational level of those women. Both the workplace and education exposes people towards both enlightened egalitarian ideas and experiences (Bolzendahl & Myers, 2004). All in all, it remains unclear whether it is the parents’ educational level or the mother’s employment status – or both - that is the actual influential factor. However, as the effect through which education level and employment status seem to relate to support for gender equality differ, it is relevant to look at both parents’ education level and mothers’ employment status. I therefore hypothesise that the higher the parental educational level is for early adolescents (Hypothesis 2a) and the higher the maternal employment status (Hypothesis 2b), the higher adolescents’ support for gender equality is.
Class composition effects – do they matter?
Previous studies on civic attitudes and knowledge have found that, next to parental factors at the individual level, parental resources also matter at the classroom or school level.
The aggregation of the individual socioeconomic status of students to the school level showed a relation with civic attitudes (Schulz et al., 2018). These class composition effects
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 12
are explained by the socialisation process that takes place among peers in a classroom, as well as by differences between schools with a lower or higher socioeconomic status. First of all, peer interactions influence the development of support for gender equality (Leaper &
Farkas, 2015). Children might be under pressure from peers to conform to the norms of a group (Halimi, Davis, et al., 2020). As a result of this peer pressure, children can start to form similar norms (Harris, 1995; Sánchez Guerrero & Schober, 2020). When students are in a school with a more negative gender culture, their support for gender equality is lower (Halimi, Consuegra, et al., 2020). Furthermore, peers could – just like parents – act as role models whose behaviour children copy or follow (Leaper & Farkas, 2015). Peer conformity is increasingly important in early adolescence, as this is the age in which adolescents start to spend more time with peers than with their parents, and peer relations become closer and more important (J. Grusec & Hastings, 2015). Overall, socialisation of support for gender equality can take place among peers, and the socioeconomic status of the class as a whole might have a relationship with the support for gender equality among individuals. However, the socioeconomic status of a school or classroom in itself might matter. Perry (2013), as in her literature review, found that teachers have different expectations for classes of students with a lower socioeconomic status, and therefore might teach them differently. Schools with a higher socioeconomic status often also have more resources to improve their teaching and school climate (Perry, 2013). Therefore, the socioeconomic composition of the school is a predictor for many student outcome variables, such academic achievement (Perry &
McConney, 2010; van Ewijk & Sleegers, 2010), language proficiency (Van Der Slik et al., 2006) or motivation (Hornstra et al., 2015).
Students from different socioeconomic backgrounds tend to be distributed unevenly across schools, as a result of which socioeconomic segregation is an issue in many European countries (Gutiérrez et al., 2020): some schools have a higher average socioeconomic status
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 13
than others. Students whose parents have a higher educational level or whose mother’s employment status is higher, are likely to be in the same school - these two indicators are often part of the measurement for socioeconomic status and/or composition. Sandoval- Hernandez and colleagues (2018) studied the impact of socioeconomic composition of the school on gender attitudes. They found the socioeconomic composition of the school to be associated with gender attitudes in 22 out of the 37 countries they measured, with a higher socioeconomic composition of the school indicating more support for gender equality. I predict the parental educational level and/or mother’s employment status to matter at the individual level, as indicators of the socioeconomic status of a student. Considering socioeconomic compositional factors as well, I expect that there is a relation between the classroom average level of maternal employment and/or parents’ educational level and support for gender equality. The support for gender equality is higher in classrooms in which on average students have a higher maternal employment status (Hypothesis 3a), or whose average parental educational level is higher (Hypothesis 3b).
What impact does an open and supportive school environment have?
More attention has been put on schools, suggesting they play a crucial role in the development of civic attitudes and knowledge (Maurissen et al., 2018). Schools function as mini societies in which students get together and it is seen as their first experience with how democracy works (Torney-Purta et al., 2007). An element of the school, which is perceived to be crucial in this experience, is an open classroom climate. Students are better able to develop their civic knowledge and attitudes when they experience an open classroom climate
(Campbell, 2008). In a classroom with an open climate, debate on political issues is fostered, and ‘how [..] students evaluate the discussion opportunities during classes’ (Claes et al., 2017, p. 224) is relevant to the development of civic attitudes. This deliberative process gives students the room to develop their own preferences (Habermas, 1996). Moreover, it helps
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 14
them develop and improve their civic knowledge and skills (Alivernini & Manganelli, 2011;
Hooghe & Dassonneville, 2012; Persson, 2015), and it has been shown to increase their political and social engagement (Campbell, 2008) and political trust (Dassonneville et al., 2012). Deliberation and open classroom climate are furthermore perceived to create tolerance for other groups (Enslin et al., 2001), one of the central values of democracy. Overall, a deliberative classroom is perceived to promote democratic and civic values and attitudes, specifically on tolerance. Previous studies on the effect of an open classroom climate on the support of gender equality found a relation with an open classroom climate (Barber, Torney- Purta, et al., 2015; Sandoval-Hernández et al., 2018). I therefore expect that when students perceive a more open classroom climate, their support for gender equality will be higher (Hypothesis 4a). Furthermore, the perception of an open classroom climate at the school level could have an impact as well. An aggregated score looks at the ‘contextual classroom effect rather than the respondents’ own perceptions of the climate’ (Persson, 2015, p. 589): it is not just the aggregate of the individuals’ score, but represents the context of the classroom climate (Ichilov, 2007). There is some discussion on whether the perception of an open classroom climate at the school or classroom correlates to civic attitudes, yet so far, this discussion has been mostly methodological in nature (Barber, Sweetwood, et al., 2015), and therefore I hypothesise that a higher average classroom climate perception will relate to higher support for gender equality (Hypothesis 4b).
In addition to open classroom climate, teachers matter. Just like peers and parents, they function as role models for their students (i.e. Paredes, 2014), through which they might influence students’ support for gender equality. The way teachers interact with students from a certain group, such as girls or ethnic minorities, could influence the way students perceive this group and may influence their support towards equality (Diazgranados & Sandoval- Hernandez, 2020). Not only are teachers role models, teachers play an important role in
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 15
creating a safe and comfortable environment for their students (Hamre & Pianta, 2001).
Having a positive relationship with your teacher makes a student feel secure, which creates room for students to have discussion on civic aspects and bring them into contact with other ideas (Wanders et al., 2020). A previous study on support for ethnic minorities in twenty-two European countries found a strong effect between teacher-student relationships and equality attitudes (Diazgranados & Sandoval-Hernandez, 2020). So far, the only empirical study which considered support for gender equality, Sampermans and Claes (2018) found, using data from seven European countries, that positive teacher-student relationships are related to more support for gender equality. However, studies on civic attitudes and knowledge in general has shown that a positive teacher-student relationship matters (Isac et al., 2014;
Wanders et al., 2020). As social justice theory expects - and previous empirical studies show - teachers to transfer democratic attitudes of equality, I hypothesize that students who
experience positive teacher-student relationships show more support for gender equality (Hypothesis 5a). I also take into account the effect of the average perception of the teacher- student relationships in a classroom, as ‘being part of a group which evaluates this relation as good might have an effect on your personal experience in classroom discussions’ (Maurissen et al., 2018, p. 9). Therefore, I hypothesise that an on average better teacher-student
relationship in the classroom will result to more support for gender equality for the student (Hypothesis 5b).
Compensation or acceleration for differences in support for gender equality between boys and girls
As we have reasons to expect that girls are generally more supportive of gender equality than boys, the question rises whether educational experiences are able to compensate for these differences, or that they will have an even further accelerating effect. Acceleration theory suggests that students whose background, such as their gender, is already advantaged
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 16
for the development of civic knowledge and attitudes benefit even more from the civic opportunities they experience in a school setting, whereas disadvantaged students are unable to benefit from the educational experiences such as an open classroom climate, thus seeing the differences between these group increase (Kahne & Middaugh, 2009). Compensation theory claims the opposite: educational experiences could help compensate for these differences between boys and girls, improving civic attitudes and knowledge more for students whose civic learning previously was disadvantaged due to their home environment (Neundorf et al., 2016; Sampermans & Claes, 2018). Studies with regards to different civic attitudes, knowledge or skills show different results, most find a compensating effect
(Campbell, 2008; Neundorf et al., 2016; Sampermans & Claes, 2018), whereas some find an accelerating effect (Kahne & Middaugh, 2009). Maurissen et al. (2018) studied 22 European countries on to what extent the school context is an important contributor to open classroom climate. They found that girls perceived a more open classroom climate, yet classroom climate seemed to reduce the differences between boys’ and girls’ civic attitudes, whereas teacher-student relationships did not show a relation. As most empirical evidence has been found for a compensational effect, I expect that the relation between educational experiences and the support for gender equality is stronger for boys than for girls (Hypothesis 6), with the indicators for educational experiences being the individual perception of an open classroom climate and of the teacher-student relationships. In the following conceptual framework (Figure 1), all hypotheses are summarized and visually presented.
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 17
Figure 1
Conceptual Framework
SUPPORT FOR GENDER EQUALITY AMONG EARLY ADOLESCENTS 18
Methods Data
I used the International Civic and Citizenship Education Study (ICCS) 2016 to explore factors influencing support for gender equality. The ICCS 2016 is a study which has been developed with the goal of exploring students’ knowledge, skills and attitudes with regards to civic issues and education (Schulz et al., 2018). The ICCS 2016 study measures all relevant factors with regards to this study: the support for gender equality, as well as a large variety of student and classroom characteristics. The study was conducted in 24 countries worldwide, in which the study aimed to survey over 150 schools in each of these countries.4 Per school, one classroom participates in the study, in which both students and teachers are sampled. The ICCS study targets students to be around 14-year-old (Grade 8). The data selection for ICCS 2016 in Europe took place between February and June 2016. Students from one classroom per school as well as up to 15 teachers from that school were sampled.
The sampling took place in a stratified two-stage cluster sampling procedure. Students filled in a questionnaire on citizenship competences and knowledge, as well as on their
background. Teachers filled in a teacher questionnaire, and school principals a school
questionnaire, to provide information on their school as well as their teaching practices. More information on the procedure of ICCS 2016 can be found in the international report (Schulz et al., 2018).
In this thesis, 5 European countries were selected: Belgium (n=2931), Denmark (n=6254), Finland (n=3173), the Netherlands (n=2812) and Sweden (n=3264). The countries selected have similar scores on the Gender Equality Index and are therefore considered comparable in level of societal gender equality (Barbieri et al., 2020)5. I chose the European
4 In the country selection in this thesis, the Netherlands (n=123) did not meet this number.
5 See Appendix A for Gender Equality Index scores of these countries.
19
Union Gender Equality Index for measuring gender equality at the country level, as this index uses a broad range of indicators to come to their scores, and offers some benefits compared to other indices such as the Gender-related Development Index and the Gender Empowerment Measure, especially with regards to conceptualising gender equality (Bericat, 2012). Dotti Sani & Quaranta (2017) previously found that there are large cross-national disparities between countries with higher and lower level of societal gender equality and support for gender equality at the individual level. As I study the school and student characteristics that could be related to support for gender equality, I opted to leave out the countries with diverging scores on the Gender Equality Index, to make sure there are limited biases on country level. In these five countries, a total of 18.434 student were sampled, of which 49.4%
was female.
Variables
Dependent variable: Students’ endorsement of gender equality The level of support for gender equality in the ICCS dataset is measured by the answers of students on six items related to ‘students’ endorsement of gender equality’ in the ICCS dataset. Students were asked to what extent they endorsed the following statements regarding equality: ‘men and women should have equal opportunities to take part in government’, ‘men and women should have the same rights in every way’, ‘women should stay out of politics’, ‘when there are not many jobs available, men should have more right to a job than women’, ‘men and women should get equal pay when they are doing the same jobs’, and ‘men are better qualified to be political leaders than women’. The students answered these questions with a scale from one to four (1= strongly agree; 2 = agree; 3 = disagree; 4 = strongly disagree). A summary scale variable of these six items was created in the ICCS 2016 dataset, which shows a more comprehensive picture than just one these statements (Köhler et al., 2016). The scale was created via item response theory (IRT) with weighted likelihood estimates (WLE) with a
20
mean of 10 and a standard deviation of 2 within each participating country (Köhler et al., 2016). Furthermore, the scale variables were recoded, a higher score indicating more support for gender equality. This variable across all participating countries had high reliability (Cronbach’s alpha = 0.78).
Independent variables
Parents’ education level: Students were asked what their mother and/or father’s highest level of education was, ranking from not completing lower secondary education till a doctoral degree. This was recoded by ICCS using the International Standard Classification of
Education (ISCED), to ensure comparability across countries. The recoded variable indicates the highest educational level for either parent, ranking from 0 to 4, with a higher number indicating a higher level of education.
Maternal employment status: The maternal employment status was measured by two items.
Students were asked the following open-ended questions: ‘what is your mother’s or female guardian’s main job?’ and ‘what does your mother or female guardian do in her main job?’
The answers to these questions were recoded by the ICCS team (Schulz et al., 2018). First, they coded the answers based on the ISCO-08 classification of the International Labour Organisation, after which they transformed these classifications into scores using the International Socio-economic Index of Occupational Status. This scale ranks from 16 to 90, with scores below 50 points indicating a low till medium occupational status, whereas 50 points or above indicates a medium till high occupational status. The higher the score, the higher the occupational status. The ICCS data set does not measure whether a mother is a
‘stay-at-home’ mother, leaving out a relevant category for this study. Whether or not a mother works is a relevant factor, as theory prescribes us that women who are part of the workforce show more support for gender equality than women who are not (Bolzendahl &
Myers, 2004).
21
Gender: Students were asked to indicate their gender. The questionnaire they filled in was binary, thus they indicated whether they were a boy or a girl. This makes this variable a dummy variable, with only a score of 0 or 1 possible. I use boy (0) as reference category.
Open classroom climate: Students filled in six statements on their perception of openness in classroom discussions. These statements were ‘teachers encourage students to make up their own minds’, ‘teachers encourage students to express their opinions’, ‘students bring up current political events for discussion in class’, ‘students express opinions in class even when their opinions are different from most of the other students’, ‘teachers encourage students to discuss the issues with people having different opinions’ and ‘teachers present several sides of the issues when explaining them in class’. Students answer categories were ‘never’,
‘rarely’, ‘sometimes’ and ‘often’. The six statements were recoded into a scale for the students’ perception of the classroom climate and a higher score indicates more openness in the classroom. The scale was created via IRT with WLE, with a mean of 10 and a standard deviation of 2 within each participating country (Köhler et al., 2016).
Teacher-student relationship: Students filled in their agreement with five statements on their perception of the teacher-student relationships in school. These statements were ‘most of my teachers treat me fairly’, ‘students get along well with most teachers’, ‘most teachers are interested in students’ well-being’, ‘most of my teachers listen to what I have to say’ and ‘if I need extra help, I receive it from my teachers’. The students answered these questions with a scale from one to four (1= strongly agree; 4 = strongly disagree). These five statements were recoded into a scale for the students’ perception of teacher-student relationships in school, a higher score indicating more positive teacher-student relations. The scale was created via IRT with WLE, with a mean of 10 and a standard deviation of 2 within each participating country (Köhler et al., 2016).
Classroom average variables: The variable parents’ education level, maternal employment
22
status, open classroom climate and teacher-student relationship were aggregated to the classroom average.
Control variables
Immigration background is another factor that correlates with support for gender equality: students with an immigration background showed less support for gender equality (Seguino, 2011; Thornton et al., 1983). I therefore also control for immigration background (immigration status/no immigration status). Immigration is a contagious variable for which many different definitions are used. The ICCS dataset defines students as coming from an immigrant family when all of their parents are born abroad (Schulz et al., 2018). When one or none of their parents are born abroad, students are not from an immigrant family.
Not just the educational level of the parents could have an impact on the support for gender equality. The level of educational attainment of individuals also correlates with the support for gender equality (Thijs et al., 2019). I therefore control for the expected
educational attainment level of students as well.
Descriptive statistics
Table 1 portrays the descriptive statistics of all variables measured in this analysis, portraying both minima, maxima, mean scores and standard deviation (SD).
Table 1
Descriptive Statistics
Variable Min. Max. Mean SD
Student characteristics
Students’ endorsement of gender equality -4.32 0.89 0 1
Gender (1 = Female) 0 1 0.51 0.500
Parents’ highest educational level -3.48 0.97 0 1
Mother’s employment status -2.03 2.59 0 1
Students’ perception of classroom discussion -3.71 2.85 0 1
23
Table 1 (continued).
Students’ perception of teacher-student
relationship -3.64 2.01 0 1
Immigrant (1 = immigrant family) 0 1 0.084 0.278
Students’ expected educational attainment -2.42 0.91 0 1 School characteristics
Classroom average parental educational level -5.07 2.47 0 1 Classroom average mother’s employment
status
-3.88 3.66 0 1
Classroom average openness of discussion -4.62 4.37 0 1 Classroom average of students’ perception of
teacher-student relations at school
-4.53 5.47 0 1
N (countries) 5
N (schools) 1015
N (students) 14695
Source: International Civic and Citizenship Study 2016.
Note. With the exception of two dummy variables, all variables are standardized with a mean of 0 and a standard deviation of 1.
Analysis
All the analyses were conducted using SPSS 26.6 I performed a linear multilevel regression analysis with country fixed effects. I chose to perform a multilevel analysis, as data was measured at both the student and classroom level. This means the data is nested:
students (L1) within classrooms (L2), which calls for a multilevel analysis (Marsh et al., 2012). The data is not independent of each other, as the students share a classroom and thus the data is very likely correlated, which makes multilevel analysis the correct approach. I merged the different ICCS files for different countries using the IDB Analyzer, a software programme developed by the researcher behind the ICCS dataset (IEA). For more
information on the merging procedure with the IDB Analyzer, see the ICCS 2016 User Guide
6 I would have preferred to use software more suitable for multilevel analysis, however, I was unable obtain that type of software (such as Mplus or HLM) at student prices. I therefore chose to use SPSS. This did mean compromising in some areas, which I will explain in this analysis section.
24
(Köhler et al., 2016).
After this, I checked the dataset for missing values. I cleaned cases with missing values using listwise deletion. Missing values might not be random: for example students from a disadvantaged socio-economic background are more likely to skip questions related to their parents’ occupational status (OECD, 2009). However, leaving the missing values in could lead to biased estimates. Even more important, I make sure that when comparing models in the analysis, I compare the same sample of students. I therefore chose to delete missing values. After deleting cases with missing values, the sample was reduced from 18.434 to 14.695 students.
SPSS does not have the functionality to use different sampling weights for different levels as advised (Köhler et al., 2016). I therefore chose to test my models unweighted. As a result of this, the results must be interpreted very carefully and are not final: the standard errors could possibly be biased (Heck et al., 2014) and results should be replicated using different statistical software.
The independent variables7 were standardized to overcome differences in the range of the outcomes of variables. The variables were standardized by subtracting the mean and dividing it by the standard deviation. The first step of standardization (subtracting the mean) is similar to grand mean centring mean (!!" − !#) of variables, but standardization takes it one step further, to make comparing predictor after analysis possible. Centring is a critical
element of multilevel analysis, to make the interpretation of results more meaningful: it adjusts the variables for differences among students within schools (Ma, X., Ma & Bradley, 2008).
Prior to running the analysis, I performed tests to see whether any of the assumptions for multilevel linear regression analysis were violated (see Appendix B). A plotted P-P plot
7 With the exception of dummy variables, as standardizing dummy variables makes them difficult to interpret.
25
of the residuals showed the data is not perfectly normally distributed but has a negative skew (Figure 11). However, square root transformation (Figure 12) nor reflect and logarithmic transformation (Figure 13) did not improve the normal distribution (see Appendix C). I therefore chose to move forward with the data with standardizing being the only transformations to it. None of the other assumptions were violated.
Before testing the hypotheses, a correlational analysis was performed. After that, a null model was created to obtain intra class correlation. In the first model, country-fixed effects were added to account for any variance that could be the result of country differences.
In the second model, the control variables were added. In the third model, all other student variables were added to test the student characteristic hypotheses. In the fourth model, school variables were added to test the school characteristic hypotheses. In the fifth and last model, the same-level interaction hypothesis was tested. All variables were added with a fixed slope and random intercept. The results are portrayed in the following chapter.
26
Results Correlational analysis
Table 2 portrays a correlation matrix to test for possible correlations between the variables. For the dependent variable, students’ endorsement of gender equality, correlations with all other variables were visible and significant (between r = -.059 and r=.338, p <.01).
Put differently, the higher adolescents’ parents’ educational level, maternal employment status, expected educational attainment, perception of open classroom climate and perception of teacher-student relationships, the stronger they endorse gender equality. If the classroom average parental educational level, maternal employment status, support for gender equality, perception of open classroom climate and perception of teacher-student relationships is higher, the support for gender equality of an adolescent is also stronger. Furthermore, the correlation between gender and support for gender equality is stronger for girls. There is a negative correlation between immigration status and support for gender equality, which indicates that if a student is from an immigrant family, the support for gender equality is lower. Note that this a very weak correlation (r = -.059).
Multilevel analysis
All results of the multilevel analysis are summarised in Table 3. Before elaborating on the results, I assess the data structure, to check the suitability of a multilevel analysis. The intraclass correlation was estimated in the null model (Table 3). This showed that 6.9 percent of the variance in adolescents’ support for gender equality can be explained by variability between schools, and 93.2 percent of the variance in the support for gender equality can be attributed to student or country. Intraclass correlation levels above 5 percent are considered substantial (Heck et al., 2014) and thus need multilevel analysis, due to the variability found between schools.
27
Table 2
Correlation Matrix for Study Variables
Variable 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
1. Students’ endorsement of gender equality
-
2. Gender .338** -
3. Parents’ highest educational level
.140** -.038** -
4. Mother’s employment status .137** -.031** .439** -
5. Students’ perception of open classroom climate
.211** .075** .070** .075** -
6. Students’ perception of teacher-student relationship
.215** .065** .047** .065** .308** -
7. Classroom average parental educational level
.164** .274** .393** .274** .106** .059** -
28
Table 2 (continued).
8. Classroom average mother’s employment status
.166** .372** .290** .372** .128** .085** .738** -
9. Classroom average perception of open classroom climate
.154** .123** .108** .123** .388** .192** .274** .330** -
10. Classroom average of students’ perception of teacher-student relations at school
.132** .086** .063** .086** .202** .367** .161** .232** .522** -
11. Immigrant family -.059** -.155** -.110** -.155** .019* -.034** -.056** -.074** .005 -.040** - 12. Students expected
educational attainment
.222** .236** .353** .229** .146** .144** .312** .236** .135** .099** .051** -
N (countries) 5
N (schools) 1015
N (students) 14695
* Correlation is significant at the 0.05 level (2-tailed).
** Correlation is significant at the 0.01 level (2-tailed).
29
Before elaborating on the results of the models, I will portray the goodness of fit for these models, which was tested using log likelihood tests (Table 3). Compared to Model 0, adding country dummies (Model 1) significantly improved the model (!!(3) = 308.35, p <
0.01). Adding the control variables in Model 2 improved the model again (!!(2) = 617.84, p
< 0.01), compared to the previous model. Model 3 was again better than Model 2 (!!(5) = 2544.19, p < 0.01) and Model 4 was better than Model 3 (!!(4) = 63.12, p < 0.01). Model 5 furthermore improved Model 4 (!!(2) = 179,599, p < 0.01).
In the first model, I added country fixed effects to control for any variance due to differences between countries rather than between schools and students. After adding the country fixed effects, the variance in support for gender equality that can be explained due to variability between schools reduced to 4.6 percent. According to Heck and colleagues (2014), this variance is not substantial. However, as the data has a nested structure, I decided to move on with multilevel analysis to account for the data structure. The country fixed effects were all significant. The average support for gender equality was highest in Sweden, followed by Denmark, Finland, Belgium, and last the Netherlands. These differences between countries are accounted for by using the country fixed effects.
In the second model, the control variables immigration status and expected
educational attainment were added. I find that immigration status has a negative, statistically significant effect. With regards to the other control variable, expected educational
attainment, the results are also in line with theoretical expectations. Students who are expected to obtain a higher educational level, show more support for gender equality.
In order to test H1, concerning the relation between parental educational level and maternal employment status, and support for gender equality, Model 3 contains the student characteristics that measure these items. The results show that there is a statistically
significant relation for these variables. These findings indicate that the higher their parents’
30
educational level and the higher the level of maternal employment, the more support students show for gender equality while controlling for students’ expected educational attainment among others. Students’ family background was positively related to the support for gender equality, thus confirming Hypotheses 2a and 2b.
To test H4a and H5a on adolescents’ experiences regarding open classroom climate and teacher-student relationship and its relationship with support for gender equality, Model 3 also contains the student characteristics that measure open classroom climate and teacher- student relationships. The results show that these also have a statistically significant relation with support for gender equality. Students who experience a more open classroom climate and a more positive teacher–student relationship show greater support for gender equality, thus confirming Hypotheses 4a and 5a. Moreover, as expected in Hypothesis 1, the support for gender equality has a relation with gender: support for gender equality is higher for girls than boys.
In Model 4, the school characteristic variables were added. The classroom average parental educational level and maternal employment status have a statistically significant positive relation with support for gender equality. Put differently, being in a classroom with peers whose parents on average have a higher educational level, or whose mothers on average have a higher employment status, relates to support for gender equality, even while
controlling for a student’s own expected educational attainment and immigration status, therefore Hypothesis 3a and 3b are confirmed. The findings show no significant relation between average open classroom climate perception or classroom average teacher-student relationship and support for gender equality. Therefore Hypothesis 4b and 5b are not confirmed.
In the fifth model, the interaction hypothesis (H6) was tested. This test was performed to find whether the educational experiences of students, as indicated by their perception of an
31
open classroom climate and teacher-student relationships, could play a positive role in the differences in support for gender equality between boys and girls. The analysis did not show a significant relation between open classroom climate and gender with regards to their
support for gender equality, which indicates there are no significant differences in the relation between support for gender equality and an open classroom climate for boys and girls.
However, the analysis did find a significant negative relation between the perception of a teacher-student relationship and gender with regards to their support for gender equality. This indicates a stronger relation between boys and teacher-student relationships and their support for gender equality (Figure 2). Hypothesis 6 is therefore partially supported.
Figure 2.
Interaction Effect of Teacher-Student Relationships and Gender on Support for Gender Equality
-1 -0,8 -0,6 -0,4 -0,2 0 0,2 0,4 0,6 0,8 1
TSRs low TSRs high
Support for gender equality
Boys Girls
32
Table 3
Multilevel Models of the Predictors of Adolescents’ Support for Gender Equality
Variables M0 M1 M2 M3 M4 M5
Intercept -0.017** (0.014) 0.242** (0.025) 0.234**
(0.024)
-0.124**
(0.022)
-0.136**
(0.022)
-0.135**
(0.022) Level 1
Gender 0.647**
(0.015)
0.646**
(0.015)
0.648**
(0.015) Parents’ highest
educational level
0.044**
(0.008)
0.033**
(0.009)
0.033**
(0.009) Mother’s employment
status
0.060**
(0.008)
0.048**
(0.009)
0.049**
(0.009) Students’ perception of
open classroom discussion
0.081**
(0.008)
0.078**
(0.008)
0.089**
(0.011) Students’ perception of
TSRs
0.147**
(0.008)
0.146**
(0.008)
0.233**
(0.011) Control
Immigrant status (1 =
immigrant) -0.273**
(0.029) -0.226**
(0.027) -0.216**
(0.027) -0.219**
(0.027) Expected educational
attainment
0.198**
(0.008)
0.132**
(0.008)
0.125**
(0.008)
0.124**
(0.008) Level 2
Classroom average parental educational level
0.037**
(0.014) 0.037**
(0.014)
33
Table 3 (continued).
Classroom average mothers’ employment status
0.042**
(0.013)
0.041**
(0.013) Classroom average
openness of discussion
0.000 (0.011) 0.001 (0.011) Classroom average
TSRs
0.001 (0.011) 0.003 (0.010) Interaction (Level 1)
Gender * Open
classroom climate -0.029
(0.015) Gender * Teacher-
student relationship
-0.182**
(0.015) Country fixed effects1
Belgium -0.375** (0.033) -0.377**
(0.031)
-0.289**
(0.028)
-0.285**
(0.028)
-0.281**
(0.028)
Denmark -0.143** (0.027) -0.112**
(0.027)
-0.149**
(0.024)
-0.136**
(0.024)
-0.134**
(0.024)
Finland -0.342** (0.032) -0.321**
(0.031)
-0.247**
(0.030)
-0.225**
(0.027)
-0.223**
(0.028)
Netherlands -0.542** (0.035) -0.384**
(0.033)
-0.321**
(0.027)
-0.292**
(0.031)
-0.292**
(0.031)
School ICC 0.069 0.046 0.029 0.020 0.017 0.016
-2*log likelihood 41364.827 41056.482 40438.641 37894.447 37831.324 37654.725
N (countries) 5
N (schools) 1015
N (students) 14695
*p < 0.05 ** p <0.01
1 Reference country is Sweden.
34
Conclusions
The aim of this study was to analyse which student and school characteristics have an influence on the support of gender equality among early adolescents. Research on the factors that play a role for support for gender equality has thus far focussed on which individual factors, such as students’ family background characteristics matter, whereas school
characteristics have been neglected (Kågesten et al., 2016). This study therefore considers the role of the school. First, I looked at individual characteristics, which have previously been found relevant for their support for gender equality. Next, I took into account both an open classroom climate and teacher-student relationships, as previous studies found these to have influence on civic attitudes, such as support for gender equality (i.e. Campbell, 2008;
Dassonneville et al., 2012 and Diazgranados & Sandoval-Hernandez, 2020 for open classroom climate; i.e. Sampermans & Claes, 2018 and Wanders et al., 2020 for teacher- student relationships). Furthermore, I questioned whether the composition of the classroom mattered. Peers influence each other – does the composition of the classroom in terms of parental resources and educational experiences impact the support for gender equality? By looking at both individual characteristics of students, which previously have been found relevant for their support for gender equality, in relation to student characteristics, I aim to find what exactly the role of the school in support for gender equality implies.
The results show that indeed, as previous studies have shown us, students’
background matters. Girls portray a stronger relation with gender attitudes than boys,
indicating more support for gender equality among female students than boys. The higher the parents’ educational level or the employment status of the mother, the more support for gender equality students portray. Parental resources are strong indicators for support for gender equality. Regarding student experiences in school, students’ perception of how open the classroom climate is and how good their teacher-student relations are, also has a positive
35
impact on their support for gender equality. This resonates with previous research on civic attitudes and school characteristics (Barber, Torney-Purta, et al., 2015; Claes et al., 2017;
Sampermans & Claes, 2018; Treviño et al., 2017; Wanders et al., 2020). This means that the individual experiences of students in a school also matter, or more particularly, in their classroom. Lastly, the results show that the way a classroom is composed also influences the support for gender equality of individual students. If the classroom has, on average, more children with higher educated parents or mothers with a higher employment status, this will have a positive impact on the support for gender equality of the student, whilst controlling for the student’s own background situation. With regards to whether the classroom average experience of an open classroom climate portrays a relation: there is no indication that a shared experience of an open classroom climate or a classroom average positive perception of teacher-student relationships has a positive relation with the support for gender equality.
Furthermore, both significant classroom compositional effects are small. This could be due to the small portion of variance that is explained by the classroom - only a small portion of differences among students in their support for gender equality comes from the way their classroom is composed. This is in line with previous findings on attitudes with regards to ethnic minorities (Diazgranados & Sandoval-Hernandez, 2020).
To answer the first research question, I find that both student and school
characteristics are related to students’ support for gender equality. Factors from students’
background, as well as their experiences in school, show a strong relation with support for gender equality. There is some indication that classroom composition matters as well, with regards to parental resources and its effect on the classroom, yet not in terms of educational experiences.
I also consider whether educational experiences for adolescents can compensate for differences in support for gender equality between genders. As expected, there are visible
36
differences in the support for gender equality between boys and girls. When looking at one of students’ educational experiences – an open classroom climate – no significant differences are found. The experience of an open classroom climate and its role on the support for gender equality is similar for boys and girls. However, with regards to the teacher-student
relationship, differences between boys and girls are found. The relation between teacher- student relationship and support for gender equality was stronger for boys than for girls. This means that a positive teacher-student relationship is likely to have a stronger role in the support for gender equality: if the male student perceives the relationships to be more positive, it is more likely that their support for gender equality will be higher. For female students, this relation is less strong, and the perception of teacher-student relationships plays a smaller role in their support for gender equality.
Discussion
There were several limitations in conducting this thesis. First and most of all: some caveats are in place regarding the generalizability of the findings, given potential sampling biases at the student and school level. SPSS is not able to apply weights to the data when using multilevel analysis. Because the data was tested whilst unweighted, the results might be skewed and not generalizable outside of this study. In general, SPSS is not the preferred statistical analysis software for multilevel analysis, and I would definitely see the benefits of replicating this study using different software.
Second, a potentially confounding factor was not included in the present study, hence its role has not been controlled for, namely religious affiliation. Among adults and
adolescents, religiosity has been found to have an effect on support for gender equality, with stronger religiosity showing more ‘gender inequitable beliefs’, consistent over religions (Halimi, Consuegra, et al., 2020; Kretschmer, 2018; Marsh et al., 2012). However, several countries participating in the International Civic and Citizenship Education Study (ICCS) did
37
not administer religious affiliation, among which Denmark, Finland and Sweden. This is why I chose to leave out control variables for religion. I did include these countries, as they have similar Gender Equality Index scores as Belgium and the Netherlands (Barbieri et al., 2020).
Although religious affiliation has previously shown a negative effect on the support for gender equality, I do not expect that including religious affiliation as control variable would have drastically altered the patterns I have reported. A previous study (Davis, 2007) found parental socialization to be of more importance than religious affiliation. Yet, it would be of importance to verify this claim in new research, and I would advise countries partaking in the ICCS study to include questions regarding religiosity, in order to improve studies using this data set.
Other limitations mainly result from the selected categorizations in the ICCS study, and in most other studies on the formation of attitudes in adolescence. The gender binary is consistent in this research domain, especially in studies on support for gender equality.
However, gender is not a binary concept and therefore should not be considered as such. The same goes for the focus on heterosexual parents. Most studies focus on the role of same-sex couples as parents, whereas other family compositions are reality as well. Sutfin et al. (2008) found that young children between four and six years old with lesbian parents hold less traditional attitudes towards gender equality. The composition of the family thus could have impacted this study. Yet, as there are – to my awareness - no previous studies on the impact of opposite-sex couples and other family compositions on support for gender equality among early adolescents, I am unsure what impact this could have had on the study.
Lastly, the ICCS study did not measure whether mothers were unemployed or what the household division at home was. Rather, it measured what kind of employment status a mother has. Whether a mother is unemployed or employed (Ex & Janssens, 1998), the amount of hours the mother works (Cunningham et al., 2005) or the way parents divide
38
household tasks (Platt & Polavieja, 2016), could impact support for gender equality of
adolescents. The more equal the division of paid and unpaid labour (i.e., household tasks), the more support for gender equality children hold. These types of questions would therefore have been a welcome addition to the data.
There are several suggestions to be made for further research. First, qualitative research would be beneficial for studies on the support for gender equality. At this point, most studies look at students’ perception of a classroom and its climate, or the teacher- student relationships within the school, through quantitative methods. Especially
questionnaires such as the ICCS study are heavily used. Yet, there are reasons to use other research methods to study civic attitudes, which are more qualitative of nature. First, questionnaires offer limited personal information to direct further learning, whereas
classroom observations or portfolios do (Daas et al., 2016). If we wish to improve support for gender equality, it would be beneficial to use research methods which are able to direct further learning of the student. Second, it still remains unclear how casual mechanisms in civic attitudes and knowledge improvement works. For instance, we see there are diverse opinions on the causal mechanism behind open classroom climate and civic competences.
Whereas Campbell (2008) argues it is the diverse and conflictual opinions in a group that leads to improved civic knowledge, others believe deliberative processes are behind improvements in civic knowledge or attitudes (i.e. Enslin et al., 2001). Qualitative studies could offer insight into the causal relations.
Second, this study makes me wonder what school characteristics could actually compensate for differences between boys and girls, rather than accelerate these. This study did not find any statistically significant results on the impact of an open classroom climate, and found an accelerating effect for teacher-student relationships. Are there any factors within the school that could compensate for differences in support for gender equality
39
between boys and girls, and which interventions might be necessary? More research into these factors could induce more support for gender equality among all genders.
Third, longitudinal studies in the support for gender equality have previously been performed (i.e. Bryant, 2003; Burt & Scott, 2002; Paul Halpern & Perry-Jenkings, 2016).
However, these were mostly performed in either the United States or the United Kingdom and more importantly, focused mainly on socio-demographic factors that might influence support for gender equality and. Longitudinal research into school characteristics that could impact gender equality over time has thus far not been performed. This type of research would add to a better understanding of the role of the school in the development of gender attitudes among youngsters. Specifically, longitudinal research could take into account the fact that school has an ongoing influence on children, as most children are in school for the majority of their childhood.
Lastly, factors outside of school such as communities or associations which students are part of, could also impact support for gender equality. Contact theory shows us that attitudes towards certain groups can be impacted by getting into contact with these groups (Janmaat, 2012). Furthermore, this type of research could add to the work already done in this thesis, where I studied the effect of peers via classroom composition. Adolescents who meet other youngsters outside of school, such as within voluntary associations or political youth groups, could be influenced by them through contact and peer effects. Studies into non- school organisations could provide more insights into the role of peers outside of schools.
Implications for practice
This thesis is one of the first studies to look into the role of schools in the support for gender equality among early adolescents. The school does play a – although small – role in the support for gender equality, and currently mechanisms to increase support for gender equality are limited. This study offers some first insights into the role teachers can play in a
