Barriers to the Full Participation of Girls in Robotics: A Case Study of a South African Community of Practice

The lack of participation of girls in programming robotics is well documented in the international literature on the subject. The Inspired Towards Science, Engineering and Technology (I-SET) is a community engagement project that promotes science engagement in predominantly disadvantaged communities. In this paper, we present the findings of an action research project. The primary objective of this research project is to monitor and improve this community of practice. This research paper aims to explore whether girls encounter barriers to full participation in robotics competition teams. While educational parity has been reached in STEMI-related subjects at the school level, young women are far less likely to select STEMI fields at the tertiary level compared to young men. The I-SET community of practice is well-situated to promote STEMI fields to girls from an early age. Barriers to their full participation in the I-SET community of practice included male dominance in the everyday activities and speech acts of the teams; a lack of technohabitus; and alienation. Broader constructions of gender identity in schools and the communities that they serve presented a major challenge to coaches keen to actively include girls. These challenges were further compounded by gendered divisions of labour that left little time for girls (particularly those from poorer backgrounds) to practice.


Introduction
The World Economic Forum has predicted that as the fourth industrial revolution (4IR) unfolds, there will be a significant shift in employment patterns towards advanced robotics, artificial intelligence, advanced materials, biotechnology, and genomics (World Economic Forum, 2016).Jobs in these sectors require a strong educational background in science, technology, engineering, mathematics, and innovation (STEMI) related fields.Women make up only 28% of the STEMI-related workforce (United Nations Educational, Scientific and Cultural Organisation [UNESCO], n.d.) and there is a concern that as the 4IR takes root, women will be left behind (World Economic Forum, 2016).There is a clear need to promote STEMI-related education to build a habitus of technology and to prepare girls for STEMI studies at the tertiary level.STEMI-related activities such as robotics competitions motivate learners to explore and engage in STEMI from an early age.
The core business of the community engagement project called Inspired Towards Science, Engineering and Technology (I-SET) is to promote STEMI learning at all school levels to encourage learners to follow careers in this field.This is achieved through a range of STEMI activities that include supporting teams for robotics competitions.Noting international trends and local initiatives such as Technogirls, and the low number of girls selecting STEMI subjects at the tertiary level, I-SET embarked on a phased action science research project to explore and improve the experiences of girls in I-SET.The objective of this part of the project was to identify barriers to the full participation of girls in robotics teams from the perspective of robotics educators.We found that girls frequently do not contribute to the programming of robots and face several challenges to their full participation in robotics team communities of practice.

Research Context
The introduction of educational robotics at primary and secondary schooling levels establishes an interest in STEMI-related studies (Schina et al., 2020).Girls who participate in robotics activities at school are also more likely to study STEMI subjects at the tertiary level (Kucuk & Sisman, 2020).Robotics learning at school promotes critical thinking skills and lifelong learning (American Association of University Women Educational Foundation [AAUW], 2000; Kucuk and Sisman, 2020).Involvement in educational robotics also has a host of psycho-emotive outcomes including self-motivation, self-determination, and self-efficacy (Arís & Orcos, 2019;Aristawat et al., 2018).By participating in robotics and robotics competitions, learners are empowered to begin the process of becoming full citizens in e-culture (AAUW, 2000).
The performance of South African learners in terms of STEMI-related subjects reflects the legacies of an unequal apartheid education system.Numerous initiatives have been created to address this inequality.I-SET is a community engagement flagship project of the College of Science, Engineering and Technology at the University of South Africa (Unisa).I-SET is a community of practice that promotes and inspires science learning through the introduction of robotics activities at the school level in mainly disadvantaged contexts.The community includes educators, learners, students, and visionary community leaders.The core business of the I-SET project is to bolster student performance in maths and science in South Africa by providing opportunities to equip and support robotics coaches (educators, students, and community leaders) and teams, both in-person and online.I-SET has developed and maintains several robotics learning MOOCs as well as a related online practical course.Robotics activities are presented at science expos, workshops, and non-profit organisations (NPOs) to create awareness and equip learners.I-SET runs a mobile science lab to take the activities to schools that cannot travel especially in poorer and outlying areas.I-SET Robotics offers these robotics activities in several of South Africa's indigenous languages.A substantial portion of the work of I-SET is centred on the support of teams to enter a range of robotics competitions including the For Inspired Towards Science, Engineering and Technology (FIRST) Lego League competition, the World Robot Olympiad (WRO), Virtual GEAR competition, Conquesta Quiz and Robothon Robotics challenge.In addition, the "I-SET Buddies" is an internship programme for students.The students build competencies and experience through presenting and mentoring at I-SET robotics workshops.

Literature Review
While there is a multitude of STEMI skills and affective benefits associated with educational robotics, access to these advantages may be limited when it comes to girls and particularly girls from poorer backgrounds.Researchers have reiterated that, globally, there are significantly fewer girls than boys involved in robotics (AAUW, 2000;Kucuk and Sisman, 2020) and the roles that girls play within teams at robotics competitions tend to be those that do not involve programming skills (Schina et al., 2020).The different explanations for this can be grouped into two broad perspectives.The first is a pedagogical perspective which argues that girls have different learning styles to boys and that these should be considered in teaching praxis and material design.The other perspective is that broader social norms underpin gendered power dynamics that play out in the robotics context.
Proponents of the learning style model to explain the lack of girls in robotics argue that girls learn differently from boys and that the field of robotics education needs to be adjusted to accommodate these learning styles.Ardito et al. (2020) and Schina et al. (2020), for example, argue that girls learn more socially.They learn through collaboration and social interaction (Schina et al., 2020) and the dynamics of friendship are more important in robotics learning (Ardito et al., 2020).The second branch of this position relates to robotics materials.Hartmann et al. (2007) posit that one of the reasons girls are less attracted to robotics is because the equipment is more orientated to the creation of boy play interests.This argument is premised on the idea that boys tend to build car-style robots while girls tend to build animals and humanoid figures (Hartmann et al., 2007).The Roberta initiative, based in the European Union, is a programme that has a similar conceptual foundation.In this project, the basis of which is to improve the number of girls in robotics, themes and experiments are designed to embrace the interests of girls (Schina et al., 2020).
This learning styles and interests model harkens back to earlier work in feminist pedagogy that argued that curriculum content and teaching styles needed to shift to include girls and women (Belenky et al., 1986).Later feminist pedagogy, however, rejected these concepts because they engage a limited range of feminine identities (Hayes & Flannery, 2000;English & Irving, 2015).Girls and boys have a range of interests and learning styles (Richardson, 2015) and to construct the interests of either gender as having a particular focus, falls into an essentialist trap that potentially reinforces broader social constructions of gender.
For several authors, identifying girls as 'the problem' does not account for how gender norms influence the choices and behaviours of boys and girls in the field of technology (AAUW, 2000;Fuller et al., 2013).Zorn et al. (2007) for example, argue that the field of robotics cannot be divorced from broader social stereotypes.The problem lies rather in the decoupling of the broader social constructions (that play out in the pedagogic context) associating masculinity with technology.These gender stereotypes are normalised in the didactic situation through unconscious biases (Zorn et al., 2007) which are played out in microaggressions towards girls in different learning contexts.These include the power dynamics that are played out in group behaviours, informal jargon (Witherspoon et al., 2016), sneering (both overt and subtle), the domination of boys and the unwitting encouragement thereof.These broader social constructions of technological identities intensify as children grow older (Witherspoon et al., 2016), and these microaggressions in the educational setting are profound barriers to the full participation of girls in robotics.
Learning to build and programme robots in a team community of practice is a social process, but it is in this sociality that gendered power dynamics also play out.While the rhizomatic model of learning in a community of practice is empowering and effective, communities of practice are also entrenched in the power dynamics of broader society (Lave, 1993) and may serve to reproduce normative social dynamics.Communities of practice should function through the benevolent guidance of old-timers of newcomers into the learned identity of their field.Communities of practice, however, can be laden with exclusionary and conflictual dynamics (Fuller, 2007).Drawing on Bourdieu's notion of habitus, Mutch (2003) posits that social practice is largely established outside communities of practice and that, unless there are interventions, this social practice will be reproduced within the community.Particularly, where participants are of a similar social background, they may respond to the dynamics of the community in established patterns that may be alienating to newcomers.Several authors in the field of gender and robotics have engaged with this reading of habitus to theorise the differential engagement of boys and girls in robotics.Witherspoon et al. (2016), for example, argue that the backbone of robotics culture is informal opportunities for computing experience.This mundane knowledge frequently plays a limited role in the life worlds of many girls and reinforces gendered dynamics in robotics communities of practice.

Gender, School Performance and STEMI in Tertiary Education
South Africa is dealing with an unequal education system that was premised on ensuring the conformity and subservience of learners.The result of this system is a lack of educational resources and poorly trained educators.These features of the education system are further reinforced by poverty.The performance of South African learners in international benchmarking exercises such as the Trends in International Mathematics and Science Study (TIMMS) reflects this history and ongoing educational inequalities.Of the countries that participated in this benchmarking exercise, South Africa was ranked in the bottom three for both grades 4/5 and 8/9 in Mathematics and Science (coming in last for grade 8/9 Science) (see Table 1) (Mullis et al., 2020).The effects of poverty on performance in the subjects are demonstrated in an examination of the results between fee-paying schools in more affluent areas and non-fee-paying schools in poorer areas.The achievement gap in mathematics between these schools is 109 points at the grade 4/5 level and 80 points at the grade 8/9 level (Department of Basic Education, Republic of South Africa [DBE], 2020a, 2020b).For science, the gap is 150 points and 107 points for the respective grades (DBE, 2020a(DBE, , 2020b)).
While the scores and ranks in TIMMS 2019 are poor, they represent an improvement from previous iterations (DBE, 2020a, 2020b) with a very large improvement from the period directly after apartheid to the present.This improvement has, however, slowed down in recent years.
In terms of the exit level secondary schooling results, whilst the overall mathematics results improved, the physical science results worsened (Mzekandaba, 2022).The pass rate for mathematics increased from 65.8% to 2020 to 69% in 2021, whilst the physical science results decreased from 75.5% to 2020 to 74.7% in 2021 (DBE, 2022).To pass, candidates need to receive a mark of 40%, but to enter science-related subjects, most universities in South Africa require learners to attain at least 60%.In their report on the final year results, the Department of Basic Education (2022) indicates that the situation has improved in terms of the number of learners achieving this benchmark.
In 2015, the South African Department of Women reported that poor performance in maths and science appeared to be further compounded when it came to gender.In interrogating these statistics, researchers have subsequently argued that these figures do not consider the larger school dropout rate of boys (Spaull & Makaluza, 2019;van Brookhuizen & Spaull, 2017).Assuming that for the large part, male scholars who drop out of secondary school achieve lower marks, Spaull and Makaluza (2019) correct for this higher dropout of weaker students by removing the same amount of lower achieving girls from the cohort before analysing the data.These results show that, in 2018, there is a pro-girl advantage in Mathematics and Physical Science (Spaull & Makaluza, 2019).This trend was borne out in the 2021 results.Girl candidates attained 65.7% of distinctions, including distinctions in Mathematics, and Physical Science (Parliamentary Monitoring Group, 2022).An analysis of the 2019 TIMMS results also supports the research findings.The TIMMS statistics for South Africa show that in 2019, there was a marginal lead of girls in both subjects in grades 8/9 and a significant lead for girls in grades 4/5 (Mullis et al., 2020).Despite these girl positive indicators in Mathematics and Science, two worrisome key indicators persist.The first is the all-important 60% benchmark required for university entrance.While the mean results for girls show higher achievement, the median results vary between boys and girls with more boys achieving results at the higher and lower end of the marks range (Spaull & Makaluza, 2019;van Brookhuizen and Spaull, 2017).This means that more boys are achieving above sixty per cent in maths and science.The related second indicator concerns the selection of the field of study at the tertiary level and the gendered distribution of employees in STEMIrelated industries.Despite the pro-girl advantage in achievement across all subjects in the school exit assessments, and advantages in terms of human capital indicators (Mosomi, 2019), there are significantly fewer girls enrolling for STEMI-related subjects at the tertiary level (boys, for example, were 72% more likely to enrol in an Engineering degree in the year after matric) (Spaull & Makaluza, 2019).
In the 2021 World Economic Forum Global Gender Gap report, there are clear indications that educational attainment is very close to parity in South Africa, however, significant pro-male attainment in STEMS, particularly in the field of Engineering, Manufacturing and Construction, persists (see Table 2).Mosomi (2019) points out that there has been a widening of the gender wage gap in higher-earning professions and that this is primarily associated with the industry of employment.This is particularly true of the mining sector which frequently draws professional employees from the field of Engineering, Manufacturing and Construction.One of the core   concerns in closing this gender wage gap, then, is to ensure more women enter the workplace in STEM-related fields.For this, a habitus of STEMI-related activities needs to be established at all lower levels of education.
There are several initiatives to improve the quality of education and the quality of life of girls at the school level.The Department of Education has instituted the Adolescent Girls and Young Women (AGYW) Programme with the objectives of increasing retention in school, decreasing HIV incidence and teenage pregnancy, and increasing economic opportunities (DBE, 2021).They have also partnered with UNI-CEF in their Girls and Boys Education Movement (G/BEM) initiative.This initiative aims to "encourage girls and boys to work together as equals and to foster respect for the human dignity and rights of both sexes" (DBE, 2021, p. 4).The TechnoGirl Initiative is part of its gender empowerment programme that has an exclusive focus on science and the mentorship of girls in STEMI fields (DBE, 2021;UnicefUSA, n.d.).The I-SET robotics project is well-situated to add to these initiatives.

The I-SET Community of Practice
The work of the I-SET community of practice is sustained through a small group of salaried staff and the volunteerism of the community of practice members.At the core of the I-SET community of practice is a set of old-timers who have worked within the community from its inception.The old timers are made up of coaches and ex-team members (mentors).These experienced coaches and mentors guide newer members to the community of practice.Each team constitutes a community of practice and can be regarded as nested communities of practice within the broader I-SET community.The coaches are responsible for robotics teams that compete in robotics competitions.The learners that age out of competitions are encouraged to volunteer and return to the community as referees, judges, or mentors.Within the broader I-SET community of practice, the teams that participate in the First Lego League competitions, and newcomer coaches are encouraged and mentored by more established (old-timer) coaches.The I-SET Buddies are an integral part of the community of practice.The Buddies act as interlocutors of technology, presenting workshops (including indigenous language workshops), mentoring and providing support to coaches and teams.
Robotics lends itself to the overlapping fields of Vygotsky's (1978) situated learning, playful learning (Zosh et al., 2017), and problem-based learning (Conde et al., 2021).When the zone of proximal development (Vygotsky, 1978) is well-established and functional, the learner enters the field of the robotics community of practice as a newcomer and is apprenticed to old-timers in the community.Thus, a learner may enter a team peripherally and begin a rhizomatic process of learning how to programme through interacting with their coach and more established members of the team.The objective, and a mark of full participation in the community, is for learners to engage in all aspects of the competitions.The active learning of robotics provides an immediate feedback-oriented learning system, learners enter the field of robotics (and develop their programmer identities) in consultative and oftentimes playful ways.The coach acts as the facilitator of the different activities, but since there is feedback immediacy, the engagement of the learners with one another in problemsolving establishes their capacities.Learning is a process of identity formation, and this occurs as participation in the communities of practice deepens.As their programming skills become more established, learners move closer towards full participation in the programming work of the team, and their identity status as team members is concretised.Through this process, they also become programmers: an identity shift takes place through their interaction with the team community of practice.As the learners develop their identities as programmers, they are also establishing their identities as scientists.

Methodology: Action Science Research
Action research is an approach that solves a problem through the research process.It is frequently used in educational and healthcare contexts and involves "conducting systematic enquiries to help [practitioners] improve their own practices, which in turn can enhance their working environment and the working environments of those who are part of it…" (Koshy et al. 2011, p.2). Action research can be used in both positivist and qualitative research.Action science is an offshoot of traditional action research techniques.It is an approach to action research "which integrates practical problem-solving with theory building and change" (Friedman, 2011, p.131).Critical participatory action research (Kemmis et al., 2014;McTaggart et al., 2017) also provided a methodological frame for this project.The objectives of this approach to action research include collaborative and participative self-reflection and a commitment to social inclusivity.
This research is part of a broader action science research project on the communities of practice associated with the I-SET robotics project.The participation of girls in robotics emerged through a series of observations of robotics competitions and discussions with various robotics coaches.Within Kemmis and McTaggart's (1988) oft-quoted self-reflective spiral of action research, the action research process for the first phase of this part of the project was envisaged as a thorough initial reflection on the situation; participative processes to explore different solutions and build a plan for improvement; enact the plan and observe whether the interventions worked; reflect on the plan and move to the second phase of the action research to improve on the initial plan.In this paper, we explore the initial reflection on the situation and for this aspect of the research, we conducted interviews, focus groups and surveys with adult participants of the community.The data collected was analysed using the principles of constructivist grounded theory (Charmaz, 2014(Charmaz, , 2017a, b), b).Grounded theory has been successfully used in conjunction with action research in several studies (Canlas & Karpudewan, 2020;Duckles et al., 2019;Teram et al., 2005).Dicks ( 2007) describes action research and grounded theory as complementary methods which build upon one another, while Azulai (2021) points out the similarities between constructivist grounded theory and participative action research.
In a later edition of The Action Research Planner (Kemmis et al., 2014), the authors question the self-reflective spiral of action research and argue that the reality of action research is often not easily categorised in these different phases."The stages overlap, and initial plans quickly become obsolete in the light of learning from experience.In reality, the process is likely to be more fluid, open and responsive" (Kemmis et al., 2014, p. 18).Similarly, we found that the initial participative reflection on the theme of the participation of girls in the different activities of the teams had the outcome of building awareness and support for the issue of the very low involvement of girls in programming.The interviews and participative processes involved lively discussions on the issue as well as potential solutions and the educators came back to us with reflections of their own in their coaching practice.
The size of the I-SET community of practice is dynamic.There are about 150 coaches and mentors on the WhatsApp group.Each coach coaches at least 10 learners.Coaches, mentors, and aged-out adult team members were interviewed.Once the data gathered from these interviews had reached saturation in terms of the different themes, we developed an online questionnaire survey to gather further information on the perceptions of the members of the community of practice.This online questionnaire also served to triangulate the data from the interviews.We received 88 responses to the survey.The questions which pertained to gender were open-ended and these were once again coded in terms of the norms of constructivist grounded theory.We drafted an article based on the data and presented it in a workshop, to the participants of the I-SET community of practice, for member checking and to extend the discussion where appropriate.

Gender and the I-SET Community of Practice
The I-SET community has a good record in terms of the participation of girls in robotics teams participating in the FIRST Lego League robotics competition, which improved in the pre-COVID statistics (see Fig. 1).
These national participation statistics are aligned with the gender participation of the learners in the robotics team of the I-SET community of practice survey.In the survey, forty-two per cent (42%) of the respondents reported that their teams were made up of fifty per cent (50%) or more girls.There are currently limited numbers of all-girls schools participating in robotics.Seventeen per cent (17%) of the respondents reported that they ran all-girl robotics teams, primarily due to the placement of some of the teams in all-girls schools (see Table 3).Worryingly, a quarter of the teams were more than 80% boys, however, there are very few teams from all boys' schools.Several of the respondents indicated that they intended to grow their teams to a more equitable balance in the number of boys and girls in teams.
When asked what the challenges were to the participation of girls in the different robotics teams, we encountered different responses in the different methodological forums.In the online survey, only six of the participants indicated that there were none or that the challenges for the boys and girls were the same.
The challenges faced by girls are the same as the one for boys.(Survey respondent, 12/11/2020).No challenges at all they are equally enthusiastic.
(Survey respondent, 12/11/2020).The girls work very hard to learn and code for the robotics.
(Survey Respondent, 12/10/2020) In interviews, two coaches of all-boys teams responded that they had not thought about the challenges girls encounter in teams at all and that the question itself had made them rethink why girls had not joined their teams and that they would strategise to attract their participation in the future.
The rest of the respondents in the survey indicated that there were issues with the participation of girls in the teams.This was also found in the interviews and focus groups, where all the participants indicated concerns about gender in the teams that they coached or had participated in.Despite the positive shift in the number of girls participating in the robotics competitions supported by the I-SET community, the balance between the number of girls and boys is only a small part of the question of gender in the communities of practice of the teams.More important, is the identity work of the team.For example, there are two components to the Challenge Programme of the First Lego League robotics competition.The first aspect involves the design and programming of the robot, and the second is the research and presentation of a project to the judges.Before COVID, teams participated in live events.Teams prepared pit areas, had a competition name and identity, had pit-area decorations, and interacted with other teams participating in the competition.Part of the team identity included a war cry, cheerleading for their team and encouraging other teams.These three different aspects of the First Lego League robotics competition, however, present a gendered division of labour.The design and coding of the robot were dominated by boys and the research and cheerleading aspects were frequently done by girls.
They usually take a back seat when coding despite the push from a female coach.
(Survey Respondent, 21/03/2021) They often don't get a chance to build and program as they are seen as the "presentation writers" and "poster makers".
(Survey Respondent, 12/14/2020) They are seen as the support characters that cheer for the team.(Survey Respondent, 12/14/2020) Girls are more likely to learn solely the theoretical part but then coming to practical, they used to be hidden by boys [boys dominate] and [that] is indeed the main challenge.
(Survey Respondent, 11/28/2020) Girls think it is only for boys, or that they won't be able to do it.I.In mixed teams, boys usually dominate the girls, especially with the programming.
(Survey respondent, 12/11/2020) We asked the participants why they thought these differences existed.Utilising the techniques of thematic content analysis, we grouped their responses into three major themes.The first theme was centred on the perceptions of girls in the teams and habitus (their access to training and their previous engagement with robotics and games).The second set of responses highlighted the dominance of male voices in teams.The third theme engaged feelings of alienation, lack of recognition and a lack of role models.

Perceptions, Identity, and Habitus
In line with the findings of Hartmann et al. (2007) and Witherspoon et al. (2016), many of the participants understood the lower participation of girls in programming activities as a function of gendered social constructions.These responses were centred on issues of access and the habitus of technology.One of the participants argued that girls did not have access to opportunities for "building skills and engineering concepts" (Survey respondent, 12/15/2020).This lack of technological familiarity resulted in a reluctance to handle and work on programming the robots.
They fear technology or machines… Mostly programming because they think it's difficult.
(Survey Respondent, 12/10/2020) The majority are boys, the girls are a bit scared and think it is for the boys.
(Interviews, 09/05/2020) The participants expanded on the theme of the social construction of gendered roles, indicating that boys had access to, and are exposed to a range of technologies in their household from an early age.This access and exposure form a habitus of technology which explained the different capabilities and identity roles of girls and boys in teams.
[ For other participants, the lack of habitus went beyond access and exposure and included the somatic interaction with technology from an earlier age.Technological ease develops through play and tinkering with technology or facsimiles from early childhood.This leaves a technological imprint which often translates to a more intuitive engagement with technology at a later stage of childhood.
Most girls also haven't had practical play time with more technical components, as boys are more likely to have had.
(Survey Respondent, 12/14/2020) It's also a background where the girls come from.Culturally…they're not technical.For them…it's the first time we meet anything technical in the robotics.Usually, that's something that's reserved for the boys, even at home there's a technical thing, they'll ask the boy can you do that TV? Can you do the VCR, [fix] the phone?
(Participant Workshop, 27 January 2022) Several of the participants in the workshop also commented that there is a division of labour in many of the households of the girls where they were expected to assist with the care of younger siblings and household cleaning chores in the afternoons when they returned home.This division of labour means that there is little time for afterschool extension activities and most of the teams met outside of regular school hours.
One of the coaches lamented that she had lost several potential girl team members because they were expected to act as caretakers and complete other homecare tasks which drew them away from participation in teams.
One other barrier is gender roles.That then results in boys having more time on their hands than girls.Enough time to do more than women.Cos "our place is in the kitchen."(Participant Workshop, 27 January 2022) These gendered roles in the household not only impacted the amount of time that girls had for robotics, but they also shaped their identities in the microphysical dynamics of robotics teams.
What I have observed is that girls tend to be more interested in the building, tidying up and making sure that the pieces are sorted and ready at all times.But it essentially goes back to the cultural values instilled in us by our parents.(Participant Workshop, 27 January 2022) The effect of this division of labour and social norms governing femininity, along with the lack of technological habitus, translates into a lack of confidence and the belief that robotics was not part of the life worlds of girl children.
Lack of confidence as some believe that it is only boys who can do certain things such as coding.
(Survey Respondent, 12/10/2020) Lack of confidence in a male-dominated field.General lack of 'prettiness' associated with robotics.(Survey Respondent, 12/14/2020) The coaches that are committed to including girls in robotics teams and ensuring that they are involved in all the activities, experienced the challenges of the social construction of femininity and the attendant division of labour and roles in teams as major challenges.While the girls' involvement in research and team cheering were essential elements of the team, their lack of involvement in coding remained a problem that many struggled to address within the team communities of practice.These dynamics were compounded by class.Poorer families do not have access to cheap domestic labour as more affluent households in South Africa and, with broader constructions of gendered roles, girls play an important role in maintaining households.
Many of the coaches also felt that it was very difficult to push against the normalised gender roles which permeate school cultures.This cohort of coaches perceives the problem of the limited participation of girls in coding activities as having roots in a broader systemic issue which plays out in the school context.Without a broader interconnected systemic intervention at the scale of the school, other institutions, and the wider community, it is difficult to turn around these ossified norms at the scale of the robotics team.

Male Dominance
Paechter (2006, p. 14) makes sense of the concern that girls come into the robotics team communities of practice with an already formed gendered identity in relation to technology.She argues that gendered identities are learnt within a range of interrelated communities of practice, where "…we can see children as moving between successive age-related communities of masculinity and femininity practice while gradually becoming less peripheral members of wider, adult-centred gender communities".The robotics teams function as "local communities of practice" which are part of the constellation of communities of practice that inscribe and uphold gendered norms around technology.The coaches frequently commented that "boys tend to dominate the design and programming of the robots while girls present the research" (Survey respondent, 21/03/2021) and argued that this was both a function of the gendered norms that girls entered the teams with as peripheral participants and the gendered inscription of the identities of boys and girls within the team.Several research participants, including those who had competed in teams when they were younger, commented on the entitlement expressed in the speech acts and behaviour of boys.
The bigger louder male personalities make the girls retreat.(Survey Respondent, 21/03/2021) The boys got to choose first and the most preferential role in the team is the pro-grammer…It was quick as they could, and there was no opportunity or space for the girls to even take charge.So that happened….
(Participant 8, Focus Group) Girls and boys are sandwiched between the gendered norms learnt in broader communities of practice and the gendered norms within the communities of practice of the robotics teams.The dominance that the boys express in the teams is further complicated by the lack of technological habitus of the girls.
The language is sometimes condescending.Like "how come you don't know this, don't you game?" (Survey Respondent, 12/10/2020) The premise of many initiatives to attract more girls to robotics programming activities is that the attitudes of girls need to shift.Few initiatives focus on boys and the way that they interact with girls in teams.While there is a focus on teamwork, there is little theoretical (curriculum-based) engagement with the team members as to what constitutes good teamwork behaviour.This is rather a function of the apprenticeship of newer members of the team by the coach and established team members.While, for example, one of the core values of FIRST LEGO League is, "we respect each other and embrace our differences", without an explicit engagement, throughout the I-SET community of practice, with how to critically question the gendered dynamics of teams and how to shift these dynamics to ensure more inclusive teams, it is difficult for teams to develop an awareness and practice that is more inclusive of girls.

Alienation and lack of Recognition
In her discussion of the military as a gendered community of practice, Taber (2011) explains that old-timers define identities and competencies associated with full membership.Women, she argues, do not have "a clear trajectory to full participation" because the identity of full membership is premised on a set of masculine norms.Thus, the identities of women in the army are circumscribed and remain peripheral.
Similarly, several authors argue that the identity of the programmer is frequently portrayed as a male in media (AAUW, 2000;Hartmann et al., 2007;Zorn et al., 2007) and images of STEM-related competitions.While there are many initiatives to change the face of technology, the norms and signification of programming remain masculine.Many coaches observed that in this context, girls quickly became alienated within teams and spent their time in team practices either hanging back or simply leaving.

Conclusion
The objective of this research was to identify the parameters of the issue of gender and robotics within the I-SET community of practice.Many of the findings reflect those of the existing literature.The field is dominated by male voices.While girls are not disadvantaged in STEM-related subjects at school, girls feel alienated and mostly take up tasks that are not related to programming or leave the robotics communities of practice.One of the limitations of the I-SET community of practice is that most robotics activities currently take place outside of school hours.Established divisions of labour, which are compounded by issues of poverty, mean that girls have little time for extramural robotics teamwork.Since girls are far less likely to take up STEM-related subjects at the tertiary level, the work of I-SET and similar projects is to rethink how girls are included in their respective communities of practice.The discourse of gender and robotics learning in childhood is centred on a problematic relationship between girls and technology.The question of girls and their relationship to robotics, however, offers an opportunity for projects such as I-SET to rethink our practice.From the findings of our research, we believe that there needs to be a shift in perspective.Rather than identifying girls as the issue, we need to monitor and transform our communities of practice.Our community needs to start building a substantive strategy around how our communities of practice can function as spaces that, not only attract girls but actively promote gender equity.This strategy could dovetail with the objectives of initiatives like the Girl and Boys Education Movement.Many of the ISET coaches are also teachers in government schools and since there are plans to include robotics in the school curriculum during school time, I-SET is well situated to provide support for classroom-based robotics activities.Communities of practice can be exclusionary, but they can also act as transformative spaces for girls and boys.This requires a vision of robotics communities of practice as spaces that give expression to the idea of girls and boys working together as equals to promote a love of the field of STEMI.
Authors' Contribution Teresa Dirsuweit, the corresponding author, drew up the methodological approach, collected data, prepared the literature review, and prepared the article.Patricia Gouws, assisted in the collection of data and contributed to the preparation of the article.Both authors read and approved the final manuscript.
Funding No funding was received for this research.
Open access funding provided by University of South Africa.
Data Availability Requests to access datasets, including transcripts of interviews and focus groups may be made to the corresponding author.

Declaration of Interest
The authors declare that they have no competing interests.
Ethics Statement This research was approved by an accredited ethics committee, the UNISA College of Science, Engineering and Technology (CSET) Research and Ethics Committee.The reference number is 084/PG/2018/CSET_SOC.The principles of ethical research practices were adhered to and all participants indicated their informed consent to participate in the study.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/.

Table 2
Education and Skills Attainment in STEMI-related fields

Table 3
Average Percentage of girls in teams