Leadership practices contributing to STEM education success at three rural Australian schools

The limited research into leading STEM education in rural schools internationally tends to adopt a deficit view, with a focus on the poor achievement and aspirations of rural students, difficulties recruiting and retaining STEM teachers, and issues of isolation and under-resourcing. Counter to this trend, this paper reports on research investigating leadership practices shaping STEM education at three high STEM-performing rural schools. High-performing rural schools in Victoria, Australia were identified through analysis of state-wide final year enrolment and achievement data in STEM related senior subjects. Three rural schools with relatively high STEM subject enrolments and achievement levels were selected for in-depth study. The theory of Practice Architectures guided thematic analysis of interviews with principals, middle leaders, and teachers, facilitating a description of the ways that leadership practices interacted with the Practice Architectures evident at each school, which, in turn, enabled and constrained practices that contributed to each school’s STEM education success. Five leadership practices were identified as contributing to STEM education success at all three schools: leveraging community relationships, utilising local resources to enrich STEM learning, empowering STEM teaching staff, promoting the value of STEM education, and supporting STEM pathways. In detailing these leadership practices, this paper provides guidance to rural education leaders and policy makers seeking to improve STEM education in rural schools.


Introduction
Quality STEM (Science, Technology, Engineering and Mathematics) education is viewed as important to the future sustainability of natural environments, food and water supplies, and the global economy (UNESCO, 2015). Quality STEM education is particularly important in rural communities, given these communities are the custodians of the natural environments and the producers of resources that feed the populous and fuel the economy. Given this, it is concerning that students in rural schools in many countries have lower aspirations, participation, and achievement in STEM studies, compared to those in metropolitan regions (Echazarra & Radinger, 2019). Australian schools reflect this trend, with lower enrolments and achievement in all senior secondary science and mathematics subjects in non-metropolitan schools compared to metropolitan schools (Murphy, 2018(Murphy, , 2019. Despite this, there has been very little attention given to addressing issues in rural STEM education by the various Australian jurisdictions  and only limited research into effective STEM education in rural schools internationally. Many issues typically associated with rural STEM education are within the remit of school leaders. The limited research published on rural STEM education suggests that, internationally, rural schools struggle to: recruit appropriately trained STEM teachers (Echazarra & Radinger, 2019;Weldon, 2016), offer advanced senior STEM subjects (Lavalley, 2018;Murphy, 2018Murphy, , 2019, provide staff with appropriate professional learning in STEM (Lavalley, 2018), and access STEM education resources (Echazarra & Radinger, 2019;Lyons et al., 2006). There is a dearth of literature exploring how leaders in rural schools can manage these challenges and contribute to STEM education success. This paper responds to this paucity, presenting an analysis of the leadership practices contributing to the high STEM performance at three Australian rural schools and addressing the question: What leadership practices can contribute to STEM education success in rural schools?

Background
This study adopted a broad definition of STEM education, one that includes traditional, disciplinary bounded education in science, mathematics, technologies and engineering, through to programmes that integrate those STEM disciplines to engage students in authentic real-world inquiry (Bybee, 2013;Murphy et al., 2020). This definition is a pragmatic one, reflecting the typical use of STEM education to refer to a broad range of STEM related educational activities (Bybee, 2013). Further, it reflects the reality of Australian schools, where STEM education is typically taught in distinct disciplines (Marginson et al., 2013) in a policy context that lacks a clear position on the role of integration in STEM education .

3
Leadership practices contributing to STEM education success…

STEM education in rural schools
Internationally, rural students underperform in science (Echazarra & Radinger, 2019;Nissinen et al., 2018) and mathematics (Luschei & Fagioli, 2016) compared to students attending metropolitan schools. Rural students have lower participation rates in advanced mathematics (Irvin et al., 2017) and tend to have lower aspirations in tertiary STEM studies, even when accounting for SES (Echazarra & Radinger, 2019;Nissinen et al., 2018). These trends are also seen in Australia. Rural students are outperformed by metropolitan students in both mathematics and science testing (Thomson et al., 2019). Rural Australian students also have lower participation rates in senior advanced mathematics subjects (Murphy, 2019), and in chemistry and physics (Murphy, 2018). Finally, Australian rural students generally have lower academic aspirations (Fray et al., 2020), and are less likely to see science as important to their future careers (Lyons & Quinn, 2010).
The relative underperformance of rural students in STEM subjects internationally is commonly attributed to a range of factors. Rural students are less likely than metropolitan students to have access to advanced mathematics and other STEM subjects (Lavalley, 2018). They also have poorer access to extracurricular STEM enrichment activities (Echazarra & Radinger, 2019) known to develop aspirations in STEM careers (Franz-Odendaal et al., 2016). Relatedly, there is a need for improved STEM career education for rural students (Franz-Odendaal et al., 2016). Rural students also have poorer access to staff qualified to teach advanced mathematics and science, and restricted access to STEM education resources (Echazarra & Radinger, 2019). Rural schools have difficulties recruiting STEM teachers and providing them with professional development in STEM education (Lavalley, 2018). These factors that impact rural students internationally similarly impact Australian rural students. Australian non-metropolitan schools are less likely to offer chemistry, physics, or advanced mathematics (Murphy, 2018(Murphy, , 2019. Where they do, they are more likely to offer these subjects as multi-grade classes, or via distance education, both modes generally viewed unfavourably (Lyons et al., 2006). Rural Australian schools are more likely to staff science and mathematics classes with unqualified teachers (Weldon, 2016), and encounter difficulties providing appropriate professional development and mentoring for their STEM teachers (Goodpaster et al, 2012;Lyons et al., 2006). Finally, rural Australian STEM teachers report having poor access to resources and learning enrichment opportunities for their students (Lyons et al., 2006).

Leadership in rural schools
Many of the issues confronting rural STEM educators confront rural school leaders more generally. A review of international research found that rural school leaders encounter difficulties with recruitment, often resort to staffing classes with out-offield teachers and have difficulties organising high quality professional development (Hardwick-Franco, 2019). Economic hardship, poor resourcing, limited subject offerings, and small, composite classes are also common in rural schools 1 3 (Hardwick-Franco, 2019;Mendiola et al., 2019). Sullivan et al. (2013) found similar challenges face rural school leaders in Australian schools, highlighting the negative impact of shortages in qualified staff and instructional materials.
Internationally, there is relatively limited research into effective leadership in rural schools (Hardwick-Franco, 2019). A review of this research emphasised the importance of school leaders building relationships and working with all members of the school community to support effective instructional practices (Preston & Barnes, 2017). The way leaders work with their staff has been found to be particularly important; effective rural school principals are aware of teachers' needs and empower teachers to innovate and take strategic risks (Mendiola et al., 2019;Preston & Barnes, 2017). Finally, successful rural school leaders utilise place-based education to enrich learning, and they nurture connection to the local community and environment (Mendiola et al., 2019).

Leadership for STEM education success in rural schools
Literature addressing effective leadership of STEM education in rural schools is scarce. Much of what is available explores practices related to rural school leaders' responsibilities, without significant investigation of leadership practices. Research into effective professional learning in mathematics and science in rural contexts suggests that it should respond to local needs and be teacher led (Albion & Spence, 2013;Jorgensen, 2016). There is some evidence that rural environments, communities, and industries can be used to enhance rural STEM education programmes (Avery, 2013;Ebenezer et al., 2018). Relatedly, local resources can be used to build rural students' STEM related aspirations (Peterson et al., 2015). Some authors suggest close community relationships with rural communities can contribute to student STEM performance (Echazarra & Radinger, 2019) and staff retention (Goodpaster et al., 2012).

Conceptual framework
To identify and comprehensively describe leadership practices contributing to rural school STEM education success, a conceptual lens informed by the theory of Practice Architecture (Grootenboer, 2018;Kemmis & Grootenboer, 2008) was used. Through this lens, practices are seen as socially identifiable activities comprising three generic characteristics of practices: sayings (characteristic discourses), doings (distinctive actions), and relatings (characteristic interactions between individuals and groups). These sayings, doings and relatings are enabled and constrained by three corresponding sets of arrangements (Grootenboer, 2018): cultural-discursive arrangements (such as traditions, jargon and forms of communication), material-economic arrangements (such as resources, products, and processes), and social-political arrangements (such as relationships, networks and power arrangements). Practices are not merely shaped by these arrangements, they also impact the form of arrangements (Kemmis & Grootenboer, 2008). For example, sayings can 1 3 Leadership practices contributing to STEM education success… become traditions, doings can have material impacts, and relatings can alter social networks. The relationship between the generic practices and their related arrangements is shown in Fig. 1.
Schools have interacting levels of practice, where the practices at one level shape the arrangements that then enable and constrain the practices at the next (Grootenboer, 2018). The focus of this study was the leadership practices that shaped the arrangements that in-turn shaped the successful STEM education programmes at each school.

Method
This paper uses data drawn from a larger research programme exploring practices contributing to STEM education success in rural schools (Murphy, 2020a(Murphy, , 2020b(Murphy, , 2021.

Case selection
Three rural schools, identified as high-performing in STEM, and called here Sweeping Plains School (SPS), River Valley School (RVS), and Highlands Secondary School (HSS) were studied. Table 1 shows that in most cases these schools exceeded, and in many cases dramatically exceeded, the mean study score of nonmetropolitan schools in the Year 12 STEM subjects offered by the schools. Further, these schools exceeded the mean enrolment proportions of non-metropolitan schools in many of the Year 12 STEM subjects offered by the schools. Of note, all case schools had relatively high enrolment proportions in the advanced mathematics subjects (Mathematical Methods and Specialist Mathematics), and in Chemistry and Physics. These data suggest that these three schools have been successful in attracting their students to enrol in senior STEM subjects and in supporting high achievement in these subjects.

Generic practice
Arrangements comprising practice architectures Doings: distinctive actions Material-economic arrangements: resources, products, and processes.

Sayings: characteristic discourses
Cultural-discursive arrangements: the traditions, jargon, and forms of communication.

Relatings: characteristic interactions
Social-political arrangements: the social networks and power arrangements.
Note: Adapted from "Situating Praxis in Practice" by S. Kemmis and P. Grootenboer, 2008, Fig. 1 The relationships between generic practices and the arrangements comprising practice architectures 1 3 Leadership practices contributing to STEM education success… All three case schools were in different rural environments, as indicated by their pseudonyms, and were all more than 200 km from the nearest major city, and more than 80 km from the nearest regional city. They each had a relatively small number of students, and the students in each school had average to somewhat below-average socio-educational backgrounds, as shown in Table 2.

Participants
All three principals of the schools were interviewed, as were the STEM middle leaders and the STEM staff. Interview participants are listed by their pseudonym and role at each school in Table 3. Participation rates in interviews at each site were 46% or higher, with a good distribution of participants by gender, teaching area, and leadership and teaching roles. STEM teachers were also given the opportunity to participate in an anonymous survey.

Data collection
STEM teachers, leaders, and principals were interviewed individually where possible; however, interviews were conducted in pairs (on two occasions) and in groups (on two occasions) in response to participant availability, as indicated in Table 3. Where multiple participants were interviewed at the same time the interviewer invited each participant to respond to each question and checked for differences of opinion to limit the risk of one participant dominating the interview or 'group think' (Cohen et al., 2011). Further, an anonymous survey completed after these interviews provided an opportunity for participants to share ideas they may have forgotten or felt unwilling to share in the interviews. Semi-structured interviews commenced with these broad questions: "What do you feel are the largest contributors to student engagement in STEM at your school? What about the contributors to student achievement?" This form of questioning allows for themes that arise to be participant driven, rather than researcher limited (Gideon & Moskos, 2012). Volunteer sampling was used for the post-interview online survey, with principals distributing the link to STEM teachers via email. The number of participants in the survey are listed in Table 4. The survey included open-ended questions about the contribution Table 2 Schools selected to participate in Phase Two a Index of Community Socio-Educational Advantage (ICSEA). ICSEA calculates a school's level of educational advantage using parental education and occupation, as well as the school's location and proportion of Indigenous students (ACARA, 2015). The national average ICSEA is 1000. The ICSEA figures listed for each school are rounded to the nearest ten to prevent potential identification. Finally, school documentation including timetables, subject selection booklets, newsletters, and strategic plans, were collected providing further evidence of school communications, resourcing, and processes.

Data analysis
Informed by the conceptual framework, data from interviews, the survey, and school documents were searched for descriptions of sayings, doings, and relatings of school leaders that may have impacted the arrangements that shaped STEM education practices and ultimately led to high STEM education performance. These data were then subjected to thematic analysis (Clarke & Braun, 2017), where iterative engagement with these data revealed five emergent themes associated with the leadership practices at each school. Data triangulation was employed, contributing to the credibility of the findings (Yin, 2014); practices were described by multiple participants at each school and themes were observed across all three schools.

Results
Analysis of the leadership practices contributing to the STEM success of each of the three high STEM-performing schools in this study revealed five themes: Leveraging community relationships, Utilising local resources to enrich STEM learning, Empowering STEM teachers, Promoting the value of STEM education, and Supporting STEM pathways.

Leveraging community relationships
Leadership practices in the case schools contributed to establishing socio-political arrangements within the local community and enabled the STEM education practices of the schools. For example, SPS leaders' maintenance of relationships with six nearby schools contributed to the school's STEM success. These schools shared technical training resources that supported the delivery of a range of STEM related vocational courses to students in Grades 9-12. The network of schools also shared a staff member who coordinated the schools' careers education and workplace learning programmes. Further, the network schools supported delivery of senior subjects 1 3 via video conferencing. Kevin said, "It has been very successful, largely because, as a network of schools, we only put in front of the class staff that are strong teachers." (Principal, SPS). Finally, these relationships also supported teacher professional learning, where schools in the network shared access to external experts running professional learning workshops at their school. RVS leaders had limited networks with other schools, but extensive networks with community groups contributed to the quality of STEM education at the school. Local paramedics ran incursions at the school about anatomy, physiology, and emergency first-aid. Grade 10 students participated in school-based fire-fighting training with the local branch of the Country Fire Authority. Janet described other networking that she felt contributed to students' STEM aspirations: We've been trying to invite people in constantly. So, [local water authority] come in and they work with the kids [and] they go out and have field days … We've had… three excursions going to the [local hydroelectric facility] … The kids find it really interesting but it also broadens their horizon about what they can do. (Principal, RVS) Less directly, school leaders at the RVS used relationships with local organisations to support professional learning. Veronica (Mathematics leader) maintained contact with the local Land Care and water authorities to enhance her understanding of the local environment, knowledge that she brings into the classroom. Community relationships also impacted STEM staffing practices. RVS school leaders used local networks to find employment for the partners of new recruits and to settle their children into school, to overcome recruitment and retention issues. In one instance, the school altered processes, changing the timetable so that a local barista could work part-time as a senior mathematics teacher.
HSS leaders also leveraged community relationships to enhance their STEM education programme. They worked with the local water authority, alpine resorts, airport, and dairy industry to enhance their STEM education programme. Further, local groups sponsored students to attend STEM excursions to Melbourne and Canberra, supporting enriched STEM learning. Close community relationships also supported STEM staffing; recently retired STEM teachers were employed as casual replacement teachers in STEM classrooms and consequently provided mentoring for their less experienced colleagues.

Utilising local resources to enrich STEM learning experiences
Leaders at all three schools expanded the material-economic arrangements of their school through encouraging the use of local resources to enable a range of STEM education practices. The networking of leaders at SPS provided students access to a range of STEM resources, such as laser-cutters, 3-D printers, digital lathes, medical simulation equipment, greenhouses and other horticultural equipment, animal husbandry tools, and automotive tools. The resources available to RVS and HSS was not as extensive, with Janet, principal of RVS, describing the school's access to digital technologies as "woeful", and Tracy, the Laboratory Manager at HSS noting, "We don't have anything fancy". Despite this, both schools provided rich learning experiences by using resources available in the local environment. The large rural grounds and local environment at RVS meant that leaders could support the establishment of a mock body-farm to study putrefaction, the raising of chicks to study measurement and data, and the housing of cattle as part of the Cows Create Careers 1 (Dairy Australia, 2019) programme. HSS's STEM programme also made good use of local resources. Indeed, Jirra felt the school had developed a reputation for it, saying, "We try and utilise our community as much as possible. Sometimes you get visiting teachers or visiting people [saying] 'you're almost like a private school with a public-school name', in that we are lucky to have what we have" (Mathematics and Science leader, HSS).

Empowering STEM teaching staff
The leadership at all three schools established socio-political arrangements that empowered their STEM teachers to develop programmes that met local student needs and to drive their own professional learning. Kevin fostered a tradition of highly valuing professional learning at SPS, saying, "We encourage our staff to undertake a lot of PD [professional development], particularly senior teachers, maths teachers, et cetera." (Principal, SPS). The STEM teachers affirmed this approach, with one teacher commenting, "When asked for the opportunity to advance teacher knowledge, the administration has always provided time." (Teacher survey 5, SPS). STEM teachers also felt that the school leaders had contributed to the school's STEM success through communication support for, and resourcing, new initiatives. One teacher said, "Leaders encourage staff to try new things. Willing to finance initiatives" (Teacher survey 3, SPS). Practices at RVS had cultivated a collaborative, self-reliant approach to professional learning, where power arrangements allowed the STEM team to seek professional learning in response to local needs. Stuart, explained, "Instead of waiting for someone to come to us and do something, we would actually go out and seek information on the web and then we'd say, 'Okay, let's do that'" (Science leader, RVS). Leadership practices were felt to nurture this approach: STEM teachers said school leaders' communications encouraged staff to "try new things" and "do activities out of the box". This has led to staff being proud of their record as local innovators. Veronica said, "We do a lot of things that we're the first in the area" (Mathematics leader, RVS). A similarly empowered, collegial approach was evident at HSS, with one STEM teacher relatively new to the school commenting, "When I came here, I just got this impression that the staff worked really well together" (Britta, Teacher, HSS). School leaders saw collegiate support as key to overcoming limited professional learning resources. Jirra said, "We don't have a massive professional development bucket, because we are a small school… 1 3 The younger ones have stepped up and have obviously learnt from the older ones" (Mathematics and Science leader, HSS).

Promoting the value of STEM education
Practices of the leaders at all three schools contributed to arrangements that established the importance of engaging and achieving in STEM. At SPS, there was an established tradition of celebrating mathematics effort and achievement at school assemblies and students were encouraged to participate in external STEM activities. Karen's communications through a regular mathematics column in the school newsletter reinforced the value of mathematics, "We get sent careers in maths stuff and it's got examples of how people are using mathematics in their careers, so I just publish those on the parent bulletin." (Numeracy leader, SPS). The wider careers programme was also viewed as contributing to the school's STEM success. Kevin explained, "That really, I think, supports students when they're making choices for subjects … there's great knowledge there and recommendation regarding the need for your maths subjects" (Principal, SPS). Leadership had dedicated significant time and resources to this programme, with all students from Grade 7 onwards participating in weekly activities about careers and future pathways.
The careers programme at RVS, which commenced in Grade 8, was similarly viewed as key to STEM education success. RVS leadership also resourced a range of special STEM events that raised the profile of STEM. Janet, the principal, described the annual STEM days, "The STEM days that are set up are huge… Kids will organise all the events and the activities and, you know, the problem solving around it" (Principal, RVS). STEM evenings were also held for families: Stuart said, "40, 50 people come along to those sort of nights … So over the years … we've done lots of hands-on activities and evening activities to get them interested" (Science leader, RVS).
HSS leaders also promoted STEM through their communication and resourcing decisions. Sally said that the previous principal "firmly believed that engaging kids with pracs [practical activities] will build a love [and] thirst for those subjects" (Principal, HSS). The STEM teachers felt this ethos was perpetuated by the current leadership, "principal and assistant principal both have teaching backgrounds in STEM subjects, so are keen to see them flourish," (Teacher survey 1, HSS). Leaders at HSS further raised the value of STEM by supporting extensive extracurricular offerings, including visits to a specialist science school and a life sciences learning centre in Melbourne. Jirra felt it was worth trying to overcome the challenges to access these opportunities, despite HSS's rural location. She said, "We try and support those as much as possible… to give our kids the best opportunities and experience they can, but obviously we can't give them everything… we try and cover the costs through clever efficiency" (Mathematics and Science leader, HSS). These efforts to promote the value of STEM education were seen to have contributed to a tradition of high expectations among staff and students. Britta described this: I think there's certainly a culture of high expectations that's been built up, obviously over a number of years. The fact that kids do well is more than accepted, it's actually looked upon as that's really good that you do well. Kids that choose mathematics methods are not looked upon as being the smart kids, or the nerds, for want of a better term. (Teacher, HSS)

Supporting STEM pathways
Leaders at each of the schools established some unconventional processes and arrangements to ensure that STEM pathways were maintained and accessible to students across the school. The SPS timetable showed that mathematics learning was privileged over science, with 287 h of mathematics and only 93 h of science timetabled for Grade 7 to 9 students, whereas the average for an Australian Grade 8 student is 139 h and 126 h per year respectively (Thomson et al., 2017). Students were also encouraged to accelerate into senior mathematics, resulting in many students in Grade 10 studying Grade 11 level mathematics. Further, all Grade 9 and 10 students spent one day per week studying vocational STEM subjects as outlined in subject handbooks, including Agriculture, Allied Health, Animal Studies, Automotive, Building and Construction, Engineering, Digital Games Design, Kitchen Operations, and Textiles. Finally, leaders supported a wide array of senior academic STEM pathways by collaborating with schools nearby to run classes online. Kevin, the principal, had adopted creative measures to resource the STEM programme with appropriate staff. Unable to recruit appropriately qualified staff to the district, he retrained an Art teacher in digital technologies, and a local carpenter as a design technologies teacher.
RVS's junior timetable was similarly skewed towards mathematics, with students in Grades 7 and 8 studying Mathematics for approximately 167 h compared to 100 h of Science. Subject handbooks show that in Grades 9 and 10, science and technologies were offered as part of an elective programme, where students studied in multi-grade classes and arranged to support high-performing students to accelerate into senior STEM subjects including Chemistry, Engineering (Certificate II), and Psychology. RVS's leadership also protected the breadth of senior school STEM offerings. Physics and Chemistry was offered despite very low numbers, sometimes as multi-grade classes. Further, as previously mentioned, school leaders rearranged school timetabling to recruit a local barista as a senior advanced Mathematics teacher.
Analysis of the HSS timetable and subject handbooks revealed that leaders had not skewed their junior curriculum towards mathematics but had adopted a process that vertically organised an elective system from Grades 8 to 10 that had other unusual characteristics. Sally, the principal, believed this system contributed to the school's STEM success, "[The] elective system allows students to have some choice in their subjects, so they can pick things that interest them including in STEM sub-jects… STEM subjects are hands-on and have engaging activities" (Principal, HSS). The STEM subjects were theme based (for example, "Environmental engineering", "Forensics and psychology", and "Medical science") and students worked in multiage groups. Jirra noted that science subjects were well subscribed, "Most of our science classes have 25 kids in them… if you look at their first and second preferences, science is often first" (Mathematics and Science leader, HSS). The elective structure and processes and communications around it encouraged students to accelerate to higher levels from Grade 8 onwards. Jirra said, "We encourage kids to accelerate into classes if they would like. So, for example… I have two accelerating Year 10 s tackling Year 11 [advanced Mathematics], and they are flying, they're really loving it" (Mathematics and Science leader, HSS). Like the other schools, HSS worked hard to maintain strong senior STEM subject offerings. For example, Jirra said, We have [Grade] 12 Chemistry, which has four students in it. Now that's very small, but we thought well it's such a key part for these kids and what they want to do, we would be doing them a disservice if we didn't try and support it. (Mathematics and Science leader, HSS) Peter, a teacher relatively new to the school commented on the success of the school in maintaining these offerings, "It astounds me that a school this size has got Physics at both year levels, Chemistry, Biology and a Psychology, Specialist Mathematics, [Mathematical] Methods all running. It's amazing" (Teacher, HSS).

Discussion
This paper presents an analysis of the school leadership practices contributing to high STEM performance at three rural Victorian schools. Through thematic analysis of interview and survey data, five leadership practice themes contributing to the schools' STEM success were identified: Leveraging community relationships, utilising local resources, empowering STEM teachers, promoting the value of STEM education, and supporting STEM pathways. The findings from this paper strengthen the limited existing literature exploring rural school leadership in STEM education, while also highlighting the potential strengths that rural school leaders can capitalise on in STEM education.
The school leaders in this study confronted challenges common for rural leaders and in rural STEM education, yet they were able to successfully manage these challenges. These school leaders faced difficulties recruiting qualified STEM teachers, as is typical for rural school leaders (Lavalley, 2018;Weldon, 2016). As is also typical in rural schools (Lavalley, 2018), the leaders in this study had difficulty accessing quality professional development for STEM teachers. However, with flexible resource management and the use of local networks, these leaders ensured that STEM classrooms were appropriately staffed and that STEM teachers were professionally supported. Leaders in this study collaborated with other schools to share teaching expertise, worked with local businesses in efforts to both recruit and retain STEM teachers, and/or supported the training of locals as STEM teachers. The leaders in this study also empowered their STEM teachers to drive their own professional learning. A culture of trust, support and calculated risk-taking was nurtured by leaders at the schools, aligning with the findings of Mendiola et al. (2019). Aligning with the literature reviewed by Preston and Barnes (2017), the leaders also encouraged sharing of expertise between staff and accessed additional expertise through local community networks.
Low rural student aspirations in STEM (Echazarra & Radinger, 2019) threaten the sustainability of the senior STEM subjects in many rural schools. The leaders in this study established material-economic and cultural-discursive arrangements that supported the nurturing of student STEM aspirations, ensuring adequate student enrolments in senior STEM subjects. Careers programmes beginning in early secondary school were believed to be key contributors to STEM success at each school. This belief aligns with research suggesting early careers education is important if rural students are to pursue pathways to higher education generally (Dollinger et al., 2021). Leaders profiled and promoted STEM in various ways at each school, establishing the value of STEM education across the school community. Junior STEM programmes were arranged in ways that nurtured student STEM engagement and provided opportunities for accelerated learning in STEM, further developing the senior STEM enrolments pipeline. Even with these efforts, senior STEM class sizes at each school were small, and sometimes classes were offered online or in multigrade groups, as is common in rural schools (Hardwick-Franco, 2019). However, those efforts provided confidence that school leadership would sustain STEM learning pathways, thus further supporting students' STEM career aspirations.
Many of the leadership practices contributing to the high STEM performance of these rural schools were dependent on harnessing the local environment and community; so the practices are particularly rural in nature. Plainly, leadership practices endorsing STEM activities using local bushland, raising livestock or crops, or partnering with rural industry were dependent upon access to these typically rural resources. However, the community relationships utilised by leaders in this study can also be viewed as peculiar to rural communities (Preston & Barnes, 2017). While the demands of community connection can be a challenge for rural school leaders (Wieczorek & Manard, 2018), the leaders in this study harnessed these connections as a strength. Further, not only were many of the leadership practices revealed in this study rural in nature, but they were typically also STEM related. The rural industries (for example, hydroelectricity and agriculture) and community groups (for example, fire-fighting and land care) that these rural schools and their families worked with were generally STEM oriented. It is arguable that the school leaders in this study demonstrated leadership practices that are both unique to rural schools and particularly suited to supporting STEM education.
This study drew on data from three Victorian schools studied as part of a larger project examining high STEM-performing rural schools generally. Given the paucity of research into rural school leadership associated with STEM education, this is a significant contribution to the literature. However, as this study only considered practices at high STEM-performing schools, additional research is needed that compares high-and low-performing schools to confirm the contribution of the leadership practices identified in this study to STEM education success. As this study was based on only three Victorian schools future research should explore leadership practices at more rural schools across Australia and internationally. Further, the data used in this study were largely dependent on the perspectives of the leaders themselves and the teachers who work with them. Given the findings point to practices involving collaboration with the wider community, and the role of leaders in promoting STEM within their school community, future research should consider the perspectives of key community bodies, local businesses, school families, and students.

Conclusion
This study took a strengths approach to exploring STEM education in rural schools, rather than the more typical deficit approach. Viewed through the lens of the theory of Practice Architectures, this study has revealed an array of leadership practices that contribute to rural school STEM education success. The relatings of these leaders, with their staff and the wider community, improved access to enriched STEM learning opportunities, facilitated staff recruitment, and empowered staff to lead their own professional learning. The sayings of these leaders contributed to raising the profile of STEM and building STEM career aspirations. Finally, the doings of these leaders resulted in resources being acquired and allocated in ways that effectively supported STEM career education and STEM study pathways despite low enrolments and limited funds. These relatings, sayings, and doings tended to hang together as practices that are only possible in a rural school and that seem particularly effective for supporting STEM education. This study has identified practices that could be adopted by leaders of other rural schools to harness the opportunities in rural communities to improve STEM education.
Leadership practices contributing to STEM education success… 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:// creat iveco mmons. org/ licen ses/ by/4. 0/.