Gender and communication in children and school: aligning theory and evidence

There are many ways in which communication practices can be gendered. It is important to recognise and understand how, why, and in what contexts these can be observed. This paper focusses specifically on children and the context of formal schooling, analysing the empirical data and connecting it with theoretical analyses to provide an overall picture of gendered communication amongst children in educational contexts, especially with regard to classroom discussion. This paper demonstrates that masculine dominance in communication is present in children’s communication, and explanations for this occurrence fall on traditional gender stereotypes that still linger, to greater and lesser degrees, in all parts of our society, including education. Therefore, it is incumbent upon educators to be aware of this threat to gender equitable learning and plan accordingly.


Introduction
Communication is a key part of understanding and interacting with the world, and this is especially so for children. Moreover, communication practices can be gendered in many ways that have a substantial effect on speakers and their learning. The research focus for this paper is on how theoretical analyses of gendered communication interact with empirical observations of gendered communication. More precisely, the guiding research question is: in what ways can gender theory inform our understanding of empirical data on gendered communication in education to address this issue through teachers' pedagogical practice? This is an important issue to address, as theory and practice are not always in alignment. On the issue of gender specifically, this paper describes the ways in which this misalignment can be observed. Thus, providing evidence for the value of research which can better align the theory with the practice.
Theory in this area includes a focus on concerns around the way in which gender is described as an influencing factor in research. The concern is around essentialising women's traits an inherently female, justifying dualist modes of thinking that separate men and women; male and female; masculine and feminine. Cameron (2005, p. 81) states that language and gender research has moved in recent times from a focus on gender difference to the study of gendered identities and practices. This aligns with Mills (2004), who suggests that this research challenges the homogeneity of women as a group and previous research that categorised all women under the same umbrella. Instead, gender is viewed through an intersectional lens in locally oriented contexts. This means that while gender may be an influencing factor in the research, there are other variables such as race, class, and context that influence how language use differs based on gender. In other words, such research would not say that all women use language like this and all men use language like that. McElhinny (2014, p. 51) states that to argue that differences found in people's behaviour can be explained solely by gender fails to consider how gender and gendered practices are constructed. In language research, gender is no longer viewed from the essentialist perspective, tying women to a particular style of language use based solely on their gender. Holmes and Meyerhoff (2003, p. 8) state that the field has moved well beyond descriptions of differences between men's and women's speech. Instead, gender is viewed as a locally managed practical accomplishment (Garfinkel 1967) or an emergent performance (Butler 1990). Indeed, research suggests that both men and women use a wide range of strategies, for example, both men and women use affiliative or relational strategies which are traditionally associated only with feminine styles of interaction (Baxter 2006;Holmes 2006;Holmes and Stubbe 2003;McConnell-Ginet 2011;Mullany 2008). Ochs (1992) states that this diversity demonstrates that specific linguistic features do not encode social meanings such as gender, but that this is done indirectly through how society typically associations these features with certain positions, such as femininity.
Conversely, empirical research on gender and communication in schools primarily distinguishes between students through the category of biological sex. When empirical data is recorded, it is often recorded as boys doing this and girls doing that. This is exactly the problem of essentialising people based on sex categories discussed in the gendered research above. There are some progressive exceptions, such as Yu et al. (2021) whose empirical research identifies four different types of girls and three different types of boys, generating an empirical analysis of variations within gender rather than just across gender. Although, this is not the norm. Therefore, when attempts are made to integrate the empirical research with the theory (such as this paper), it is impossible not to talk about girls and boys in an essentialised ways because almost all empirical data has been essentialised in this way. For this reason, this paper will discuss what is typical of girls and boys, as observed through empirical data, and, at the same time, use that essentialised data to discuss the theory about the limitations of essentialising children by of gender. This paper acknowledges that this is a contradiction but sees no way around it if empirical data is to be incorporated into the analysis.
In this light, although this section will discuss the relevance of gender to communication, it is important to note that gendered communication is diverse and changes based on many other variables. However, what is particularly useful is the theorisation of gendered communication practices. This will enable an understanding of why certain communication practices are aligned with femininity as a way to better inform empirical research which may suggest that certain women and girls in certain circumstances practice feminine communication more than men and boys. Research on gender and communication moves beyond who does how much of what and into more rewarding inquiry such as the purpose of language variation between genders and how it gives rise to social meaning, including how language use gives more opportunity to certain groups of people (Meyerhoff 2014, pp. 100-101). Menard-Warwick et al. (2014, p. 486) caution against the thinking that might assume masculine dominance is no longer relevant, and instead call for greater research into gender inequalities that extend beyond stating individual differences. In speaking about gender and language in the context of education settings, they note that "what we need is not some new grand generalisation about language and gender that can be applied across educational settings, but rather renewed ways of theorising gender and relating these theories to local contexts in ways accessible to educational practitioners who wish to make a difference in their own classrooms" (Menard-Warwick et al. 2014, p. 486). This is exactly the kind of theorising that this paper attempts to demystify by understanding gendered dialogic learning practices. But first, we must gain a grasp on the literature on gender and communication.

Empirical evidence on gendered children and educational communication
This section will delineate the empirical evidence demonstrating a difference in communication based on gender in children, young people, and the educational contexts in which these young people are communicating.
The research of Eliot (2013) and Halpern et al. (2011) has found that the evidence indicating that sex differences in the human brain have implications for learning is extremely weak (cf. Helgeson 2017, p. 266). Recently, a meta-synthesis of 30 years of research on human brain sex differences revealed little to no difference in male compared to female brains (Eliot et al. 2021). There has been a focus on the brain and its role in learning broadly in education in recent years (Helgeson 2017, p. 169), and this has crossed over to the area of gender and education where some speculate that brain differences between male and female students can explain some behavioural differences (Lenroot and Giedd 2010;Semrud-Clikeman et al. 2012). It is important to note the similarity to this line of gender difference based on the brain to the ideas from 50 years ago that focussed gender differences based on biology (cf. Harding 1986, p. 179). These both take a line of essentialising the female experience as universal: all females have the same biology/brain, therefore females (as a homogenous group) are different to males in the same ways. Notably, this line of essentialising gender differences logically leads to the conclusion that these differences are unchangeable: if gender differences are based on biology or the brain, and we cannot change those aspects of ourselves, then the gender differences that are caused by them cannot be changed. In the educational context, this idea of essentialising males and females comes up again in this new way: the brain. The research makes it clear that this educational theory of brain differences causing variation in behaviour based on gender also lacks evidence. Instead, the theoretical focus here will be on the ways in which society influences gendered behaviour, language, and communication, which aligns more with the evidence base (de Vries and Sodersten 2009;Ngun et al. 2011).

Dominance and care
Masculine dominance is a common thread across many classroom settings. The research of Goodwin (1980) finds that boys tend to issue direct commands whereas girls used inclusive suggestions for action. He elaborates on this difference in saying that it reflects different peer group organisation; with boys' groups based on power and girls' groups based on solidarity. Boys' groups tend to be leader-dominated and hierarchical, while girls rely on negotiation. Davies (2003, pp. 128-129) states that girls create a sense of unity through their language, while boys display an anti-school, anti-female cultural attitude, restricting their freedom of language use to conform with the masculine gender role. She suggests that girls tend to be more successful in school because of this collegiality in learning, whereas boys sabotage their classmates' participation due to the ideal of masculine power, authority, and dominance. This is especially harsh on non-masculine boys (Besnier and Philips 2014, p. 129). This demonstrates that although much of this research on masculine dominance focusses on boys enacting it and girls being disadvantaged by it, masculine dominance cannot be universally applied to all male students. In young children, this dominance and care dualism is also evident. Kyratzis (2001) finds that pre-school boys often reference acts of aggression (e.g., ''smash this girl!'), while the girls' language and play suggests valuing lovingness and graciousness. The meta-analysis of Leaper and Smith (2004) finds that girls produce significantly more affiliative speech, which is defined as language used to establish or maintain connections with others. They also find that there is a significant difference in favour of boys' using measures of assertiveness and directive language. Bablekou (2009) also researched the use of dominant and submissive language in children, finding that boys use more dominant and aggressive communication styles. In relating the dominance and care dualism to research on empathy, Staley (1982) finds girls as young as 4 employing more reflective and emotional language than same-age boys. In a study of third grade children, girls adopted a negotiating role in group work, while boys took a competitive approach (Cook-Gumperz and Szymanski 2001). Gayle-Hackett (1986) researched gender and conflict situations with both adults and children, finding that there was a gender difference, but more importantly, the gender difference was greater with high school students than adults. This may indicate that certain gendered communication practices are heightened in young people more so than they are in adults. The effect of this when considering gender and classroom communication will be important in understanding how age is a moderating variable in gender differences in communication that may make gendered communication differences increase in severity (Cupach and Canary, 2009, p. 238).
Understanding some correlated factors of gender difference in communication, we can turn to personality characteristics. Xie et al. (2011) found that in children in grades 4 through 12, there existed differences in aggression favouring boys. On the other side of the aggression spectrum, adolescent girls scored higher in degrees of conscientiousness and agreeableness, and these effects were similar across cultures in the 23 countries in which research was conducted (De Bolle et al. 2015). Hicks et al. (2008) and Steinmayr and Spinath (2008) have also found that agreeableness was greater in girls than boys.
In terms of play and organisation, girls are likely to play in pairs and boys in larger groups. Same-sex play organisation may socialise children into a pattern of gender stereotypical interaction styles Helgeson 2017, p. 279). The dominance of boys has been discussed above, but moreover, the dominance specifically targeting femininity is also an important factor to consider. Maccoby (1998) states that boys' groups are more exclusionary than girls', and that boys' sense of masculinity leads them to reject associations with femininity, while girls do not reject all associations with masculinity. Research on the effects on this samesex play organisation states that girls will find it difficult to influence boys, meaning girls will interact with boys less often (Maccoby 1998;Mehta and Strough 2009). The prosocial play of girls, in contrast to the aggressive play typical of boys, leads girls and boys to form different styles of communication, and the style of communication that is formed from the prosocial play typical of girls is a style that makes it hard for those communications to penetrate more aggressive and assertive communication styles typical of boys (Helgeson 2017, p. 279;McCloskey 1996). This communication style typical of girls is associated with fostering connection through agreement, turn-taking, acknowledging feelings, and peer teaching (Whiting and Edwards 1988). It fosters a communication style that leads girls to make more polite suggestions rather than strong assertions: polite assertions are usually well received by others with the same communication style. Boys, conversely, tend to interrupt, threaten, refuse to comply, name-call, and try to top one another's stories (Helgeson 2017, p. 281). This leads boys to become nonresponsive to polite suggestions, making the communication style typical of girls ineffective. Girls find their typical communication style effective with adults and other girls, but not with boys. This has been applied as a cause to explain why girls spend more time in close proximity to authority figures, such as teachers, than boys do. "It was first thought that girls stayed closer to teachers because of their affiliative nature. However, girls stand near teachers only in the presence of boys. Girls likely believe that an adult authority figure will temper boys' dominant behaviour" (p. 283).

Language use
Boys have been observed speaking more than girls at each grade level. This is an interesting feature to note, as girls are regarded as having greater linguistic skills, both in an academic testing situation (ACARA 2018; OECD 2019a) and in social situations (Eckert 2000). Yet despite this perceived prowess in language, boys tend to speak more often in class. Boys' dominance in classroom dialogue has been shown through quantitative research that examines the amount of time that boys spend talking in comparison to girls, as well as the different types of contributions that girls and boys make to classroom dialogues. For example, in a study by Aukrust (2008), boys were observed speaking more than girls at different grade levels (1,3,6,9). Aukrust (2008) videotaped 26 classrooms across four grade levels for one week each. Teachers were asked to choose a whole-class discussion where students would be able to share their views and opinions. These discussions were recorded, transcribed, and analysed to identify: (1) frequency of participation, (2) turn allocation, (3) spontaneous utterances, and (4) overlapping utterances. Aukrust (2008) found that boys contributed 61% of all utterances compared to 39% from girls. On average, boys made 10.3 utterances compared to 7.3 utterances for girls. Worryingly, this gender difference appears widen as students get older and progress through their schooling. That is, the gender difference was lowest with Grade 1 students, with an average of 7.3 utterances versus 8.3 favouring boys (or 42% compared to 58%). For the Grade 9 cohort it was 3.3 to 7.6 utterances per student favouring boys (or 27% compared to 73%). Boys also made many more uninvited comments: in Grade 1, girls made 58 uninvited comments compared to 88 for boys; in Grade 6, boys made triple the number of uninvited comments (62 compared to 187); in Grade 9 the numbers were 10 to 85 favouring boys. Boys also had more overlapping utterances (i.e., talking over others) at every grade level, with a difference of 14 to 59 in Grade 9. Further, boys not only speak more, but take the floor more often than girls. Girls tended to wait to be assigned a speaker role appointed by the teacher, with girls receiving nearly double the speaking turn allocations initiated by the teacher than boys, despite having fewer speaking turns overall (both teacher-allocated and self-initiated).
Similarly, Keogh et al. (2000, p. 33) report that "boys dominate both the amount and type of verbal interaction (p. 33)," while Swann and Graddol (1988) find boys speak more words per turn, have more turns per interchange, and have more interchanges. Julé (2005) analysed linguistic space as it relates to gender in seven-year-olds and reports that there is a clear difference in the amount of talk produced by girls and boys, with boys dominating at a rate of 88% to 12% over forty hours of observation in the single classroom used for this study. Similarly, Blair (2000) finds that there is a clear masculine dominance in public talk through her research of students in eighth grade. Moreover, Collins and Johnston-Wilder (2005) note the disproportionate number of girls compared to boys who have spent whole days in school not participating in the kinds of talk that their teachers see as central to learning. In the tertiary classroom, and where female students exceedingly outnumber male students, Galvin et al. (2013) report that masculine dominance still persists in classroom discussion. These quantitative analyses demonstrate the dominance of boys in classroom discussion. By the numbers, it can be seen how much boys dominate. For this paper, the analysis will not focus on question of if boys dominate classroom discussion, in the way that these studies have investigated that question quantitatively. Instead, this paper focuses on why and how boys dominate, and relating that why and how to the gender dualistic thinking, stereotypes, and norms that allow this masculine dominance to persist.
Though this research suggests differences in language use, it is noted that girls' achievements in the researched classrooms was equal to or higher than boys. Therefore, Aukrust (2008) suggests that though there are observed differences, these interaction patterns do not put either girls or boys at an educational disadvantage. For clarity, the use of educational disadvantage from Aukrust specifically refers to academic results. However, it is noted there are other educational disadvantages that are non-academic. It is important to understand that girls may be performing equal to or greater than boys academically and still be disadvantaged by educational practices. Aukrust (2008, p. 249) seems to acknowledge this when she says that although there seems to be no educational disadvantage (read: academic disadvantage), girls "seemed less permitted to challenge conversational rules and expectations, which may have impacted the repertoire of the conversational strategies they acquired." There is also data that suggest a significant wellbeing gap between girls and boys in education. For example, data for all OECD countries found that 11% more boys than girls (at age 15) report being satisfied with their lives (OECD 2019b, p. 154). The OECD concluded that "these results imply that boys were more likely than girls to report a greater sense of meaning in life," and differences in this area are more strongly pronounced between gender than between socio-economic status or immigrant background (p. 166). The OECD also finds that "in all [OECD] countries and economies, girls were more likely than boys to report sometimes or always feeling sad" (p. 176) and "in almost every education system, girls expressed greater fear of failure than boys, and this gender gap was considerably wider amongst top-performing students" (p. 188). Furthermore, "male students reported a significantly greater sense of belonging at school than female students" (De Bortoli 2018, p. 3). As De Bortoli notes, this worldwide trend is also prevalent in Australia. Australian male students were − 0.03 below the OECD average for students' sense of belonging, whereas female students were − 0.22, a difference of 0.19. For context, "the difference in the sense of belonging for Australian indigenous compared to non-indigenous students is 0.11, with results of − 0.12 for non-indigenous students and − 0.23 for indigenous students" (p. 18). Thus, the gap in sense of belonging for girls compared to boys in Australia is nearly twice as big as the gap between indigenous and non-indigenous students. And girls' sense of belonging is nearly as low as that of indigenous students objectively (− 0.22 compared to − 0.23). Moreover, girls' sense of belonging has declined over time. Between 2003 and 2015, there was a 20 percent decrease in the number of girls who disagreed with the statement "I feel like an outsider (or left out of things) at school", compared to a 12 percent decrease for boys (De Bortoli 2018, p. 24). So while the twenty-first century has seen some encouraging decreases in gender inequality in terms of academic performance, PISA results show that in our schools the situation for girls' wellbeing, connectedness, and belonging is actually be getting worse, both objectively worse and worse in comparison to boys.

Gender stereotypes (empirical research)
The previous sections on gender, children, and education have focussed largely on the empirical data indicating masculine dominance and its manifestations, and this section now turns to a focus on the conception of gender stereotypes as contributing to this masculine dominance. Leaper (1994) explains that this is tied to the norms of behaviour expected from children within gender segregated peer groups. She states that the gender-typed activities of girls foster nurturance, affection, and social sensitivity, whereas boys' activities emphasise aggression, competition, and dominance. This can be said to lead to girls' caring and considerate disposition, and boys' competitive and dominant disposition in other areas such as classroom discussion (Goodwin and Kyratzis 2014, p. 511).
There is evidence of a backlash effect for non-gender-conforming behaviour with children. This gender counterstereotypical backlash helps to maintain gender-typed behaviour in boys and girls. Children police the social landscape and evaluate peers in regards to the conforming to designated values (Eder et al. 1995;Goodwin 2006). This occurs for a number of status markers such as race, class, and ability (Evaldsson 2005). The ways in which children often enforce these values are related to their evaluation and assessment of peers, using pejorative descriptions and negative categorisations of activities and people (Goodwin and Kyratzis 2014, pp. 516-517). The concerns of this paper are when it relates to gender. Girls' policing related to gender-typed behaviour includes the avoidance of fighting or the physical (Evaldsson 2007), while boys' policing relates to physical vulnerability, emotion weakness, cowardice, or asking for a teacher's intervention (Evaldsson 2002). Children police the talk and behaviour of same-sex peers in order to maintain gender-stereotypical behaviour. This may lead to greater expression of gender-stereotypical behaviour. Goodwin and Kyratzis (2014) describe what they call the high school heterosexual social market. This market describes that acquisition of social capital and social status within peer groups. Boys acquire this social status primarily through accomplishment in masculine domains, such as sports and overt competition. They accomplish masculinity through developing physically through sports and the appearance of strength, while also engaging in aggressive, competitive talk about masculine subjects. Conversely, girls become marginalised in these activities and dominate other spheres of activity and accomplishment, such as the technology of beauty and personality; experimenting with cosmetics, clothing, hairstyles, and the development of cute or clever personalities (p. 530). The association of sports and competition to masculinity pushes boys down a masculine pathway in order to access this social market at school. While the association of beauty and cute personalities to femininity pushes girls down this feminine pathway. This market perpetuates cycles of gender stereotypical behaviour for both girls and boys.
Sex segregation, through same-sex play or social pressure, increases gender stereotyping and therefore makes mixed-sex communication more difficult (Halpern et al. 2011). This is evident in research with single-sex schools, where girls tend to experience greater pressure to conform to gender roles than in coeducational schools (Drury et al. 2013). The implication in single-sex learning suggests that sex is an important category in learning and therefore the pressure to conform to that heightened value of sex in a single-sex setting increases (Howard et al. 2011). The Gender Segregation Cycle of Martin et al. (2012) suggests that the greater the time spent interacting with same-sex peers leads to an increase in gender-stereotypical thinking, fewer positive attitudes toward other genders, and more negative attitudes toward other genders. This then perpetuates same-gender interaction more. The concept of the Gender Segregation Cycles predicts that girls and boys will develop difficulties interacting with one another (Helgeson 2017, p. 282), which is supported by empirical research showing that there are greater difficulties in mixed-sex compared to same-sex groups in learning (DiDonato et al. 2014).

Self-esteem and confidence
Another important factor is that of self-esteem and self-confidence. Girls are less confident than boys during adolescence and this has been shown to limit their input during classroom discussions. Diseth et al. (2014) detail their study of 6-8 th grade students, describing that despite better academic performance girls had lower levels of self-esteem. A meta-analysis by Kling et al. (1999) demonstrates a difference across all ages (including adults) in self-esteem in favour of males, and most importantly, that this difference varies greatly by age, with adolescence being the time when the greatest gender gap in self-esteem was measured. The more recent meta-analysis of Zuckerman et al. (2016) found similar outcomes, with self-esteem in favour of males, and that the difference is greatest during adolescence. The comparatively lower levels of self-confidence girls display explains why boys perceive themselves as more capable, and therefore why their teachers may also perceive them in that way, thereby creating the conditions which allow boys to dominate classroom discussion.
Girls' self-confidence as related to their relationship to different activities and subjects is also interrelated with gender stereotypes. Määttä and Uusiautti (2020) assert that the decreases in confidence girls face as they enter adolescence is related to girls noticing conflicting expectations in their environment. This is especially so in girls' growing awareness that they are expected to possess feminine-associated traits, such as passivity, sensitivity, and altruism, and that they are not expected to demonstrate masculine-associated traits such as competitiveness and traditional notions of autonomy. Thus, not surprisingly, research shows that girls often underestimate their abilities, while boys often overestimate them, and, consequently girls worry more about their grades than do boys. This is despite the fact that girls do better than boys in school academically (Helgeson 2017, p. 234). Moreover, this phenomenon is exaggerated when the task is masculine-associated. For example, there are a number of studies which demonstrate that girls are less confident in STEM domains despite equal performance (Blanch et al. 2008;Else-Quest et al. 2013;Jansen et al. 2014;Muzzati and Agnoli 2007;Wach et al. 2015). This is because of the gendered nature of school subjects, whereby maths and science are seen as masculine because they espouse many of the masculine-associated traits, notably Reason, rationality, objective, and abstract notions of knowing, thinking, and understanding (Bleazby 2015;Kessels et al. 2014;Leaper et al. 2019;Yu et al. 2021). Girls become less interested as well as less confident in domains perceived to be masculine (Park et al. 2011;Teese 1998;Teese et al. 1995;Teese and Polesel 2003). Additionally, masculine perceptions of subjects have a direct influence on girls' achievement in those domains (Kessels et al. 2014;Leaper et al. 2019). This research suggests that girls may be even less likely to participate in classroom dialogues in traditionally masculine disciplines (e.g., physics) or about traditionally masculine topics (e.g., war).
Girl's lower self-confidence can inhibit risk taking or make them more cautious, which can impact participation in classroom dialogues. Girls have a tendency to self-consciously limit their response to conventional moves, playing it safe in classroom discussion rather than taking risks with their thinking and input because of lower confidence levels which infect their thinking with the idea that they do not want to be exposed as 'incorrect' (Glasser 2007, p. 219). Burgess (1990, p. 91) suggests that girls' achievement, self-esteem, and willingness to take an active role are particularly endangered in mixed-sex environments. This is supported by others who suggest that girls feel more comfortable taking risks and making mistakes in girlsonly discussions (Davies 2005). One reason for this may be due to girls' greater use of acknowledging moves compared to protest moves (i.e., greater use of supporting interlocuters in dialogue over disagreeing with them; Saunston 2007, p. 316). This can be seen in the research of Broughton (2002, p. 19). As she focusses on a particular adolescent girl in her research, she notes that the girl becomes quiet and loses her voice when among her male friends but feels free to act spontaneously without fear amongst her female friends. The student herself reflected that she had allowed herself to be constructed by others, and subsequently performed a public self that conformed to their expectations. Fisher (2014) also focusses on girls fitting a profile whereby they present as shy and reticent in the classroom but reveal loquacious personalities in interviews with the researcher. It was noted that no boys fit this profile, only girls in the study. These examples link to the idea of the classroom being a representation of the outside world. If that outside world has particular notions of what gender is for girls, they may be influenced in the classroom to perform in a way that meets that society's or classroom's expectations of gender.

Theories of gendered communication
The variability in communication based on gender in the above sections describes the difference but is not able to articulate reasons and solutions. This section will focus on theory that can help to explain the causes and solutions to a gender divide in communication.

Gender as performance
As noted in earlier sections, there has been a distinct move away from essentialising gender, which is to say that there has been a move away from regarding biological sex as the sole contributing factors in gender differences. Researchers such as Katz-Wish and Hyde (2014) and Marecek et al. (2004) would even argue that the study of gender is not worthwhile because it is impossible to be divorced from its context. What this all means is that gender should not be regarding as a static and unchanging quality of an individual, but a dynamic social construct (Helgeson 2017, p. 20). This means a move away from static quantitative research that describes girls as more empathetic and boys as more domineering. Instead, the focus will be on the theorisation of the factors that contribute to a statistically significantly greater number of girls displaying empathy than boys and vice versa for domineering behaviour. Additionally, how notions of empathy get tied to femininity and how notions of dominance get tied to masculinity. This can better help to explain why it is the case that when quantitative studies are conducted we often observe more girls displaying empathy and more boys displaying dominance (p. 20). What this comes down to is the idea that "gender does not reside inside a person but instead resides in our interactions with people" (p. 65). Researchers have applied this idea and now more commonly recognise gender as a continuum rather than a categorical variable (Carothers and Reis 2013).
Much of this theory is taken from the work of Butler (1990) who describes gender as something that one continually does, rather than someone one innately has. Similarly, Goodwin (1990) argues that activities, rather than cultures, groups, individuals, or gender, should be the basic unit of analysis in communication and interaction. "Gendered patterns in language variation emerge as a direct function of what gender means as a social category that is assigned to individuals in a society" (Meyerhoff 2014, p. 92). This is regarded as the performativity of gender: people perform gender in certain ways, for example, by communicating in ways that align them with either masculinity or femininity. Some authors, such as Constantinople (1973) and Spence (1985) have argued that all individual characteristics contribute to a bipolar continuum with masculinity and femininity at either end, and individuals are placed on this continuum to indicate their degree of masculinity or femininity (Spence & Buckner 2009, p. 115). Bem (1974) was early in making this claim, and further, that most men and women are not wholly masculine or feminine but fall somewhere in between this continuum. This means that the performance of masculine and feminine communication styles do not necessarily correlate to biological sex. However, as has been seen in the sections above, data suggests that more girls perform feminine communication styles, such as displaying care and empathy, and more boys perform masculine communication styles, such as displaying dominance and aggression. This means that it is not the case that masculine communication styles causing harmful effects on feminine communication styles translates absolutely and unequivocally to boys' communication styles causing harmful effects on girls. However, the empirical research discussed above does show that girls tend to use feminine styles more than boys, and vice versa. Therefore, in the current state of society, it is the case that when masculine communication styles dominate feminine ones, it tends to mean more of a disadvantage for girls than for boys.
Specifically within communication and language, Lazar (2014, p. 196) argues that gendered communication practices are drawn upon in everyday interaction to perform gendered identities. Cameron (2014, p. 293) also regards language and communication as a resource for the performance of identity. This affects how speakers choose to perform gender through their communication practices. Helgeson (2017, p. 181) notes that while speech communities can operate with opposing sets of norms, for example masculine and feminine communication norms, the communication norm associated with the socially dominant group tends to exert pressure on speakers to adopt this style, and this happens to an even greater degree in formal contexts. Therefore, the performance of gender in communication will have different effects on masculine compared to feminine gender communication. This leads into the next section on gender and power in communication if the power resides with the masculine.

Gender and power in communication
Power is a key element in the discussion of gender and communication. Boys tend to dominate discussion, be more assertive, and interrupt more than girls. How this plays into elements of power and gender is necessary to elaborate on. Foucault (1977) described the relations of power, showing how power is substantially discursive in nature. Power is embedded in all networks of relations (Lazar 2014, p. 144). However, there are ways in which gendered subjects are affected by power in various ways (Ramazanoglu 1993). Connell (1995) argued that the subjugation of women is a necessary feature of the development of masculinity. Therefore, the development of masculine communication styles is likely to disallow the expression of feminine communication styles. Gal (1995, p. 171) notes that language use is systematically related to persons, power, and the desirable moral order. If the cultural norm is that masculinity dominates femininity, the desirable moral order expressed through language is that girls should be silenced and boys should be dominant. Giles and Coupland (1991) explain how the style of one's language can affect how much influence they are able to exert over others. Beyond being in a genuinely authoritative or influential position, how a person uses language affects their perceived credibility as an authoritative or influential person. This is known as communicator credibility (Burgoon 1989). Hosman (1989) researched how the linguistic forms of hesitations and hedges (e.g., 'sort of' or 'kind of') lead to perceptions of low power low competence, reducing communicator credibility. When a person is a low-status speaker in a particular context, they tend to use more of these lowpower linguistic forms than otherwise. O' Barr (1981) researched inexpert witnesses in court, finding that low-status speakers tended to use more hedges, intensifies, tag questions, declaratives with rising intonation, polite forms, and hesitations. Highstatus speakers avoided these forms. These types of linguistic features are described as low-power or powerless speech (Bradac and Mulac 1984a, b;Erickson et al. 1978). These low-power linguistic forms are observed in female speakers more often than male speakers (Lakoff 2004;Leaper and Robnett 2011;Tannen 1990). This indicates that there is a perception that feminine speech styles are forms of powerless speech. The link between gender, power, and communication here is explicit, with the dominant masculine-speech-style group holding the power. West and Zimmerman (1983, p. 111) discuss what constitutes a dominant group. They describe how the data on males interrupting females is not merely an indicator of the power differential, but it is a way of 'doing' power; power is exerted through interruption. Beyond 'doing power,' it is also a way of 'doing gender,' insofar as the use of interruption based on gender implicates the use of power in what it means to be a man vis-à-vis a woman. Scott (1986Scott ( , p. 1069) describes gender as an axis for the organisation of inequality.
With children, Maccoby (1998) explains her argument for why divergent interaction and communication patterns exist based on gender. She claims it is a result of the greater strength of boys' playgroups compared to girls'. She says that boys are bound together more closely than girls and this leads to more exclusionary play where peer-group acceptance becomes the most important concern. Boys tend to engage in greater status-oriented communication because they have a greater need for recognition from other boys. She writes that "clearly, an essential element in becoming masculine is becoming not-feminine, while girls can be feminine without having to prove that they are not masculine" (p. 52).

Gender stereotypes and role conforming
The empirical section on gender stereotypes above describes how children face backlash for gender non-conforming behaviour. This section will focus on the theory behind why it tends to be the case that girls conform to feminine gender roles and boys conform the masculine ones.
Many theorists have attempted to explain the development of gender conforming behaviour and roles in children (e.g., Block 1973;Bussey 1983;Heatherton & Frankie 1967;Kohlberg 1966;Langlois & Downs 1980). Spence and Buckner (2009, p. 113) describe how many of these theories trace back to Freud (1925), in which he outlines the process of sex-role identification where children attempt to emulate or model themselves after their sex or a particular representative of it (such as a same-sex parent). Sex-role identification is responsible for the development of the full spectrum of masculine and feminine characteristics, beliefs, and behaviours that boys and girls are expected to exhibit.
Similarly, the concept of gender schema states that people tend to organise and process information according to the gender with which this information is stereotypically associated (Spence and Buckner 2009, p. 114). Bem (1981) articulates that gender schemas are applied differently by different individuals, and each varies in the degree to which they use gender schemata to organise and evaluate information. Masculine men and feminine women are said to be more gender schematised than others. This theory draws on the theory of sex-role identification in childhood, with variations in gender schematic thinking attributable to the strength of sex-role identification in childhood. Spence and Buckner (2009, p. 117) state that "those who are strong in their sex-role identification, it is presumed, engage in an active (though not necessarily conscious) effort to shape themselves in their characteristics, attitudes, and role behaviours to resemble the expected member of their own age and sex." Expected gender roles and behaviours are enacted and renewed in both institutions and social practices (Lazar 2014, p. 186), and these affect the gender schematic thinking of individuals. Connell (1987) argues that institutions are substantively structured in terms of gender ideology. This is demonstrated through analysis of empirical data in Holmes (2005Holmes ( , 2014, whereby she found that gender was always salient, even where it was seemingly not a primary variable, it always had the potential to create and reify acceptable roles and behaviour for women. Notably, this same potential for gender restrictive behaviour was absent for males in the study. Lazar (2014, p. 187) states that this kind of underlying gender pressure is pervasive of many institutions, cultures, and discourse, and both men and women contribute to their reproduction. Spence (1985) ties this to her theory of gender identity, where a psychological sense of maleness or femaleness is central to a person's self-concept. Then, as children gain increasing exposure to gender roles and stereotypes, they attempt to confirm their gender identity by adopting those roles and stereotypes, which tend to include the most visible stereotypes related to dress, toys, and behaviours. Connecting with both gender schemas and sex-role identification, the models in the child's social environment, such as parents, siblings, peers, teachers, as well as celebrity idols, influence a child's adoption of stereotypes related to gender identity and therefore children differ in the degree to which they resemble societal stereotypes (Spence and Buckner 2009, p. 120). Cognitive Development Theory supports the gender identity theory. Kohlberg (1966) states that once children acquire their gender identity, they are highly motivated to behave in ways that are perceived as consistent with that identity. There is also evidence that this process is the same for transgender children, as they are motivated to behave in alignment with the gender norms of the gender they identify with and in opposition to the sex they were assigned at birth (Olsen, Key, Eaton 2015;Tobin et al. 2010). Given the theorisations about children adopting gender conforming behaviours based on their gender identity, it will be important to look at how children acquire perceptions of gender stereotypical behaviour. Mulac and Bradac (2014, pp. 98-99) detail how research demonstrating a greater use of empathy by women than men reflects what is valued about them in the local contexts where the research was conducted. The display of empathy and cooperation with others reflect what is seen as appropriate for women more than men. She notes that the display of empathy may be used by women in order to establish authority and belonging as a person who is not gender nonconforming, thereby avoiding the backlash distributed to gender non-conforming people. In 1922, Jespersen (1990, p. 210) stated that "women exercise a great and universal influence on linguistic development through their instinctive shrinking from coarse and vulgar expressions and their preference for refined… veiled and indirect expressions." Cameron (2014, p. 282) asserts that this generalisation is a reflection on the role women were expected to play at the time. This is exaggerated by the word 'instinctive,' suggesting that these were inherently female traits and anyone who fails to display them is not wholly female. This gender schema may have influenced the development of girls at the time to feel pressure to be refined, veiled, indirect, and shrink from vulgar expressions. And indeed, the women who were already doing that would have acted as gender-typed role models for their female children to model themselves on. Sherzer (1987) suggests that in any community, gendered linguistic behaviour may be acquired through the force of a social norm, where women attempt to conform to the social norm for their gender or their behaviour is policed to enforce gender conformity.
Gender-role socialisation describes adults more directly socialising children to conform to gender roles through the environment, rewards, and models that shape a child's behaviour. The research of Whiting and Edwards (1988) on interpersonal behaviour found two key differences in gender: nurturance (helping) in favour of girls and egoistic dominance (coercion, competition) in favour of boys. They conclude that these differences were the result of the different environments in which girls and boys grow up. Montgomery (2008, p. 181) describes how it is often the case that girls are encouraged to be communal, express concern, and control their assertiveness, while boys are encouraged to be agentic, assertive, and control their expression of emotion. Pollack (2000Pollack ( , 2006 describes the pressure boys face to keep control of their emotions and maintain emotional distance from others, pressure that is caused by socialising boys to be tough, aggressive, dominant, and to restrict their emotions. This aspect of male gender-role socialisation has been linked to the expression of aggression towards the self and others (Feder et al. 2007). Timmers et al. (1998) suggest that gender-role socialisation may explain gender differences in the expression of emotions, as girls are brought up to be caring, deferential, and empathetic to others, this leads them to exhibit emotions that may strengthen relationships and inhibit emotions that could harm them.
Research on parents' behaviour as aligning with gender-role socialisation provides support for this theory. Parents have been shown to use more supportive speech with girls and more autonomy-supporting speech with boys (Tenenbaum and May 2014). They have also been shown to watch boys more while getting actively involved with girls more (Clearfield and Nelson 2006). This may suggest that parents believe that boys should be independent but girls require help (Helgeson 2017, p. 185). Parents tend to use the language of emotions more with girls than boys (Anzar and Tenenbaum 2015). They encourage boys to suppress their emotions (Levant and Rankin 2014), especially in the expression of sadness (Cassano and Zeman 2010). A meta-analysis suggests that boys are less aware of their emotions and have more trouble identifying and explaining them (Levant et al. 2009).
Alongside parents, peers also influence gender-role socialisation. As discussed in previous sections, both adults and children face backlash for gender non-conforming behaviour. In children, the prominence of same-sex play is thought to be a contributory factor in gender conforming behaviour (Golombok et al. 2008). Backlash against gender non-conforming behaviour may take the form of harassment and bullying. Gender conformity and same-sex friendship is associated with less harassment by peers, while gender non-conforming behaviour is associated with greater harassment from peers, and this is especially so for boys (Lee and Troop-Gordon 2011). Cameron (2014, p. 294) emphasises that while speakers have agency in the linguistic moves they make, their choices are constrained in many ways, including the gendered meanings which we ascribe to certain ways of speaking. These meanings detail what it is to be normal, and there is backlash for non-normality and non-conformity.

Gender non-conformity
Delineating the specifics of gender stereotypes in language and communication, Parsons and Bales (1955) differentiate between two types of roles. Instrumental roles exhibit goal-oriented behaviours aimed at accomplishing tasks, while expressive roles aim at promoting group harmony and cooperation. These roles would become linked with gender, and men assumed instrumental roles and women assumed expressive roles (Spence and Buckner 2009, p. 111). Evaldsson (2005) describes how fighting back and being able to take it are recognised to be commonplace to the experience of boys (cf. Danby and Baker 1998;Eder et al. 1995;Morgan 2002;Thorne 1993). Moreover, toughness (Smith 2010) and competition (Evaldsson 2002;Willis 1981) also fit this category.
On girls, Horner (1972) notes that competence, independence, and intellectual achievement were inconsistent with norms of femininity but consistent with the norms of masculinity. This could lead to a dilemma for girls in how they approach achievement. Helgeson (2017, p. 229) discusses why women might withdraw from achievement behaviour because they are concerned with conforming to the stereotypes of their gender identity. Fried-Buchalter (1997) researched this explicitly, confirming that girls demonstrated greater correlation between achievement and negative peer reactions, social isolation, and pressure to live up to others' expectations than boys. Bell (1996) found similar results, with high school girls declaring that achievement and affiliation were opposites, which she refers to as the smart versus social dilemma.
Moreover, the consequences for non-conformity are also severe and need to be taken into consideration when thinking about why it is the case that boys and girls (in practice) do, think, act in ways that are different. A recent experience this author had with a student confirms the pressures from the social world on gender conformity and the consequences for straying from the norm. A group of girls who left primary school after Year 6 and entered high school in Year 7 relayed some of the pressures they faced. One of the girls noted that the previous year, her group used to play running games with a large group, including boys. However, she stated that now they were in high school they no longer were able to do so, despite them all stating that they want to. The explicit reason given for this was the school uniform. In her primary school there was no school uniform and they could wear clothes of their choice, but in high school they had to wear a uniform which included a dress. They stated that they could not play running games in the school uniform dress. When queried about the full details of the school uniform, they noted that the school did, in fact, offer shorts for female students. However, the girls stated definitively that it was not socially acceptable for any of the girls at the school to wear the female uniformed shorts without backlash. This led these particular girls to all wear the school dress every day. This story is not an isolated occurrence, with research suggesting girls in Australian schools are less physically active when required to wear school uniforms (Norrish et al. 2012;Watson et al. 2015). A reading of this anecdote is that the girls were pressured to conform with the social norms of femininity associated with their gender, which meant not wearing the school shorts, and they chose to conform despite their explicitly expressed desire to continue to play running games at school. In essence, they were forced to choose between their desire to be physically active and their femininity. The pressure of conforming to the gender norms of their localised social community outweighed their other desires. This is an example of gender-role strain.
Gender-role strain is the concept of performing gender-role expectations that have negative consequences (Helgeson 2017, pp. 65-66). In previous sections the backlash effect of non-conformance has been discussed. This section will focus on the psychological theories that give rise to gender conformance. Pleck (1995) distinguishes between two different types of gender-role strain. Self-role discrepancy theory is about gender-role strain that arises when a person must live up to the gender role that society has constructed, for instance, being strong for men or being pretty for women. Socialised dysfunctional characteristic theory is about when the gender roles that society instils contain inherently dysfunctional personality characteristics. Helgeson (2017, p. 66) describes the example of a dysfunctional gender role for men being the inhibition of emotional expression, and for women being dependency, both of which are not healthy.
Pollack (1998,2000,2006) suggests that gender roles are more rigid for boys than girls. Boys' gender roles disallow the expression of vulnerability, and indeed, all things feminine. Oransky and Marecek (2009) explored this explicitly in their research where a theme emerged that boys avoided self-disclosure and displays of emotion for fear of lacking masculinity and being taunted by other boys. While this might initially appear to suggest that boys have it worse than girls, it must be taken into consideration that part of the more rigid gender role for boys is the dominance and disassociation of the feminine, which directly impacts girls. So even if it is the case that boys' gender roles are more rigid, this rigidity still negatively impacts girls. This negative impact is demonstrated by Coston and Kimmel (2012) who describe ways in which men respond to threats to status. They describe men exaggerating their masculinity, including becoming sexist and derogating women, as well as emphasising their bravery, physical strength, and aggression. Bosson et al. (2009) similarly describe how threats to masculinity lead to more aggressive behaviour.
For women, Gillespie and Eisler (1992) identified five areas of gender role strain: (1) fear of unemotional relationships; (2) fear of physical unattractiveness (3) fear of victimisation; (4) fear of behaving assertively; and (5) fear of not being nurturant. This strain is associated with depression (Helgeson 2017, p. 69). Parent and Moradi (2010) created the Conformity to Feminine Norms Inventory, detailing nine feminine norms (1) niceness, (2) importance of relationships, (3) thinness, (4) modesty, (5) being domestic, (6) caring for children, (7) romantic relationships, (8) sexual fidelity, and (9) investment in appearance. While these gender roles are broad, they have specific connections to communication and language. For women, niceness, modesty, being caring, and not being assertive all relate to the way women tend to use language. This pressure to conform affects the way women tend to use language. The same can be said of men, with aggression, derogation of women, and even bravery aligning with linguistic features that men tend to use more of including assertiveness, interrupting female partners, and being more open to challenge others' ideas.
Gender-role strain is closely connected with stereotype threat. This is the idea that when someone is threatened with a reduction in their gendered status they will perform worse. For example, a girl who perceives doing well in mathematics as being a threat to her femininity will perform worse in that subject. Walton and Spencer (2009) performed a meta-analysis on stereotype threat research, finding that stereotyped threatened groups do in fact perform worse than nonstereotyped groups, but the stereotyped groups' performance improves when that threat is removed, even on identical tasks, demonstrating that the stereotype threat is the salient factor, not the task itself. For example, a stereotype threatened group may be female students taking a science test and before the test it is stated that science is a masculine domain. Therefore, the female students perceive a threat to their femininity if they do well. Although these groups performed better when there was no threat, Cupach and Canary (2009, p. 233) described how some stereotypes are so pervasive that they can have an affect even when they are not explicitly acknowledged. Campbell and Collaer (2009) researched this phenomenon in a study comparing men and women in their visual-spatial skills. They found that even when they threat was not acknowledged, women still performed worse. However, when the threat was neutralised by a statement telling students that women and men usually perform at the same level on this task, the scores of women and men were similar. Stereotype threat can affect a variety of domains, including athletic performance for girls (Hively and El-Alayli 2014). In the academic domain, Fogliati and Bussey (2013) find that stereotype threat has adverse effects in maths for girls and they also disengaged from the task. This kind of disengagement may indicate that girls are less inclined to pursue domains, such as maths and science, when they perceive a stereotype threat to their femininity (Cheryan 2012;Helgeson 2017;Simpkins et al. 2005Simpkins et al. , 2006. In children, Trautner et al. (2005) describe three phases of stereotype development. Firstly, prior to age 5, children acquire information about gender-related characteristics. Up to 18 months of age, children begin showing preference for gender-stereotyped toys (Serbin et al. 2001). Between 18 and 24 months, girls link masculine and feminine toys with other boys and girls respectively (boys do this around 31 months) (Poulin-Dubois et al. 2002;Serbin et al. 2001). By age 3, girls are knowledgeable about stereotypes (O'Brien et al. 2000). The second phase, between ages 5 and 6, children rigidly apply this acquired information to sex. Helgeson (2017) states that children are more likely than adults to rely on target sex when making a judgement about a person. Martin and Ruble (2009) describe how children rely more on category-based expectancies (such as gender), rather than individualising each person. Thirdly, by age 8, children use individuating information rather than sex alone. However, this individuating information can also act to reinforce gender-role stereotypes. Helgeson (2017, p. 97) describes: This is, older children will infer that Hannah would like to climb trees rather than play with dolls because they see that Hannah dresses in jeans and a t-shirt [rather than a dress]. This is, older children will be less likely to rely on target sex to infer behaviour, but they use their knowledge of gender-role stereotypes to generalise from one aspect of gender-role behaviour to another. Older students take into consideration the individuating information but that individu-ating information comes from gender-role stereotypes. Beliefs about gender roles-masculinity and femininity-may be more rigid than beliefs about sex.

Conclusion
The results of this analysis demonstrate that there are many ways in which children's communication is gendered. These patterns of gendered communication are important considerations in assessing children's development as well as pedagogical strategies involving classroom discussion. Gender stereotypes play a role in the language, motivation, and performance of children, and the cost of non-conformity is high. In order to comprehend the phenomenological world of children, consideration of the gendered nature of that world and how they communicate within it is vital. Teachers ought to take heed of the ways in which classroom discussion can manifest in gendered ways, how children produce gendered ways of speaking and interacting, and how the teacher themselves can also do this. Explicitly acknowledging and being aware of the ways in which gender can manifest, and therefore discriminate, within the classroom and within classroom discussion will help to ensure that gender equitable pedagogical practices are being used.
In practice, the pedagogical work required of teachers is to be aware of the ways in which gender impacts within classroom discussion. For teachers, being of aware that if dialogues become competitive, argumentative, or aggressive, this privileges masculine ways of communicating and will likely marginalise many girls from the discussion. Being sensitive to the pedagogical moves that are inclusive to all students is necessary for equity. Ensuring that a supportive and collaborative discursive classroom is created is more inclusive. This collaborative discussion should still be critical, analytical, and full of disagreement, making it a productive academic dialogue. But it does not have to be competitive and argumentative. There are many pedagogical models that teachers can access to ensure this inclusivity, including Dialogic Instruction (Nystrand 1997), Dialogic Enquiry (Wells 1999 (Murris et al. 2016). Teachers should also be aware of those masculine characteristics, such as assertiveness, that they will more often see in boys and girls. This will help them understand that while many boys will assert themselves into the discussion, many girls will wait for the teacher to give them the floor. This difference says nothing in itself about the willingness of students to participate, but is reflective of dominant stereotypes and ideals of masculinity and femininity. And for most young people, they will prefer to align with the gender role associated with their sex. The way that students express their gender through action is a matter for pedagogical concern. Teachers should be accountable for how inclusive their teaching is to enable all students to participate, succeed, and excel.
Funding Open Access funding enabled and organized by CAUL and its Member Institutions.