'I Have Seen STEM in Action and It's Quite Do-able!' The Impact of an Extended Professional Development Model on Teacher Efficacy in Primary STEM Education

Interest in Science, Technology, Engineering and Mathematics (STEM) has gained momentum due to increasing calls for a more STEM-literate society. As teaching integrated STEM poses curricular and pedagogical challenges for most generalist primary teachers, professional development (PD) is essential to support them to develop appropriate knowledge and efficacy to teach STEM. This paper presents a qualitative study of 17 primary teachers, 2 school principals and the PD facilitator. It explores the perceived impact of a customised three-phase STEM PD program on teacher efficacy in STEM education using Bandura’s (1977) sources of efficacy as an analytic lens. The findings illustrate how particular features of the PD model were identified as contributing to the development of participating teachers’ performance accomplishment, emotional arousal, vicarious experiences and verbal persuasion. In particular, there was consensus that the developmental structure of the STEM PD program, in terms of teachers’ assuming increasing ownership for their STEM learning, incrementally enhanced teacher efficacy in STEM education. Implications of this study for STEM education and in particular STEM PD are discussed fully.


Introduction
As STEM understandings, skills and dispositions are increasingly sought-after, STEM education has received enormous investment over the past two decades (Cotabish et al., 2011;Kloser et al., 2018;Morrison et al., 2015). Concurrently, there has been a steady increase in STEM policy and curriculum development (Quigley et al., 2017). As STEM is cross-disciplinary and situated in real-world problem-solving contexts (Livstrom et al., 2019), there is growing conviction that STEM disciplines 'cannot and should not be taught in isolation' (STEM Task Force, 2014, p. 9). Rather, an integrated STEM model, that situates learning in contexts where multiple disciplines are addressed together consistently and authentically, is recommended (Kloser et al., 2018).

Early Years and Primary Teachers' Preparedness to Teach STEM
Early exposure to integrated STEM education starting in pre-school or before is commonplace. Research reveals that early exposure supported by knowledgeable teachers can enhance the development of foundational STEM knowledge and sustained interest in and positive attitudes to STEM (Nadelson et al., 2013;Nesmith & Cooper, 2019;Peters-Burton et al., 2019;Quigley et al., 2017). While the role of the primary teacher is crucial in STEM implementation (Margot & Kettler, 2019) and many primary teachers are enthusiastic about integrated STEM practices (Hamilton et al., 2021;Hourigan et al., 2022;Smith et al., 2015), teaching STEM can be daunting for primary teachers who are generalists (Goodnough et al., 2014). Given the knowledge demands (Lehman et al., 2014;Margot & Kettler, 2019), many feel ill-prepared to provide integrated STEM experiences where learners work collaboratively to solve open-ended real-world problems (Cotabish et al., 2011;Nesmith & Cooper, 2019;Peters-Burton et al., 2019).
Primary teachers acknowledge that limited personal learner experiences of the STEM disciplines (e.g. science and engineering) negatively affect their efficacy beliefs in teaching these subjects (Mulholland et al., 2004;Nesmith & Cooper, 2019). In addition to the knowledge demands, teachers require robust understandings of integrated STEM pedagogies (Margot & Kettler, 2019). In Park et al.'s (2017) study, teachers identify shifting from teacher-led pedagogies to more learnerled instruction as challenging. Primary teachers also report time and scheduling challenges (Lesseig et al., 2016). Furthermore, while primary teachers demonstrate high efficacy in supporting reform-based mathematics learning (Goldenberg et al., 2015;Star, 2015), they report low efficacy in integrating mathematics within STEM tasks (Holstein & Keene, 2013). Given teachers' tendency 'to teach what they were taught', many teachers' limited experience of integrated STEM as learners is a barrier (Nadelson et al., 2013, p. 158). The challenges outlined are enduring as evidenced in Kurup et al.'s (2019) study exploring Australian pre-service primary teachers' integrated STEM beliefs, understandings, and intentions. The findings show that pre-service teachers lacked strong understandings of integrated STEM and received limited opportunities to engage with or teach integrated STEM in schools. Nonetheless, teachers' beliefs regarding the importance of integrated STEM motivate their classroom implementation (Bell, 2016). Primary teachers eager to teach integrated STEM have done so devoid of tried and tested models (Epstein & Miller, 2011). In Kurup et al.'s (2019) study, pre-service primary teachers demonstrated strong beliefs regarding its importance, reporting their intentions to teach STEM in the future.
The above research suggests that limited understandings and experiences of integrated STEM education among primary teachers resulted in low teacher efficacy (Kurup et al., 2019). To address this, innovative practice across the teacher education spectrum, including initial teacher education and teacher professional development (PD), is crucial (Barak, 2014). This study examines the effect of a needs-led integrated STEM PD program on participants' teacher efficacy in STEM education.

Teacher Efficacy in STEM Education
Self-efficacy refers to one's belief in his or her ability to produce a desired outcome and can be developed from four sources (Bandura, 1977): 1. Performance accomplishment: mastery of past experiences or previous success in the same or similar task, 2. Emotional arousal: physiological arousal felt from completing the task in the past, 3. Vicarious experiences: learned information about the task by watching others, 4. Verbal persuasion: what others tell you about the task and your ability to complete the task, especially if you view the persuader as credible.
Teacher efficacy is defined as a teacher's belief in their ability to influence student learning (Guskey & Passaro, 1994). It is both situation and content specific and influenced by factors including subject area, grade level, and learner characteristics (Tschannen-Moran et al., 1998). Teacher efficacy is a strong indicator of classroom practice (Cakiroglu et al., 2012). Previous studies explored teacher efficacy in individual STEM disciplines such as science (Murphy et al., 2007), mathematics (Gresham, 2008) and engineering (Bagiati & Evangelou, 2015;Hammack & Ivey, 2017). High teacher efficacy in a distinct STEM discipline does not imply high teacher efficacy in another STEM discipline (Hammack & Ivey, 2017) or in integrated STEM. Teacher efficacy in integrated STEM is dependent upon teachers' knowledge for teaching (Committee on Integrated STEM Education, 2014) and has been found to be related to student persistence and retention in STEM subjects (Painter & Bates, 2012) and overall improvement in student learning (Nadelson et al., 2012;Yoon et al., 2014).
Despite an absence of research examining the impact of STEM PD on primary teachers' efficacy in teaching integrated STEM, this area merits exploration given the acknowledged impact of teacher efficacy on classroom practice. According to Bandura (1977), the most influential source of teacher efficacy is performance accomplishment, or mastery experience. Engaging in integrated STEM, as both a learner and teacher, develops mastery experiences. These experiences provide evidence of a teacher's capacity to engage students with integrated STEM, directly impacting teacher efficacy (Bandura, 1977). Bandura identifies vicarious experiences as the second most important source of teacher efficacy. It involves opportunities to observe the engagement of others, including teacher educators and peers, in integrated STEM classroom practice. Teacher efficacy is also influenced by verbal persuasion, the receipt of information and formative, supportive feedback from others, about integrated STEM education practice. The emotional, physical, and psychological well-being of a teacher can influence their feelings about teaching particular content or within certain contexts. Teachers with a high sense of teacher efficacy most likely view their state of emotional arousal as energising, whereas those with self-doubt or low teacher efficacy may regard their emotional arousal as a limiting factor (Bandura, 1977). The impact of integrated STEM on learners may also affect teachers' emotional arousal.
Teachers need access to PD that enhances their STEM knowledge, skills and dispositions, thus empowering them to provide quality integrated STEM to their students (Bruce-Davis et al., 2014;Cotabish et al., 2011;National Research Council, 2013). Despite widespread support for integrated STEM education, 'comparatively little attention has been given to the content of STEM teacher preparation or professional development' (Rinke et al., 2016, p. 300). Research suggests that across all career stages, primary teachers report increases in their STEM knowledge and teacher efficacy following participation in PD (Nadelson et al., 2013). PD that provides opportunities for engagement in mastery experiences of STEM contributes to teachers' efficacy (Velasco et al., 2022). Kelley et al.'s (2020) study used Bandura's framework to evaluate high school teachers' efficacy for integrated STEM instruction after engaging in a PD program. They recommended that future research explore best practice in quality PD to enhance integrated STEM education. To date, there is a lack of research investigating the effect of STEM PD on primary teachers' efficacy in teaching integrated STEM.

Features of PD That Support Teacher Efficacy in STEM Education
This section presents a range of PD features identified as fundamental in STEM PD research. In a meta-analysis of integrated STEM PD, Lynch et al. (2019) found that PD had better outcomes (including improved student learning) when it supported teachers in using curriculum materials, sought to improve teachers' STEM knowledge for teaching and understandings of how students learn, incorporated summer workshops, and included teacher meetings to discuss and troubleshoot classroom implementation. Both Baker-Doyle and Yoon's (2011) and Kilpatrick and Fraser's (2019) studies also acknowledged that peer collaboration and sharing within STEM PD programs, through support networks and teacher learning communities, can foster improved teacher collaboration and learning. Kilpatrick and Fraser's (2019) study concluded that effective STEM PD that is needs-led, focusing on teachers' prior experiences, knowledge and beliefs, placing the teacher in the role of learner, facilitates them to experience the learners' perspective. In addition, the American primary teachers in Parker et al.'s (2015) study identified modelling by the PD facilitator, where they model the role of the teacher, as an essential component of the STEM PD. These experiences develop teachers' STEM knowledge and understandings of how students learn STEM content (Goodnough et al., 2014;Parker et al., 2015). Research concurs that engaging in integrated STEM, first as a learner followed by opportunities to implement these STEM practices in the classroom, provides a stronger foundation for enactment (Estapa & Tank, 2017;Nesmith and Cooper, 2019).
There is evidence for the benefits accruing from extended STEM PD. Brown and Bogiages' (2019) research with second level science and mathematics teachers found that multi-year STEM PD proved effective for early-career teachers. These teachers reported changes in their dispositions towards STEM after engaging in workshops across two years.
Situating PD within teachers' own school is also recommended as it allows the PD to be customised to meet the school culture and curricula (Shernoff et al., 2017). Research concurs that the peer coaching approach, the opportunity to apply PD learning in the real world of the classroom while being supported by a mentor, assists teachers in applying their newfound learning in the classroom setting (Owens et al., 2018;Parker et al., 2015;Showers & Joyce, 1996). However, sufficient time is required to support genuine collaboration between teacher participants and the PD facilitator during peer coaching. Hence, the pairing of extended PD with peer coaching has been found to create optimal changes in teachers' STEM instruction (Cotabish et al., 2011).
Each of the above features of effective STEM PD has been shown to contribute to the development of teacher efficacy in STEM education. For example, opportunities to engage with integrated STEM tasks as learners during PD workshops provides performance accomplishment and emotional arousal from the learner perspective, whereas facilitator modelling supports the development of vicarious experiences from the teacher perspective (Bandura, 1988). Equally, opportunities to engage in classroom implementation during PD promote performance accomplishment and emotional arousal, whereas peer coaching can develop teacher efficacy through verbal persuasion where the teacher receives frequent and detailed feedback (Beattie et al., 2016). The collaborative forum of meetings also provides access to meaningful verbal persuasion. This study explores study participants' perceptions of the impact of a STEM PD program that incorporates recommended PD features on teacher efficacy in STEM education.

Context of This Study
In Ireland, the location of this study, there is a dedicated primary mathematics curriculum. However, science is positioned within Social, Environmental and Scientific Education (SESE), alongside history and geography (Department of Education and Science [DES], 1999). Technology and engineering are not formal primary curricular areas. While the STEM movement is in its infancy, STEM education has progressively received increased attention in Ireland. The recent STEM Education Policy Statement 2017-2026 advocates 'a culture of collaboration for professional learning in STEM education' (DES, 2017, p. 13). Despite this, little formal guidance in the form of integrated STEM curricula or support presently exists. However, in the recent draft primary curriculum, one of the broad curricular areas is 'Mathematics, Science and Technology Education' (National Council for Curriculum and Assessment [NCCA], 2020). This proposal may stimulate national STEM PD for primary teachers. In the interim, Irish primary schools' and teachers' participation in integrated STEM practices and STEM PD is voluntary and varied.

Focus of This Study
While there is robust research on the features of effective PD and on teacher efficacy in teaching the individual STEM disciplines (Menon & Azam, 2020), there is limited research on teachers' perceptions of what successful PD 'looks like' (Owens et al., 2018, p. 370). This study reveals the impact of a STEM PD program that was developed, implemented and evaluated within the Irish primary school system on teacher efficacy in STEM education. The four sources of efficacy (Bandura, 1977) are used as a lens to evaluate the features of the STEM PD program that influenced teacher efficacy, thus addressing the following exploratory research questions: 1. How do study participants perceive changes in teacher efficacy in STEM education as a result of engaging in a STEM PD program? 2. What features of the STEM PD program do study participants identify as sources of teacher efficacy?

Methodology
The STEM PD Program The STEM PD program was designed by the PD facilitator to support implementation of integrated STEM education at primary level. It was coordinated and hosted by a regional education centre. Throughout the PD program, the use of simple and available resources, such as recyclable materials, was promoted to ensure STEM education was accessible to teachers and pupils. The program ran over three school years and was broken into three phases ( Fig. 1), incorporating key PD features, namely extended engagement, facilitator modelling, active participation, classroom implementation, peer coaching, and collaboration. These features provided teachers with ongoing opportunities to experience efficacy sources outlined by Bandura (1977) thus positively affecting their teacher efficacy in STEM education. Each PD phase required teachers to engage with workshops (in the regional education centre) and classroom implementation.

STEM PD Program: Phases 1 and 2
The regional education centre had identified teachers' lack of confidence in teaching science as an area of concern, a finding consistent with previous research in Ireland (Murphy et al., 2007) and the UK (Holroyd & Harlen, 1996). As establishing and supporting teacher efficacy in science was considered an essential pre-requisite, phase 1 intentionally supported teacher efficacy in science, with some integrated STEM ideas introduced. In phase 2, the focus of workshops moved to integrated STEM ( Fig. 1). In both phases 1 and 2 ( Fig. 1), in the face-to-face workshops, teachers engaged with science (phase 1) and integrated STEM activities (phase 2) themselves, taking the role of learners and/or teachers. The active engagement supported teachers to appreciate that integrated STEM practice was characterised by its learner-centred, active, collaborative, problem-based features. Within the workshops, performance accomplishment was developed by supporting teachers' experiences of science (phase 1) and integrated STEM (phase 2) in facilitated workshops, thus positively impacting their emotional arousal. Observation of the PD facilitator leading the science and integrated STEM instruction during PD workshops promoted vicarious experiences. From phase 1 onwards, teachers were supported (through workshop ideas and peer coaching) to implement science and integrated STEM lessons in their own classrooms between workshops, providing access to performance accomplishment and emotional arousal. Peer coaching opportunities, where the PD facilitator visited teachers' classrooms, initially co-teaching with teachers and providing feedback, offered meaningful vicarious experiences alongside verbal persuasion. Opportunities for peer sharing and feedback during subsequent workshops also facilitated verbal persuasion.  From phase 2 onwards, teacher participation increased, with teachers taking increasing ownership of their learning, considering and implementing various adaptations to STEM activities experienced in workshops, leading the instruction during peer coaching and providing feedback to peers.

STEM PD Program: Phase 3
Given the requirement for increasing teacher participation across the PD program, the final phase, referred to as 'Peer-Teach' marked a further shift where teachers moved into the role of STEM leaders. Accordingly, in phase 3, classroom implementation preceded the PD workshops ( Fig. 1), facilitating teachers to develop and trial integrated STEM education in their classrooms, which they subsequently shared with peers during workshops. While the classroom implementation fostered mastery experiences and emotional arousal, workshops promoted vicarious experiences as teachers gained insights into peers' integrated STEM ideas and approaches. This sharing forum also supported verbal persuasion as teachers received feedback from both the PD facilitator and peers.
These PD features served to enhance teachers in developing their teacher efficacy in STEM education. Figure 2 summarises how the four sources of efficacy (Bandura, 1977) were developed throughout the three phases of the STEM PD program.

Participants
In total, 60 primary teachers participated in the STEM PD program. Participation in the PD program, and in this research, was voluntary. The nature of PD engagement was dependent upon school size. In large schools (LS) (> 10 class teachers), three middle-class teachers from each school attended PD workshops during school time. In small schools (SS) (2-6 class teachers), teachers across all class levels, including teaching principals, attended PD workshops after school time. Most SS teachers were multi-grade, teaching 2-3 class groups, whereas LS teachers were Grade 3 or 4 teachers. Sampling was purposive selecting '…information-rich cases' (McMillan & Schumacher, 2001, p. 400), i.e. primary school teachers, principals and the PD facilitator. Seventeen primary teachers (including one teaching principal) whose teaching experience ranged from 6 to 35 years (mean of 13 years) participated in this research study. Two non-teaching LS principals and the PD facilitator also participated. The PD facilitator was a retired primary teacher and principal, with years of experience teaching integrated STEM, and an established science and STEM education PD designer and facilitator.

Data Collection
Data collection was mixed-methods, consisting of both qualitative surveys and semi-structured interviews. A sequential approach was used in the data collection process. The first stage involved written pre and post surveys administered to the participating teachers at the start and end of the STEM PD program. The second S138 stage, conducted at the completion of the PD program, involved interviews with teachers, principals and the PD facilitator. While these instruments were used to collect data regarding participants' beliefs about the nature and implementation of STEM education more broadly, this paper focuses on the items and data that reveal participants' perceptions of the impact of the STEM PD program on teacher efficacy in STEM education (Appendix A & B).
Given the absence of established scales to gauge teacher efficacy in integrated STEM education, the qualitative survey and interview items were intentionally open-ended to provide insight into experiences and perceived effects of the STEM PD program on participants' teacher efficacy in STEM education alongside particular PD features identified as sources of efficacy (Cohen et al., 2000).

Data Collection Stage 1: Surveys
All primary teachers who attended the first PD workshop were asked to participate in the study. Post-surveys were completed following the final PD workshop. Surveys were completed anonymously but coded to enable matching of teacher data. A total of 12 teachers completed pre-and post-surveys prior to and on conclusion of the STEM PD program. The surveys consisted of open-ended items to facilitate teachers the freedom to focus on areas and issues they deemed most relevant (Appendix A). Pre-survey items asked teachers to detail their STEM understandings, rate their teacher efficacy in STEM education (out of 10) and explain their rating as well as their motivations to participate in the STEM PD. A variety of post-survey items captured the perceived impact of the STEM PD on participants' teacher efficacy in STEM education and particular features of the PD model considered most influential.

Data Collection Stage 2: Interviews
Preliminary comparative analysis of pre and post teacher surveys informed the design of interview guides for the next data collection phase. Teachers and principals in two LSs and one SS who had participated in the PD program agreed to engage in the second stage of data collection. Semi-structured interviews were conducted on conclusion of the STEM PD program by two of the research team who were not engaged in any aspect of the PD. There were 12 interview participants (eight teachers, one teaching principal, two administrative principals, and the PD facilitator). Four of these teachers also completed the surveys. Different interview protocols were prepared for the teachers, principals and PD facilitator (Appendix B). The teacher interviews served to build on data collection stage 1, adding further depth to the data thus strengthening the findings. Interview items with principals uncovered their perceptions of the impact of the PD program on teachers, whereas the PD facilitator interview items revealed the nature and perceived effects of the PD program. The semi-structured approach, while ensuring a consistent approach across interviewee subgroups, allowed interviewers to clarify responses and probe further as appropriate and interviewees the opportunity to raise issues of importance to them. All interviews were approximately 45 min and were audio-recorded. Recordings were transcribed and identifying information redacted to provide anonymity.

Data Analysis
The researchers were aware of the limitations of self-report data and the potential mismatch between one's perceptions and reality. Furthermore, data in the form of opinions and beliefs may contain a certain degree of bias. However, this paper intentionally focuses solely on participants' 'lived experience'. Despite this, measures to ensure the trustworthiness and rigour of this qualitative study S140 included engaging with the study over a prolonged period and collecting data using different data collection instruments (survey and interview), thus promoting methodological triangulation from multiple sources (primary teachers, school principals, PD facilitator). All anonymised transcripts reflected verbatim accounts of participants' opinions and reflections (Creswell, 2009;Suter, 2012).
Analysis of the qualitative data was deductive, where Bandura's (1977) four sources of efficacy were used as a lens to guide the analysis of data with respect to particular PD features that affected teacher efficacy in STEM education. The research team's data analysis approach was informed by Braun and Clarke's (2021) guidelines: 1. Data organisation: The researchers collated the data from the instruments (surveys and interviews) and participants (teachers, school principals and the PD facilitator). The pre and post surveys (n = 12) were paired to gauge perceptions of changes in teacher efficacy in STEM education. Survey participants are identified as 'P' in the findings and distinguished using numbers 1-12, e.g. Participant 5 (P5). Although four teachers completed both surveys and interviews, these data were not linked. Pseudonyms are used for all interview participants. All data was anonymised. This document was shared with all researchers. 2. Data familiarisation: All research team members read the cumulative data set several times to become familiar with the data. In a meeting, all observations were shared and discussed. 3. Coding: Initially, two researchers worked alone to code the collated data set using the four sources of efficacy (Bandura, 1977) as a lens. They then met to share and compare their coding. This process uncovered common indicators of efficacy sources, e.g. recognising mastery of STEM as a learner as a source of performance accomplishment. The points where coding differed were of special interest, requiring further consideration and meaning making. 4. Refining and Defining indicators: The two researchers worked together to refine and define indicators for each of the four sources of teacher efficacy from the identified codes. The organised data (grouped according to the efficacy indicators and matching data) were shared with the research team. To reduce the potential that findings were influenced by personal biases, the research team met to discuss coding decisions. After discussion, the research team agreed on the coding and identification of indicators for each efficacy source. Table 1 illustrates the key indicators for respective efficacy sources and provides exemplar data. The use of multiple researchers during data analysis promoted investigator triangulation. 5. Writing Up: The researchers agreed to present the findings using the four sources of efficacy as signposts. Within the confines of this paper, a sample of illustrative quotes served to provide the reader with insights into the qualitative findings. The writing up process provided the researchers with further opportunities to clarify and recognise evidence of teacher efficacy in STEM education as a result of participation in the PD program. Physiological arousal as STEM teachers through observation of pupils' experiences There's a lot of problem solving, sometimes, you know they're … planning at the desk and they might be 20 min into it, and then you can see them all sitting back, they're deflated and they just can't solve the problem. And next thing one person is up again. So, it's patience, it's problem solving, it's organisation, its determination because a lot of children nowadays struggle to see through things (SS: Teaching principal)

Vicarious experiences
Developmental 3 Social Support network More collaboration between staff members, pupils and developed good home/school links (P2) It's the cross fertilisation of thinking and ideas and approaches (PD facilitator)

S142
In addition to the qualitative data analysis, quantitative data collected from the survey item 'Rate your efficacy in teaching STEM (out of 10)' were compared. This item, which was part of both the pre-and post-survey (Appendix A), provided further insights into the perceived impact of participation in the STEM PD on teacher efficacy. After completing initial descriptive analysis of the paired pre-and postsurvey data (Appendix C), it was observed that both data sets were not normally distributed. Consequently, the Wilcoxon signed ranks test was used to investigate the difference in teacher efficacy in STEM education before and after engaging in the PD program. Quantitative findings served to complement qualitative findings.

Findings
The findings are presented in two sections reflecting the study's research questions.

Perceived Changes in Teacher Efficacy in STEM Education
This section addresses the first research question, revealing the perceived impact of STEM PD participation on teacher efficacy in STEM education. There was consensus that prior to engaging with the STEM PD, teachers possessed low STEM teacher efficacy:

Lack of confidence in this area and knowledge of resources (P5).
Not confident about what it involves (P10). The biggest issue I found, was their lack of confidence in science, they only see the 'S' in STEM (PD facilitator).
This was also reflected in their STEM teacher efficacy self-ratings, with 10 of the 12 teachers rating themselves between 2 and 4 out of 10 (median = 4) in terms of their perceived ability to teach STEM effectively in the pre-survey (Appendix C). However, at the end of the extended STEM PD program, perceptions of increased teacher efficacy in STEM education were palpable in both survey and interviews responses: Yes, I am more confident as I have seen STEM in action and it is quite doable! (P8) Greater confidence in setting STEM challenges and allowing children to develop and expand their own ideas (P4) The ease of teaching it [STEM], it's just the most simple, no brainer… things that we should have been doing for the last 40 years (LS2: Ava) Teachers' confidence has definitely increased. After the workshops and visits from John [PD facilitator], the teachers definitely feel more comfortable (LS1: Principal) This was also echoed in quantitative findings. Firstly, all teachers' self-ratings had increased in response to the post-survey item, with 10 of the 12 teachers rating themselves between 7 and 9 on the same scale (median = 7.5). Eight of the 12 teachers' ratings increased by 4-5 points on the scale (Appendix C). In addition, the Wilcoxon signed-rank test found that increases in teachers' STEM teacher efficacy self-ratings after participation in the three-year STEM PD program were statistically significant (T = 37, z = 7, x, p, 0.01). Overall, the findings suggest that participants perceived that PD program participation positively impacted teacher efficacy in STEM education.

Features of the STEM PD Program That Supported Teacher Efficacy
This section addresses the second research question, i.e. the particular STEM PD program features that study participants identified as sources of STEM teacher efficacy.

Features of PD That Supported Performance Accomplishment
The PD facilitator acknowledged the teachers' lack of confidence in science at the start of the PD program: 'From the minute you mention science, you can see the bodies drop, you can see a reaction within the room…' Teachers also associated their poor teacher efficacy in STEM education with limited science understanding, with almost half of presurvey respondents reporting a lack of appropriate science knowledge. School principals and the PD facilitator also connected teachers' hesitancy to engage with and implement STEM with teachers' lack of personal learner experiences of integrated STEM: The biggest challenge for teachers is the lack of their own knowledge, the feeling that they don't have the knowledge (SS: Teaching principal) Teachers are afraid of their lives when they don't have the answer... (PD facilitator) Many of the teachers also reported uncertainty about engineering and technology at the beginning: 'Some of them can be bothered by the T or the E, but the M won't bother them at all' (PD facilitator).
Through active participation in integrated STEM tasks during PD workshops, the teachers learned that 'technology is anything that works for you' (SS, Deirdre) e.g., a ramp, biro. Hence, the opportunities within workshops for teachers to engage with science and STEM, initially in the role of learner and over time in the role of teacher, facilitated performance accomplishment in both science and subsequently STEM education. Teachers shared their increased confidence in their science and STEM understandings at the end of the PD program and distinctly associated this feature of the STEM PD with improvements in their teacher efficacy: …now that I understand STEM better I feel more confident and more inclined to embrace science as a subject (P11) I think the way they run it is why we're still there. …he [PD facilitator] let us off to figure it out ourselves. And I think the fact that you're making it yourself, touching the materials, and then you say 'that's grand' (LS1: Mia) S144 Participants also identified classroom implementation as a key PD feature: Between sessions, teachers had to implement. Different teachers try different ideas and then, and they always have to come back to me to the workshops with photos or models that they've done based on the content of the previous workshop. I mean I'm very insistent on that because this PD is not a show and tell and no implementation (PD facilitator) Teachers were consistent in reporting that the requirement to teach STEM in their own classrooms as part of the STEM PD program provided explicit and unique learning opportunities: 'When we brought the activities back to the classroom, what we did at the PD, we could see the value of it [STEM] even more' (LS2: Kate). Teachers acknowledged initial challenges: 'From a teacher's point of view the biggest challenge is sitting back and not giving them all the answers. It was quite hard in the early stages…We had to kind of nearly learn a different type of teaching' (LS1: Mia). Despite this, there was evidence of mastery in teaching STEM, verifying learner-centred STEM practices: I've learned to step back yourself, not be always in control, because we can have a habit of standing at the top of the classroom, and giving out everything and get them all looking good and whatever. But that's what it's about, it's to step back …. and then you're there to probe and go after them (SS: Molly) Learning is more child-centred and collaborative-I realise I don't have to be the expert and am learning in partnership with the children. Investigating and supposing are important skills, higher order skills are required. Outcomes are open-ended. Process is valued over product (P9) The teachers acknowledged the value of their own experiential learning as a precursor to classroom implementation: 'Having done these activities [in the workshops], and then myself with children in my class, I know the true educational value of the process along with the final product' (P6); endorsing the complementary nature of the teacher as learner (workshop) and classroom implementation features of the PD program in supporting teacher efficacy.

Features of PD That Supported Emotional Arousal
Participants' experiences as STEM learners themselves and observation of their pupils' experiences as STEM learners were two sources of emotional arousal, and thus contributed to their STEM teacher efficacy. There was resounding feedback from teachers indicating that they enjoyed their participation in the PD program, in particular the face-to-face workshops, which were 'practical and hands on' (PD facilitator): It was just what we needed. It was fun, our thoughts really came back to children because it was … just brilliant…it was a fantastic model really, really good (LS2: Ava) The PD facilitator recognised the importance of teachers encountering positive STEM experiences as learners: 'If they enjoy it, there's a fair opportunity that they'll transfer to the children in their classroom'. Through their own experiential learning, the teachers appreciated the need to allow productive struggle in STEM activities: I think troubleshooting is definitely one skill they develop...and to become more creative, starting to think outside the box a bit more…it's [STEM] definitely an all-rounder. It develops them in an all-round sense (SS: Molly) There's a lot of problem solving, sometimes, you know they're … planning at the desk and they might be 20 minutes into it, and then you can see them all sitting back, they're deflated and they just can't solve the problem. And next thing one person is up again. So, it's patience, it's problem solving, it's organisation, its determination because a lot of children nowadays struggle to see through things (SS: Teaching principal) Teachers' comments consistently captured their joy and emotional arousal in observing how pupils who may not generally excel academically in other lessons did so in STEM: The really artistic children you can really see their flair (LS1: Mia) Those that will generally tune off after five minutes and start fiddling or start distracting somebody would be very engrossed in this (SS: Deirdre) Teachers recognised the positive impact of their pedagogical changes on pupils' learning: Definitely problem solving, critical thinking and they learn the skill of group work, communication, and listening, and they also have to self-assess… they actually have to reflect (SS: Grace) The enhanced learning experiences witnessed by teachers affected their commitment to integrated STEM practices. They were also enthused by their pupils' motivation: When you say we're going to do STEM, there would be a big cheer in the room. That was completely unprecedented we didn't plan for it. I didn't think it would happen, but it was great. It was great (SS: Deirdre) Teachers attributed pupils' fervour to the opportunities to engage in pupil-directed actions and hands-on work. They reflected on scenarios where pupils demonstrated being as invested in the process as themselves, creating a symbiotic relationship, where the pupils' enthusiasm for and desire to engage in the STEM activities motivated teachers to engage fully and bring back new STEM activities from the respective PD sessions:  The teachers' own experience as STEM learners and their observations of pupils' enhanced learning experiences and positive attitudes to STEM culminated as a powerful efficacy source fostering enhanced teacher efficacy in STEM education. In fact, some teachers conveyed that the pedagogical approach taken to STEM education had permeated into their teaching generally: It's an all-encompassing methodology and pedagogy. Like it's all that, it has everything in it. It has everything. It has everything. It is problem solving, which is ultimately I think where we're going in education, solving problems (SS: Teaching principal) I see STEM as a methodology [teaching method] which can be used in all curricular areas (KS, Rob)

Features of PD That Supported Vicarious Experiences
Prior to participation in the STEM PD, teachers reported limited understanding of how to teach STEM: 'I don't have the confidence or knowledge in how to approach STEM' (P7). Vicarious experiences, occasions where the teachers learned about teaching STEM education by watching others, were central features of the STEM PD program. These included opportunities to observe the PD facilitator modelling best practice as teachers engaged in STEM tasks during workshops, providing them with first-hand experience of the practices they were expected to implement subsequently in their own classrooms: All my workshops are all practical...they must do it...they will break up into groups and look at and complete different STEM activities...they must test it out and do it themselves, because that's how they'll know when things are going wrong and what to expect when they're in the classroom (PD facilitator) The way the [PD facilitator] works is so tight... he has these workshops and they go back to the classroom and they try out whatever he has done in the teacher centre, and then they spread it out as best they can to the other teachers and then they're back again for more structured ideas. And you see it keeps the momentum, it keeps it sustainable (LS1: Principal) A central aspect of integrated STEM education, modelled by the PD facilitator across the PD program, was learner-centred practice. Teachers were assured they did not need to 'know the right answer', as open-ended learning values multiple approaches and solutions: We actually physically made them all ourselves. Yeah. And then what was interesting to see is how we went about it and then when you run it in your classroom, how the kids run it, and all of the end results are so different (LS1: Mia) Equally other core pedagogical elements of integrated STEM practice were modelled throughout this PD model. For example, teachers observed the PD facilitator enabling and facilitating peer learning and group discussion. During each face-to-face workshop, the PD facilitator encouraged the teachers to share their learning with each other. The PD facilitator also emphasised the importance of reflection as part of the learning process: My last question is always, 'if you were doing this again, what would you do differently, if anything?
In the post-survey, teachers their enhanced teacher efficacy (reflected in their efficacy self-ratings) was related to improved understanding of how to approach teaching STEM as a result of: I have become more confident in approaching STEM projects as I have a better understanding of the process and way of working (P8) In addition to opportunities for vicarious experiences during workshops, teachers received support from the PD facilitator during in-school visits: I do a school visit fairly early in the program...I might visit in the first year once or twice, and I might get out once more in second year ... I say I will come out for that and do it with you. So I refuse to do it on my own, they [PD teachers] have to participate with me (PD facilitator) Exposure increased due to visits from John [PD facilitator] (P11) Teachers acknowledged that the opportunity to co-teach with the PD facilitator provided unique learning opportunities that supported their teacher efficacy in pursuing a learner-centred approach: I felt I learned so much, just sitting there watching the children. It's this huge learning for the teacher to see, you know that can further, you know, influence the teaching, by just observing what's happening... (SS: Grace) Teachers recognised that they experienced a shift in teacher responsibilities across the PD program, where in phase 3, they assumed the role of STEM authorities and role models: Experience: from the facilitator and other teachers on the course (SS: Rob) …it's very step-by-step…at the start its very supportive… guiding teachers in a very structured way… But after all that they go off themselves and plan lessons and share experiences (SS: Teaching principal) The developmental 'transfer of agency' by the PD facilitator to the teachers is unpacked further in the next section.

Features of PD That Supported Verbal Persuasion
Verbal persuasion was enacted through the PD facilitator and teachers' interactions about STEM education, in particular feedback regarding their ability to teach S148 integrated STEM. In the early stages of the PD program, across the various PD structures (workshops, classroom implementation), teachers received encouragement and support predominantly from the PD facilitator. This feedback was valued given the PD facilitator's resounding credibility as a teacher, principal, teacher educator and STEM education expert: John's support and guidance was invaluable, as well as having experienced the benefits and success first hand (P10) I suppose John, if you've come across him, he's amazing. You'd listen to him all day everyday (LS1: Mia) The school principals recognised the potential for becoming over-dependent on this form of verbal persuasion, in terms of promoting sustainability: Yes, we're lucky to have John. Yes, but John isn't going to last forever (LS1, Principal) John was like what we really want, we all wanted to clone … a John for every school and we can't do that (LS2: Principal).
However, to this end, the PD model intentionally and actively sought to develop a collaborative network of teachers, where over time, levels of teacher involvement steadily increased. This lessened teachers' dependence on the facilitator as a source of knowledge: At the end, for the last 3-4 sessions [in phase 3], it is the school delivering to the schools. I don't deliver any more, they do. I support them in that and help them with ideas but they do it. I sit down as a participant and I have to build the car or whatever the same as everyone else (PD facilitator) This shift served to empower the teachers to engage in verbal persuasion. The phased increase in teacher participation was acknowledged and welcomed by teachers: He [PD facilitator] would have fed us all the information at the start exactly what he wanted to do and the lessons and then he'd pull the science out.… but maybe the second year into it…each team went off and did their own [STEM lessons] and came back and told everybody what they did. So he kind of kept pushing us then to kind of think for ourselves. And then last year, he actually got different schools to actually present to the rest…so I think it was quite a progressive course that way. He kind of wanted you to take over (LS1: Mia) Teachers explicitly commended this collaborative approach, aptly described as '…the cross fertilisation of ideas' (PD facilitator). Going forward, many teachers sought ongoing opportunities for this efficacy source, believing continued collaboration that promoted authentic verbal persuasion would sustain their increased efficacy in STEM teaching: Coming together with neighbouring schools for further discussion and support (P2) Collaboration between schools, sharing good practice (P12)

Discussion of Findings
Whereas previous literature has recognised the benefit of extended and sustained PD (Cotabish et al., 2011), this paper evidences how the recognised features of effective PD were purposefully built into the phases of an extended STEM PD program to provide the four sources of efficacy (Bandura, 1977). Despite limited prior experiences, the teachers were enthusiastic about STEM before beginning the STEM PD program. Such enthusiasm for STEM has been recognised in previous research (e.g. Hamilton et al., 2021;Hourigan, 2022;Smith et al., 2015). Teachers' self-reported low teacher efficacy in STEM education prior to engaging in the STEM PD echoes previous research which acknowledges that, as generalists, primary teachers find teaching integrated STEM daunting (Goodnough et al., 2014). The findings also support research identifying teachers' lack of experience of STEM as learners (Mulholland et al., 2004;Nadelson et al., 2013;Nesmith & Cooper, 2019) and teachers (Cotabish et al., 2011;Nesmith & Cooper, 2019;Peters-Burton et al., 2019) as contributors to low efficacy in STEM education. Similar to previous research (e.g. Murphy et al., 2007), participants recognised teachers' lack of confidence in science as a particular barrier, acknowledging the impact of low efficacy in any of the STEM disciplines in inhibiting teacher efficacy in integrating and teaching them (Bagiati & Evangelou, 2015). The four sources of efficacy were progressively developed through the three phases of the PD program. Initially the PD facilitator was positioned as the STEM education expert. However, over time, he facilitated a shift in roles where teachers increasingly assumed more ownership for their STEM learning, thus promoting sustainability through the development of a social network and community of primary STEM educators. In this way, the teachers developed efficacy differently through each phase of the PD program as the PD facilitator gradually released responsibility to the teachers. Subsequently, in both post-surveys and interviews, participants' self-report evidences positive increases in their teacher efficacy in STEM education. This research explored the various sources of efficacy, namely performance accomplishment, emotional arousal, vicarious experiences and verbal persuasion, within the STEM PD program. Participants reported many opportunities within the STEM PD program for performance accomplishment or mastery experiences in science (phase 1) and STEM (phases 2 & 3), including successful experiences as learners in workshops, scaffolded experiences as teachers in their own classrooms (phases 1 & 2) and later as teachers in the workshops (phase 3). Study participants agreed that positive experiences of science and STEM as learners, through hands-on participation in the PD workshops were sources of emotional arousal. Such experiential learning nurtured teachers' appreciation of learners' potential in STEM when supported by skilled and knowledgeable teachers (Nesmith & Cooper, 2019). The findings suggest that teachers' experiences as STEM learners enabled them to better understand how students perceive STEM, reflecting the findings of Goodnough et al. (2014), and supported them to develop and challenge their conceptions of integrated STEM (Parker et al., 2015). Participants agreed that engagement in the workshops also provided opportunities S150 for vicarious experiences, where they could learn about STEM teaching through observation of the PD facilitator modelling learner-centred pedagogies required to teach science and integrated STEM (Margot & Kettler, 2019). These findings build on previous research that recognise opportunities to assume the role of learner as an effective feature of PD (Kilpatrick & Fraser, 2019).
In terms of the classroom implementation feature of the STEM PD, teachers recognised the pedagogical challenge of shifting away from teacher-led instruction to more learner-centred approaches, reflecting the findings of previous research (Lesseig et al., 2016;Park et al., 2017). The peer-coaching feature available during classroom implementation (phases 1 & 2) was identified as a central component of the STEM PD program, providing access to meaningful vicarious experiences and enabling teachers to apply new STEM learning in their own classrooms (Goodnough et al., 2014;Owens et al., 2018;Parker et al., 2015;Showers & Joyce, 1996). The combination of the features of classroom implementation and peer-coaching supported teachers in customising their learning to best fit their own school setting and culture (Shernoff et al., 2017). In this forum, teachers received access to verbal persuasion from the PD facilitator on their science (phase 1) and STEM (phase 2) practice. Teachers valued the verbal persuasion opportunities, particularly as they perceived the persuader, the PD facilitator, to be credible due to him being renowned for his practiced pedagogical expertise (Bandura, 1977). Hence, participants perceived that the combination of peer-coaching and feedback during classroom implementation enhanced teacher efficacy in STEM education (Costabish et al., 2011).
Findings across participants suggested that the built-in classroom implementation feature across the STEM PD program, where teachers received ongoing opportunities to teach science (phase 1) and STEM (phases 2 & 3) in their own classrooms, triggered mastery experience and emotional arousal. Teachers reported that this classroom implementation facilitated them to develop an awareness of the benefits of STEM in fostering a range of skills and dispositions among pupils (Bruce-Davis et al., 2014;Smith et al., 2015). Equally, it was within classroom implementation that teachers gained insights into the benefit of providing sufficient support for learners without diminishing the cognitive demand and giving time to develop curiosity and stamina (Goldenberg et al., 2015) and the role of trial and error and struggle as a natural part of the learning process (Star, 2015). Teachers acknowledged that observation of their students' enthusiasm for and engagement in STEM instigated their recognition of the importance of integrated STEM education (Bell, 2016), and increased their motivation to teach STEM in the classroom. They recognised that the STEM PD program nurtured a cyclical relationship between students' enthusiasm and teacher motivation. Hence, in this study, teachers' emotional arousal in response to their mastery experience during classroom implementation proved an energising source of efficacy (Bandura, 1977).
The collaborative and interactive nature of the STEM PD program, where teachers received feedback from the PD facilitator and peers, was identified as a desirable feature (Goodnough et al., 2014;Parker et al., 2015). Alongside valuing feedback from the PD facilitator, teachers welcomed the vicarious experiences and verbal persuasion from other teachers following classroom implementation. They believed that this provision of frequent and detailed feedback supported teacher learning and efficacy (Beattie et al., 2016). In addition, in the final STEM PD phase (phase 3), teachers moved into the role of STEM expert, taking the lead as sources of vicarious experiences (sharing their STEM classroom practice) and verbal persuasion (providing feedback on STEM lessons and ideas) among their peers. The extended STEM PD program provided time and space for regular and meaningful interactions between the PD facilitator and teachers, developing a social network of peer support by the end of the three-year program. Participants acknowledged that the resultant social networks fostered improved collaboration and learning (Baker-Doyle & Yoon, 2011;Kilpatrick & Fraser, 2019).
Overall, the study reveals that the developmental structure of the STEM PD program that included specific features targeting particular efficacy sources facilitated teachers access to powerful performance accomplishment, emotional arousal, vicarious experiences and verbal persuasion (Lynch et al., 2019) that incrementally enhanced their teacher efficacy in STEM education (Estapa & Tank, 2017;Nesmith & Cooper, 2019). The reported impact of the STEM PD on teacher efficacy in integrated STEM education is to be welcomed given research that reveals a relationship between these teacher constructs and student learning outcomes (Lynch et al., 2019), for example students' persistence and retention in STEM subjects (Painter & Bates, 2012).

Conclusion
Given the increasing dependence on STEM-related knowledge and skills internationally (Morrison et al., 2015), STEM education has a key role to play in fostering positive dispositions towards STEM (Cotabish et al., 2011). The authors recognise primary teachers as the gatekeepers at the initial stages of the STEM education pipeline. Remarkably, this research study was carried out within an education system where there was a notable absence of a formalised primary STEM curriculum. The participants' voluntary commitment to integrated STEM implementation and PD is testament to their recognition of its value for young learners. However, while many primary teachers recognise the benefits of STEM learning, they have reported low teacher efficacy in STEM education (Holstein & Keene, 2013). The STEM PD program reported here sought to address recommendations that PD be developed to support teachers to take on the role of STEM educators (Bruce-Davis et al., 2014;National Research Council, 2013), developing and sustaining teacher efficacy (Velasco et al., 2022).
The findings of this study build on previous research examining teacher learning in STEM education following completion of extended, needs-based, collaborative PD programs with primary teachers (e.g. Nadelson et al., 2013;Parker et al., 2015) and high school teachers (e.g. Kelley et al., 2020). However, this research illustrates how the features of a specially designed STEM PD program were carefully aligned to support teacher efficacy in STEM education (Fig. 2). To our knowledge, this is one of the first studies in Ireland, and internationally, to focus on the impact of STEM PD on teacher efficacy in primary STEM education. Although Kelley et al. (2020) investigated the effects of teacher PD and S152 integrated STEM curriculum development on teacher efficacy, their study participants were high school science and engineering technology teachers.
The unique design and evaluation of the PD program using Bandura's (1977) framework as a theoretical perspective, acknowledges the importance of teacher efficacy in teachers' professional learning and practice. The research study offers a contribution to the international research by investigating how to scaffold and develop teacher efficacy through a carefully structured PD program. The authors hope that the PD program (Fig. 2), which has been shown to have a positive impact on teacher efficacy in STEM education, could be used to inform further research and STEM PD for primary teachers in other jurisdictions. The extent to which this PD model is transferable to other contexts is dependent on the compatibility of the settings and participants involved.
The authors recognise the limitations of this study. This study measures the perceived effectiveness of the program on teachers' efficacy in STEM education; it does not reveal the impact on the quality of their integrated STEM teaching or the effect of this practice on their students' understandings, skills and dispositions. In addition, the research focused on one group of teachers' experiences of the STEM PD program. Due to the small sample size, the results are not generalisable. However, the self-report data provides a qualitative and valued insight into the 'lived' experiences of the key stakeholders; teachers, school principals and the PD facilitator.
This research is timely as the enactment of primary STEM education is being supported in different ways internationally, where some countries and educational systems are further along their journey. In Ireland, conversations are starting regarding optimal approaches to support primary teachers in implementing integrated STEM education. The findings of this research provide valuable insights into the effectiveness of STEM PD as the Irish STEM policy is in the early phases of implementation (DES, 2017) and the primary school curriculum is being revised (NCCA, 2020). The current establishment and development of STEM in primary education provides an opportune time to consider the implications for the development of relevant STEM PD models. This study has implications for STEM educators, policy makers and PD developers internationally, as it signifies the need to support teachers by providing PD opportunities that encapsulate teacher efficacy in STEM education. Future study could capture the medium to long term effect of STEM PD participation on teacher efficacy in STEM education. It would also be worthwhile to explore the impact of a teacher network on experienced and novice STEM teachers' efficacy to teach integrated STEM education (Hourigan et al., 2022). Follow-up research should also measure the impact of participation in STEM PD on teachers' integrated STEM practice and their students' understandings, skills and dispositions using measures including classroom observation and student survey and/or interview.