A scoping review of interventions in primary science education

ABSTRACT

Effective science education is crucial for developing a scientifically literate citizenry, and for many, foundational primary science education experiences play a significant role in defining their long-term science trajectories. However, primary science education has been limited by a dissonance between the poor science trajectories for generations of primary students and the positive findings often reported university research; a divide that teachers are primarily responsible for bridging. This paper presents a scoping review of primary science intervention literature from the past 20 years to both describe the research outputs and analyse the evidence for the effectiveness of different primary science teaching approaches. The search yielded 142 research outputs from 26 nations with data from as many as 36,021 students. The results showed an established field with robust research designs covering all science disciplines and primary school years. Effect size analyses showed that an array of student-centred interventions covaried with large to very large improvements in science content knowledge, skills and attitudes. With the effectiveness of many student-centred approaches established in primary science education, issues of feasibility and scalability should now become a central focus for all stakeholders. Limited coverage of K-2 science in the sample was a point of concern.

KEYWORDS:

• Primary
• elementary
• science
• intervention
• evaluation
• review
• practices
• pedagogy

Introduction

In our increasingly complex, interconnected societies, the importance of a scientifically literate citizenry for meaningful economic and social engagement cannot be overstated. In the broadest sense, a scientifically literate citizen is able to apply their robust science knowledge and skills in novel contexts (Scientific Literacy) with nuanced consideration for societal impacts (Science Literacy; Bybee, 1997; Bybee, 2014; Laugksch, 2000; Roberts & Bybee, 2014). However, this view of the purpose of science education is not universally accepted and remains subject to debate (Hodson, 2003; Jimenez & Menendez Alvarez-Hevia, 2021; Linder et al., 2011). A recent report released by the U.S. National Academies of Sciences, Engineering and Mathematics (2022) reinforced the importance of effective science education in capturing and developing students’ natural curiosity. Primary teachers are particularly well-positioned to teach science in integrated ways that meaningfully consider learners’ social, cultural, and developmental needs. For example, Gresnigt et al. (2014) review of eight integrated primary science projects showed that primary educators are capable of developing their learners’ science dispositions, skills and knowledge whilst simultaneously increasing the amount of time spent on science in their classrooms. Early engagement in science is crucial to establishing positive science trajectories as there is a corpus of literature indicating learners’ science disengagement tends both to worsen after primary school and become more challenging to remediate (Ali et al., 2013; Denessen et al., 2015; DeWitt & Archer, 2015; DeWitt et al., 2014; Lindahl, 2007; Said et al., 2016).

The purpose of the scoping review presented in this paper is to describe and consolidate a selection of research outputs covering interventions in primary science classrooms in terms of demographics, research designs, and most importantly, student learning outcomes. The following subsections will orient the reader to the relevant background literature in the field of primary science education and make an argument for the scoping review presented. Initially, the state of primary science education will be outlined, followed by an argument for the importance of primary science learning trajectories and then a discussion of current classroom practices. The latter sections of this introduction will explore systemic challenges to primary science education, describe recent improvements in this space and interrogate nebulous notions of ‘best practice’ in primary science teaching. To conclude, a more thorough purpose statement will precede the research questions.

The state and purpose of primary science education

Globally, science education still has considerable room for improvement. According to Trends in International Mathematics and Science Study (TIMSS) data over the past 25 years, children’s foundational primary science education experiences have not yet enabled them to fulfill their potential (M. O. Martin et al., 1997; Thomson et al., 2020a). While the declining science achievement scores of Year 4 students from multiple Organisation for Economic Cooperation and Development (OECD) nations, such as Australia, England, Japan, New Zealand and the United States, are of concern, the most pressing issue is that 90% of participating nations are failing to meet the High International Benchmark (550): a sign of a student’s capacity to generalise their science learning beyond the classroom. Although compelling, such TIMSS data are mired by issues that limit all large-scale international assessments, such as misalignment of syllabi across jurisdictions, exclusionary sampling, methodological rigidity, and cultural bias which can all diminish the relevance of findings to students’ and teachers’ lived experiences (Baker, 1997; Bracey, 2000; Schuelka, 2013; Wang, 2001; Zhao, 2020). Such international tests also lack sensitivity to the more localised contextual differences that are central to regular classroom teachers’ experiences. For example, Year 4 Spanish students’ science achievement on a national assessment varied considerably by region, with multiple regression analyses indicating that co-official languages and regional socio-economic status explained 80% of the variation in achievement levels (González-Gómez et al., 2019). Understandably, recent research from Luxembourg has begun to investigate linguistic diversity as it relates to science education experiences in ways that large scale assessments cannot (Salloum et al., 2020; Siry, 2020; Wilmes & Siry, 2021). So, while scientific literacy is a justifiable focus, it cannot be deemed universal in its acceptance nor its measure, as the embodiment of scientific literacy will vary across nations, regions, schools and classrooms.

The importance of primary science learning trajectories

Primary science education experiences are crucial as students’ science trajectories become more intractable during the primary-secondary transition (DeWitt et al., 2014; Lindahl, 2007), with many students becoming more disaffected by their science learning as they progress through secondary school (Ali et al., 2013; Said et al., 2016); in particular, girls express far less favourable science views than boys over time (Denessen et al., 2015; DeWitt & Archer, 2015). A recent investigation of 149 Spanish primary students showed that Year 3 students held more informed views on the nature of science than older students in Years 5 and 6 (Toma et al., 2019). Similar positive science dispositions were reported in the questionnaire data obtained from 183 Spanish upper primary students (Fernández Cézar & Solano Pinto, 2017). Promisingly, the most recent iteration of TIMSS showed significant improvement in Australian Year 4 Students’ science achievement and a closing of the gap between metropolitan and non-metropolitan learners (Thomson et al., 2020a, 2020b). This aligns with reports of evidence-based practice in school and university settings (e.g., Aubusson et al., 2019, 2015; Deehan, 2022, 2021; Fitzgerald et al., 2019). Such signs of improvement have buttressed calls for more research into earlier science education (Potvin & Hasni, 2014). Clearly, the state of primary science education warrants substantial focus from educators, researchers, administrators and communities to ensure we rise to meet the challenges of 21^st century societies.

Current primary science teaching practices

Decades of consistent focus on primary science education may be having positive impacts on primary science teaching practices and classroom experiences (e.g., Aubusson et al., 2015, 2015). According to the recent U.S. National Survey of Science and Mathematics Education (NSSME), elementary teachers are emphasising science concepts through explanation, class discussions, group work and hands on activities (Banilower, 2019; Banilower et al., 2018). However, opportunities for students to develop scientific skills by engaging in scientific tasks themselves remain somewhat limited. Conversely, data from the Australian National Sample Assessment – Science Literacy (NAP-SL; Australian Curriculum, Assessment and Reporting Authority (ACARA), 2019; Australian Curriculum, Assessment and Reporting Authority (ACARA), 2013) showed high percentages of Year 6 students who ‘mostly’ or ‘always’ planned and carried out their own investigations (52%), worked on group investigations (74%) and had in-depth class discussions (61%) as part of their science lessons. Yet, the outward facing approaches of guest speakers (35%) and excursions (36%) remain underutilised. In a contradiction of earlier trends (e.g, Goodrum et al., 2001), there appears to be evidence that primary teachers are willing to embrace student-centred practices in their science teaching practice; possibly suggesting the issues in science education are related more to systemic challenges, such as time (e.g., Crump, 2005; Jenkinson & Benson, 2010), resourcing (e.g., Gonski, 2011; Rowe & Perry, 2020), socio-economic challenges (Halsey, 2018; Sullivan et al., 2018) and crowded curricula (e.g., Akar, 2018; Australian Primary Principals Association (APPA), 2014), rather than inherent teacher deficiencies. In their survey of 709 Spanish Year 6 students, Jimenez and Menendez Alvarez-Hevia (2021) found that these students were becoming more motivated by science learning that relates to their everyday lives. At the very least, teachers appear to be doing their best to produce engaging science lessons as an overwhelming majority of Australian Year 6 students on the NAP-SL want to learn more science (86%) and believe their teacher enjoys teaching science (85%), but the possible scientific literacy issue emerges with just under half (49%) viewing science as part of their everyday lives (Australian Curriculum, Assessment and Reporting Authority (ACARA), 2019). This distils the strength of primary educators in sparking student curiosity (e.g., NASEM, 2022) and emphasises the need for continued development of students’ science and scientific literacy (Bybee, 1997; Roberts & Bybee, 2014). Efforts to implement effective, feasible and scalable primary science education practices are greatly hindered by a degree intergenerational science disengagement (e.g., Breakwell & Beardsell, 1992; Howitt, 2007) and systemic issues such as overburdened teaching schedules, too many competing curricular requirements, resourcing limitations and context specific challenges (AITSL, 2021; Akar, 2018; Australian Primary Principals Association (APPA), 2014; Crump, 2005; Gonski, 2011; Halsey, 2018; Jenkinson & Benson, 2010; Rowe & Perry, 2020; Sullivan et al., 2018). Deeper interrogation of the aforementioned negative science trajectories (Ali et al., 2013; Denessen et al., 2015; DeWitt & Archer, 2015; DeWitt et al., 2014; Lindahl, 2007; Said et al., 2016) suggests that they may be catalysed by passive, teacher-centred centred primary science teaching approaches (Goodrum et al., 2001). A national review of Australian primary science education showed that over a quarter of surveyed students were frustrated by teacher centred approaches, such as note taking (Goodrum et al., 2001). This has been echoed in subsequent reviews wherein primary students felt disengaged with their science learning and struggled to make connections to their daily lives (Australian Curriculum, Assessment and Reporting Authority (ACARA), 2013; Goodrum & Rennie, 2007). Indeed, due both individual and systemic issues, passive learning can still feature too prominently in primary science education practice (Australian Curriculum, Assessment and Reporting Authority (ACARA), 2013; Australian Curriculum, Assessment and Reporting Authority (ACARA), 2019; Banilower et al., 2018; Banilower, 2019; Goodrum & Rennie, Goodrum et al., 2001).

Systemic challenges to primary science education

Long-term declines in post-compulsory secondary science enrolment (Kennedy et al., 2014; Norton & Cakitaki, 2016; Norton et al., 2018) may be caused, at least in part, by the aforementioned challenges to primary science education, such as limited resourcing (e.g., Gonski, 2011; Rowe & Perry, 2020), socio-economic divides (Halsey, 2018; Sullivan et al., 2018), overcrowded curricula (e.g., Akar, 2018; Australian Primary Principals Association (APPA), 2014), and insufficient time to attend to core teaching tasks (e.g., AITSL, 2021; Crump, 2005; Jenkinson & Benson, 2010). Such systemic challenges (e.g., Breakwell & Beardsell, 1992; Howitt, 2007) are likely contributors to issues surrounding primary teachers’ interest in and knowledge about science (Appleton, 1992, 2003; Bleicher & Lindgren, 2005; McDonnough & Matkins, 2010). Preservice primary teachers hold similar views to their inservice counterparts (Boon, 2010; Murphy & Smith, 2012). However, there is some recent evidence to suggest intergenerational science disengagement is being addressed as Pino-Pasternak and Volet (2020) found that half a cohort of Australian preservice primary teachers entered their studies with favourable science views. Despite educators seeing the value in science, there remains room for growth. For example, many preservice primary teachers lack sophisticated and technically accurate understandings of widely covered socio-scientific issues in global warming and climate change (Boon, 2010). Continued improvement of the attitudes and capacities of preservice teachers remains the central focus for many preservice primary science academics as they strive to enhance the quality of primary science education (Deehan, 2022). A pre-post-test investigation of 1070 primary students across 54 classes in Germany found that measures of teacher competence, particularly science teaching efficacy, were the strongest predictors of student science achievement (Fauth et al., 2019), which clearly illustrates the importance of teachers in improving primary science education.

Recent improvements in primary science education

The state of preservice primary science education embodies many positive developments in primary education. Over the past two decades, there has been a concerted realignment of preservice primary initial teacher education (ITE) towards authentic, school-based practices and deeper student-centred learning (Deehan, 2022; Palmer, 2008); with the field beginning to address divides between traditional and on-campus learners (Deehan, 2021). Extensive research has been published highlighting the positive impacts of various student-centred tertiary education approaches on the science knowledge (McKinnon et al., 2017), teaching efficacies (Deehan et al., 2017) and dispositions of preservice primary teachers (González-Gómez et al., 2019). A wide variety of approaches have shown to be beneficial to preservice primary science teachers for their own benefit and as models for practice in their professional teaching both in the classroom and beyond, including: alternative conceptions-based instruction (e.g., Deehan et al., 2017, 2019; McKinnon et al., 2017; Trundle et al., 2007), authentic teaching tasks (e.g., Kim & Bolger, 2017; Lewis, 2019; Wallace & Coffey, 2019), constructivism (e.g., Hume, 2012), group learning (e.g., Deehan et al., 2017, 2019), cross curricular integration (e.g., DeLuca et al., 2015; Parker et al., 2012), inquiry learning (e.g., Chen & Tytler, 2017; Saçkes et al., 2012), firsthand science teaching experiences (e.g., Kahn & VanWynsberghe, 2020; Lewis, 2019; Palmer, 2011), mentoring (e.g., Kenny, 2012; Sempowicz & Hudson, 2011), modelling (e.g., Donna & Hick, 2017; Menon & Sadler, 2018), nature of science instruction (e.g., Demirdöğen et al., 2016; Mesci & Renee’S, 2017); problem-based learning (e.g., Etherington, 2011; Ford et al., 2013), reflective practices (e.g., Aydeniz & Brown, 2017; Dalvi & Wendell, 2017), school-university teaching partnerships (e.g., Hobbs et al., 2018; Kenny et al., 2014) and student-centred investigation (e.g., McKinnon et al., 2017; T. Wu & Albion, 2019). However, despite concerted efforts in the university sector to both expand and incorporate the evidence base regarding student-centred practices in ITE programmes there is still considerable potential for growth, with some educators expressing resistance to pedagogical changes (Toma & Greca, 2018) in favour of teacher-centred approaches (Banilower, 2019; Banilower et al., 2018; Goodrum et al., 2001) and/or overly simplified notions hands-on learning (Kleickmann et al., 2016). Although efforts have been made to bridge the long-standing divide between schools and universities (Anagnostopoulos et al., 2007), via the establishment school-university partnerships (e.g., Hobbs et al., 2018) and research into preservice to inservice teaching transitions (e.g., Deehan et al., 2020), more work is needed to ensure these efforts are consolidated into improved in-school primary science teaching practice.

‘Best practice’ in primary science education

The sheer array of and complex interactions amongst many of the ‘best practices’ present a pressing professional challenge for teachers in ensuring their specific approaches are most effective for their unique student cohorts. It is hoped that the scoping review presented in this paper can empower educators in the selection, adaptation and justification of their science teaching practices. At the macro level, broad principles of best practice are embodied in sources such as Harlen’s (2015) ‘14 big ideas about and of science’ and Roth’s (2014) ‘attributes of effective primary science teaching’, both of which are based clearly in science education research literature. Aubusson et al. (2015) have drawn on a corpus of literature to advocate for inquiry learning, digital technologies, authentic experiences and other student-centred approaches in primary science education. Indeed, school-based research has shown the effectiveness of student-centred approaches, including, but not limited to: Problem-Based Learning (e.g., Bulu & Pedersen, 2010; Wu & Hsieh, 2006), Inquiry Learning (e.g., Assaraf & Orion, 2010; Liu, Horton, Olmanson, Toprac et al., 2011), Site/Place Based Learning (e.g., B. Akcay & Akcay, 2015; White, Eberstein, Scott, Ito et al., 2018), Cooperative Learning (e.g., Can et al., 2017; Tarhan, Ayyıldız, Ogunc, Sesen et al., 2013), Nature of Science Instruction (e.g., Khishfe & Abd-El-Khalick, 2002; Quigley, Pongsanon, Akerson et al., 2010), Technological innovations (e.g., Field, 2009; Rowe, Shores, Mott, Lester et al., 2010) and Cross Curricular Integration (e.g., Abdi et al., 2013; Gresnigt et al., 2014; Guthrie, Wigfield, Barbosa, Perencevich, Taboada, Davis, Tonks, Tonks et al., 2004). Furthermore, countries such as Australia (Australian Curriculum, Assessment and Reporting Authority (ACARA), 2017), the United States of America (Next Generation Science Standards (NGSS), 2013), Great Britain (Eggleston, 2018), and Singapore (M. Kim et al., 2013) have pivoted to explicitly frame science curricula around student-centred, inquiry learning approaches. Still, near consensus realignment at the macro levels does not fully alleviate the challenges at the micro levels of primary science practices and learner outcomes within classrooms. In fact, Deehan (2022) has directly called for more efforts to contextualise notions of best practice in science education for, often disempowered, generalist primary teachers who are seldom represented in the research literature.

The purpose of this scoping review

Ongoing consolidation and critique of the primary science education literature is essential if recent improvements are to are to be sustained and developed. However, the sheer and ever-increasing volume of publications makes it markedly more challenging to remain abreast of broader trends within specific fields of research, particularly for stakeholders outside of academia. Scoping reviews can be an efficient catalyst for data-informed practice within schools. Without regular consolidation via literature reviews (e.g., Aubusson et al., 2015), research trajectories in science education and education more broadly can become unfocussed and repetitive (Dyment & Downing, 2020; Potvin & Hasni, 2014). This paper aims to contribute to an underrepresented domain in primary science education (Skamp, 2020) by presenting a scoping review of classroom interventions that is more targeted than broader STEM reviews (e.g., McDonald, 2016; Murphy, MacDonald, Danaia et al., 2019; Murphy, MacDonald, Wang et al., 2019), yet retains a degree of flexibility that cannot be replicated in more restricted meta analyses (e.g., Deehan et al., 2017). In this paper, we analyse a sample from the body of primary science education research literature to answer the following questions:

• What is the state of primary science intervention research in terms of demographics, pedagogical approaches and research designs? and

• How effective are the different approaches employed in the primary science intervention research in terms of enhancing students’ science knowledge, skills and attitudes within different contexts?

Method

The purpose of this review is to identify and analyse the research evidence for science education interventions in primary school settings. As the aim is to provide an overview of the available evidence, a scoping review (Peters et al., 2020) was chosen as the most appropriate methodology. A scoping review is characterised by being reasonably open in terms of structure, meaning that it cannot provide critical appraisals and synthesised answers; rather it is best suited to providing an overview or map of evidence (Munn et al., 2018). Essentially, this review has remained somewhat open to the field by focusing broadly on the state of a field and ‘effectiveness’. While this prevents the scoping review from being exhaustive, it can make it more inclusive of different perspectives within an academic field. It is important to note that literature reviews exist on a continuum rather than having fixed categories. The authors opted against a narrative review because the absence of a structured literature search methodology would hinder replicability and cause issues of bias. A more structured systematic review was also avoided as this would have required a fixed definition of effectiveness, such as national test scores, that would not have reflected contemporary educational perspectives. Ultimately, the scoping review serves to identify and report the evidence in relation to an area of inquiry, rather than answer a narrow question (Peters et al., 2020). For example, Gresnigt et al. (2014) tighter focus on integrated curriculum in primary science education enabled them to interrogate a smaller selection of research outputs with greater depth. Alternatively, more open reviews can provide a broader, more inclusive overview of primary science practice, but are open to critique in terms of transparency and, by extension, replicability (Aubusson et al., 2015). Indeed, unlike more open ended literature reviews, the scoping review benefits from clear inclusion criteria, a systematic search process, a clear coding framework and transparent reporting for replication and critique (Munn et al., 2018); yet it retains a degree of flexibility to respond to unanticipated literature themes, research designs and data collection methods for the sake of an open approach to analysis that may not have otherwise been afforded by a more narrow systematic review. In essence, the scoping review presented in this manuscript aims to find a middle ground between more structured, restricted reviews, such as meta-analyses (e.g., Deehan et al., 2017) and less structured, more open reviews (e.g., Aubusson et al., 2015). The inclusion criteria and transparent search processes enable the reader to see how the sample meets the intended research aims, but the authors cannot claim the search to be exhaustive (Arksey & O’Malley, 2005). This necessary compromise allowed the area of inquiry to be defined prior to investigation (Munn et al., 2018), for example, the focus of the study must be an intervention for primary science education. Conversely, a more general approach may have included studies that had a mathematics intervention on attitudes towards engineering and science. Several searches were conducted using slightly modified terms, yielding fewer non-duplicate results as the search phase continued, signalling an increasingly representative sample as more outputs were collected.

For the purpose of this review, we sought to identify papers that reported empirical evidence regarding interventions in primary science education. The search was limited to peer-reviewed papers published in the last 20 years; written in English; and retrievable through journal databases. A summary of the inclusion and exclusion criteria is presented in Table 1. Selected research outputs needed to feature a science intervention delivered to primary aged students. To account for possible jurisdictional differences, primary students were broadly defined as being 12 years of age or younger, and/or in Year 6 or lower. It should also be noted that the focus on evaluative research with delineated time periods is likely to skew the search process towards quantitative and mixed methods designs. While this is appropriate for the aims of this scoping review, the authors do not wish to understate the valuable, ecologically valid contributions qualitative research makes to primary science education research.

Table 1. Inclusion and exclusion criteria.

Inclusion criteria	Exclusion criteria
Empirical research	Not empirical research Unclear empirical research Second hand or indirect research
Published from 2001 onwards	Published prior to 2001
Published in English	Published in a language other than English without an English executive summary
Focus of the study is an intervention for primary science education	Focus of the study is something other than a primary science intervention Focus of the study is a STEM intervention with ill-defined links to science
The study seeks to evaluate an intervention by measuring or exploring student science performance or achievement through data collected on at least two separate, defined occasions (i.e., pre-to-post-test)	The study does not seek to evaluate an intervention through data collected on at least two separate occasions. The study focuses on classroom/ science teaching processes/ experiences in a way that may not clearly delineate time periods (i.e., pre-to-post-test).
Intervention explicitly includes primary school-age students (i.e. 5–12 years of age and/or Kindergarten to Year 6)	Intervention is for students older or younger than the specified age and Year level ranges.

Search strategy

Between July 1st and 7 August 2021, sweeping searches were conducted through Google Scholar and Primo for publications of studies focusing on primary science education interventions. These two scholarly search engines enabled access to 257 academic journal databases, including EBSCOhost, ERIC, ProQuest Education Database, Taylor and Francis Online, Informit, JSTOR, SAGE journals online, Scopus, SpringerLink, Ulrichsweb, Web of Science and Wiley Online Library. This method of searching was selected to retrieve the greatest variety of publications on the field, sacrificing the necessary time to process the larger quantity of publications for a more representative initial sample (Arksey & O’Malley, 2005). Six Boolean search strings were developed by the authors and a research assistant with consultation from an academic librarian. The Boolean search strings are presented in Table 2. Date limiters were applied to all searches, excluding all publications prior to 2001. Results were further limited to peer-reviewed articles written in English or featuring an executive summary in English. In total, 119,500 results were found across the six search strings. To restrict the inordinate time that it would take to screen all results, a Boolean search string would be terminated if 20 consecutive results did not meet the inclusion criteria, effectively creating an initial yield of 1,820 research outputs.

Table 2. Boolean search string terms.

Boolean search string 1:	science AND (elementary OR primary OR ‘middle school’ OR ‘middle-school’) AND (educat* OR teach*) AND (project OR programme OR unit)
Boolean search string 2:	(science -‘social science’) AND (elementary OR primary OR ‘middle school’ OR ‘middle-school’) AND (education OR teaching OR instruction) AND (project OR programme~ OR unit OR lesson) AND (child~ OR student~)
Boolean search string 3:	(elementary OR primary OR ‘middle school’ OR ‘middle-school’) AND (science -”social science”) AND (student~) AND ‘post-test’
Boolean search string 4:	(science -‘social science’) AND (elementary OR primary OR ‘middle school’ OR ‘middle-school’) AND (education OR teaching OR instruction) AND (project OR programme~ OR unit OR lesson) AND (child~ OR student~) AND perform~
Boolean search string 5:	science AND (elementary OR primary OR ‘middle school’ OR ‘middle-school’) AND (educat~ OR teach~) AND (project OR programme OR unit) AND -”pre-service”
Boolean search string 6:	(science -”social science”) AND (elementary OR primary OR ‘middle school’ OR ‘middle-school’) AND (performance~ OR achievement~) AND (student~ -‘pre-service’) AND (intervention OR treatment OR experimental)

Screening process

Papers retrieved through the search strategy were screened to identify studies which satisfied the inclusion criteria. Figure 1 presents the four-phase screening process of identification, Level 1 screening, and Level 2 screening and inclusion. The identification phase screening consisted of a check of the title, keywords, and abstract by the third author and a research assistant. The initial screening resulted in 258 papers that appeared to meet the inclusion criteria. Following this initial screening, the full text of these 258 papers were double checked by the third author and research assistant, resulting in the removal of 25 research outputs that were duplicates or otherwise did not meet the inclusion criteria (Level 1 Screening). At this point, the remaining 233 publications were systematically coded and summarised by the first and third authors (Level 2 screening). The research outputs were cross-examined systematically, resulting in refinement to the coding framework (Table 3) and categorisation of approaches to primary science teaching (Table 4) to more accurately group or distinguish distinct qualities of each study. During this phase, an additional 91 articles were discarded for a variety of reasons, including but not limited to: unclear participant ages, lack of firsthand student data and broad STEM foci that could not be clearly classified as science education interventions. This multi-phase screening process, conducted by the authors and a research assistant, resulted in 142 papers meeting the criteria for inclusion in this scoping review.

Figure 1. Literature screening process.

Table 3. Coding framework.

Area of interest	Codes
Research methods	Qualitative
	Quantitative
	Mixed methods
Research design	Cross-sectional (one shot post-test)
	Quasi-experimental, pre-post-test design or multiple data collection points (including equivalent groups, single cohort & multiple cohorts)
	Quasi-experimental, pre-post design (with non-randomly assigned control group)
	Experimental design (with random assignment)
Longitudinal	Data collected after a delay period
	No data collected after a delay period
Outcome measure(s)/ concept(s)	Achievement/content knowledge
	Skills
	Attitudes
	Other
Science focus	LW – Living World
	PW – Physical World
	ES – Earth & Space
	MW – Material World
	WS – Working Scientifically
	DP – Design & Production
	NoS – Nature of Science

Table 4. Approaches to primary science teaching.

Approach	Description
Alternative Conception Targeting	Learners’ alternative conceptions can inform the design and delivery of science learning experiences (e.g., McKinnon et al., 2017) and are typically identified through diagnostic assessment at the commencement of a learning and teaching cycle. Alternative conceptions can be sourced directly from learners or through scholarly material.
Concept Mapping^a	A distinct subset of research output focussing on concept mapping as both an assessment tool and a means of structuring metacognitive learning reflection. Modes of concept mapping (Brandstädter et al., 2012), concept mapping as a tool for learning reflection (Asan, 2007) and integration of concept maps into complex digital learning spaces (Hwang, Kuo, Chen, Ho et al., 2014).
Play-Based Learning^a	Play-based learning is based upon harnessing children’s tendencies for imaginative and physical creative play in learning and recreation. Whether student or teacher guided, play-based learning is creative, spontaneous, enjoyable and process-focussed (Ashiabi, 2007; Pyle & Danniels, 2017; Sturgess, 2003). In science education research, play-based learning can occur in digital (Enyedy, Danish, Delacruz, Kumar et al., 2012) or physical (Bulunuz, 2013) spaces.
Problem/Project-Based Learning	Problem-based learning uses real-world problems as a starting point for the acquisition and integration of new knowledge into existing schemas (e.g., Etherington, 2011). This approach helps students to develop transferable skills which can be used in novel situations.
Constructivism/5Es	Learning that occurs when an individual constructs their knowledge through active cognitive (Kamii & Ewing, 1996) and/or social participation (Vygotsky, 1977) within a phenomenon or situation (Slavin, 1996). This also includes the 5Es framework underpinning the Primary Connections resources that are frequently used in Australian primary schools (e.g., Hume, 2012). Although previously deemed a guiding principle of primary science education (Deehan, 2022), constructivism has been operationalised for teaching through a variety of models and frameworks (Aubusson et al., 2015).
Cooperative Learning	Cooperative learning occurs when students work together to complete a task that would otherwise be impossible or unreasonable to complete individually (e.g., Deehan et al., 2017). This strategy is often supported with clear expectations in terms of process and output. For example, students or groups of students may be assigned discreet roles within a larger learning task (Tarhan, Ayyıldız, Ogunc, Sesen et al., 2013).
Cross-Curricular Integration (Including STEM)	An approach to teaching where two discreet disciplines are integrated to create deep learning outcomes. For example, allowing students to collect and graph data is an example of a deep integrative link between mathematics and science (e.g., Kim & Bolger, 2017). Broader STEM (e.g., Kandlhofer & Steinbauer, 2016) or deeper numeracy (e.g., Weinberg, Basile, Albright et al., 2011) foci would be classified as cross-curricular integration.
‘Hands-On/Student-Centred (NFS)’^a	Student-centred learning is characterised by affording agency to learners through hands on learning or simulations. Teachers function as responsive facilitators (e.g., McKinnon et al., 2017; Palmer, 2006). Student-centred learning can be more contained experiences where knowledge and skills can be developed for more extended inquiry sequences. This approach has previously been classified as a guiding principle of, rather than a tangible approach to primary science education (Deehan, 2022), and this can be reflected in procedurally unclear intervention descriptions (e.g., Akyurek & Afacan, 2013). In this study, this concept relates to researchers’ broad conceptualisation of interventions as “Hands-on/Student-Centred” whilst not providing further specificity (NFS). It is provided in quotation marks to signal that it is a distinct code that does not cover all areas. While it is acknowledged that hands on approaches are not necessarily student centred, this categorisation simply reflects how some interventions have been described.
Inquiry Learning	Inquiry learning is characterised by a focus on a specific outcome. It allows participants to apply skills and knowledge to seek the information needed to achieve the outcome. Learners can be afforded partial (guided) or complete (open) control of the inquiry process (e.g., Fitzgerald et al., 2019).
Lesson Sequencing^a	A small selection of research outputs focused primarily on how lessons and/or learning experiences were delivered, organised or sequenced. Research foci included: learning stations (Metcalf, 2013; Ocak, 2010), massed, clumped and spaced science lesson delivery (Gluckman, Vlach, Sandhofer et al., 2014) and self-paced learning (Martin, Durksen, Williamson, Kiss, Ginns et al., 2016).
Literacy Support^a	Literacy support approaches were categorised as discreet from other forms of cross curricular because such elements are foundational, rather than enriching, to primary science teaching. Common Literacy Support approaches include: varied science texts (e.g., Balim et al., 2016; McTigue, 2009), Concept Oriented Reading Instruction (CORI; e.g., Guthrie, Wigfield, Barbosa, Perencevich, Taboada, Davis, Tonks, Tonks et al., 2004; Wigfield, Guthrie, Perencevich, Taboada, Klauda, McRae, Barbosa et al., 2008), science language transitions (e.g., Brown & Ryoo, 2008; Brown et al., 2010) and functional literacy scaffolds for problem solving activities (e.g., Bulu & Pedersen, 2010). Dialogue in primary science teaching continues to be a stronger area of focus in the field (Diakidoy & Kendeou, 2001; France, 2021).
Narrative Focus^a	Interventions with a narrative focus organise the resources, digital technologies and/or learning experiences around a cohesive story with the aim of developing holistic, synergistic knowledges and skills within learners (e.g., Lester, Spires, Nietfeld, Minogue, Mott, Lobene et al., 2014; Ritchie, Tomas, Tones et al., 2011).
Nature of Science Instruction	The understanding that scientific knowledge is fluid and always subject to reasonable debate. Instruction in this area may orient the learner to the variety of scientific approaches beyond an experimental research design. Essentially, ‘Nature of Science’ Instruction orients learners to science epistemology (e.g., Demirdöğen et al., 2016).
Science Beyond the Classroom	Interventions where the science learning is centred on experiences and strategies that extend beyond a typical primary science classroom. Excursions to local ecosystems (Prokop, Tuncer, Kvasničák et al., 2007), outdoor learning practices (Assaraf & Orion, 2010), museum visits (Martin, Durksen, Williamson, Kiss, Ginns et al., 2016), engagement with community members (Stevens Andrade, Page et al., 2016), university excursions (Ozogul, Miller, Reisslein et al., 2019) and summer camps (White, Eberstein, Scott, Ito et al., 2018) were examples of science learning experiences that could not be facilitated with the resources, time and support available in a typical primary science classroom.
Technology^a	Any form of digital educational technology delivered to primary students with the aim of enhancing science learning. Technological innovations may include; Robots (Shiomi, Kanda, Howley, Hayashi, Hagita et al., 2015), Technology Enhanced Curriculum (Varma & Linn, 2012), Augmented Reality (Fleck & Simon, 2013), 3d Games (Lester, Spires, Nietfeld, Minogue, Mott, Lobene et al., 2014) and Learning Management Systems (Field, 2009).

^aDenotes an approach that emerged inductively during the coding process.

Coding and analysis

Initially, essential information was extracted from the 142 research outputs that met the inclusion criteria, including: names of authors, national contexts, publication years, participant numbers, participant ages and participant year levels. These descriptive details were quantified in order to understand the scope of the intervention studies reviewed (Linder & Simpson, 2017). A coding framework (Table 3) was applied in the analysis of the research methodologies and science education interventions. This coding framework addressed broad research methods, deeper research designs and longitudinal components. Outcome measures and concepts were defined as broadly as possible to account for the varied research methods included in this review. Content Knowledge encompasses student understanding of science content mostly in traditional fields of chemistry, biology, astronomy, physics and earth sciences. It should also be noted that science content knowledge is often presented as ‘achievement’ in the literature sample (e.g., Karaçalli & Korur, 2014; Keil, Haney, Zoffel et al., 2009) but the authors selected the term ‘Content Knowledge’ to avoid making inappropriate value statements. Skills-focused research investigates student capacity to conduct scientific investigations and activities with some degree of independence. Attitudes measures relate to student dispositions towards any aspect of science and/or their science learning. To minimise conceptual bias, other concepts and measures were also considered, such as reading comprehension of science texts (e.g., Reutzel et al., 2005). The science focal area codes are based on existing science education literature and the structure of the Australian K-6 Science Curriculum (Australian Curriculum, Assessment and Reporting Authority (ACARA), 2017; NSW Education Standards Authority (NESA), 2017; Potvin & Hasni, 2014). These focal areas can divided into content codes (i.e. Living World, Physical World, Earth and Space & Material World) and process codes (i.e. Working Scientifically, Design and Production & Nature of Science), which are not mutually exclusive. Results are presented in tabular form to map the data extracted from the papers, as well as descriptive form to provide details in relation to the objectives of the review (Peters et al., 2020). For the non mutually exclusive codes, science foci, and interventions, we presented the percentages based on the total sample (142) and code counts to show prominence within both the sample and the specific coding area respectively. This choice was intended to prevent a distorted view of the sample as the importance of multiple codes within research outputs could not be reasonably quantified.

Multiple cycles of both deductive and inductive analysis were used to code the primary science teaching approaches within the interventions to ensure a clear grounding in the science education literature and to ensure appropriate incorporation of unanticipated themes in the findings. The deductive phase began with the development and application of a set of a priori themes based on established primary science education frameworks (Deehan, 2022; Deehan et al., 2017) and wider literature (e.g., Aubusson et al., 2015). The eight themes were: Alternative Conception Targeting, Problem/Project Based Learning, Constructivism/5Es, Cooperative Learning, Cross Curricular Integration (including STEM), Inquiry Learning, Nature of Science Instruction, and Science Beyond the Classroom. The authors then undertook in-depth reading of the studies and used an inductive coding approach (Boyatzis, 1998) to identify emergent, salient themes. Table 4 lists and describes all approaches to primary science teaching that comprised the final coding scheme. Inter-rater reliability was primarily achieved through a series of independent and collaborative coding sessions by the research team. Coders identified interventions for further clarification which then informed the structure of the coding framework. After a coding check with the final iteration of the coding framework the authors reached consensus.

In order to compare the contextually bound impacts of interventions delivered in different educational contexts, as measured by an array of data sources, effect sizes were calculated by the research team. Effect size is a calculation of the degree to which, for two data points, the null hypothesis can be rejected (Vogt, 1999). Effect sizes were calculated for all quantitative research outputs with appropriate comparison points, either within group pre-to-post tests or between group post-tests, and sufficiently detailed descriptive statistics (i.e. numbers, means & standard deviations). For consistency, Cohen’s d was used for comparison between groups of equal size and Hedge’s G was used for groups of different sizes (Peng & Chen, 2014). Common interpretations of small (0.2), moderate (0.5) and large (0.8) formed the basis of these analyses (Cohen & Cohen, 1983; Fritz et al., 2012); which echoes similar effect size classifications in large scale assessments such as Australia’s National Assessment Programme – Literacy and Numeracy (NAPLAN; Australian Curriculum, Assessment and Reporting Authority (ACARA), 2021). This also aligns with the work of Bloom et al. (2008) who report an average yearly effect size gain of 0.23 in science knowledge for middle grade learners. An effect size of 2.0 or greater could be classified as solving the 2-sigma problem in education, meaning that the intervention could be considered as effective as one to one tutoring (Bloom, 1984). In statistical terms, an effect size of 2 or greater would mean that 98% of students would score above the pre-test mean on the post-test (Bloom et al., 2008), a clear marker of desirable growth patterns. This is not to say that all students will achieve similar science outcomes in a particular context, but rather that the overwhelming majority (98%) would be displaying significant growth relative to where the group began, as defined by the class mean; something that most educators would aspire to in any context.

Results

The results of this paper are divided into two sections to address both research questions. The first section presents the descriptive information on the literature sampled, including: publication years, contributing nations, year levels, science foci, intervention approaches and research designs. The second section provides the calculated effect size means and the largest effect size increases in the selected literature.

RQ1: What is the state of primary science intervention research in terms of demographics, pedagogical approaches, and research designs?

This review focused on peer-reviewed journal articles published in English between 2001 and 2021, i.e. the last 20 years. As Figure 2 shows, the greatest number of papers were published between 2008 and 2016. Although no 2021 publications were captured during the review period, it is possible there will be newer publications that are not included in this review.

Figure 2. Publications per year.

Papers were coded according to the country in which the study was conducted, and a total of 26 nations were represented in the review (Figure 3). The majority of studies were conducted in the United States (41.5%), followed by Turkey (22.5%) and Taiwan (7.7%). The remaining 23 nations accounted for 28.3% of the included research outputs, for an average of 1.6 per country. Of interest to the authors was the limited scope of Australian research, with only 3 papers included in the review.

Figure 3. Publications by countries of origin.

This review focuses on primary science interventions, and studies across all primary years were captured within the review. However, as shown in Figure 4, the largest proportion of studies focused on Years 3–6 (86.4%), with only 13.6% focused on the younger primary years (i.e. Kindergarten to Year 2).

Figure 4. Year levels.

A range of science foci and intervention approaches were reported. Table 5 displays the science topics which were the foci of the interventions in the studies reviewed. Most of the research outputs had more than one science focus (m = 1.96). It can be seen that Living World and Working Scientifically were most frequently the focus of the science interventions, with both represented in approximately 21% of studies. In contrast, Material World and Design and Production were the least explored, with each represented in less than 10% of studies. Most of the publications analysed incorporated more than one intervention (m = 2.08; Table 6). The majority of interventions were based on technology (16.7%) such as virtual or augmented reality experiences, online learning environments, robotics, coding, and animation. As might be expected, inquiry learning (14.6%), cross curricular integration (11.2%), and PBL (10.9%) approaches were also prevalent among the studies. Of note is the scarcity of play-based approaches, which was represented in only two studies in the review. The duration of the interventions ranged from a single one-hour session up to three years, however, the majority of the interventions took place at regular intervals over a period of several weeks.

Table 5. Science foci.

Science focus	Number of papers	Percentage of Codes (N = 278)^a
Living world	64	23.0%
Working scientifically	63	22.7%
Physical world	43	15.5%
Earth and space	38	13.7%
Nature of science	38	13.7%
Material world	20	7.2%
Design and production	9	3.2%
Unspecified	3	1.1%

^aCount is not equal to the literature sample (142).

Table 6. Intervention approaches.

Intervention approach	Number of papers	Percentage of Codes (N = 295)
Technology	49	16.6%
Inquiry Learning	43	14.6%
Cross Curricular Integration (including STEM)	33	11.2%
Problem/Project Based Learning	32	10.8%
Cooperative Learning	25	8.5%
Science Beyond the Classroom	24	8.1%
Nature of Science Instruction	20	6.8%
‘Hands-on/Student-centred (NFS)’	19	6.4%
Literacy Support	18	6.1%
Narrative Focus	9	3.1%
Alternative Conception Targeting	8	2.7%
Constructivism/5Es	6	2.0%
Lesson Sequencing	4	1.4%
Concept Mapping	3	1.0%
Play-based Learning	2	0.7%

^aCount is not equal to the literature sample (142).

The reported type of data was predominantly quantitative, with 91 papers (64.1%) representing quantitative designs. 46 papers (32.4%) reported mixed-methods designs, while the remaining 5 papers (3.5%) were qualitative studies. Only 26 longitudinal studies were reported. Methodologically, the studies were typically quasi-experimental (80.3%), with or without a control, and measured a range of outcomes including content knowledge (49.8%), skills (18.6%), and attitudes (26.2%). Randomly assigned experimental designs were identified in a significant minority of studies (19.7%). The sheer number of research outputs featuring a control group (92) greatly strengthens claims for causal and covariant relationships between the use of different student-centred approaches and improved science outcomes for primary learners. Additionally, there was an average of 221 participants per research output, although it should be noted that the average was skewed by some large projects with over 1000 participants (e.g., Harris, Penuel, D’Angelo, DeBarger, Gallagher, Kennedy, Krajcik, Krajcik et al., 2015; Keil, Haney, Zoffel et al., 2009; Sebaganwa, 2013). The effectiveness of the interventions in relation to these outcomes is examined in the next section.

RQ2: How effective are the different approaches employed in the primary science intervention research in terms of enhancing students’ science content knowledge, skills and attitudes within different contexts?

There is compelling evidence for the efficacy of various practices within the selected literature. Figure 5 presents the descriptive statistics (M, N & SD) for the relevant content knowledge, skills, and attitudes effect size data, both within intervention groups from pre-to-post (blue) and between intervention and control groups on post testing (red). The content knowledge data stands out both in terms of mean effect size (2.07) and research focus (n = 81). The mean pre-to-post test effect size of 2.06 suggests that not only has much of the field solved the 2-sigma problem (Bloom, 1984), but the reported content knowledge increases are approximately 900% higher than what may be considered normal progression (Australian Curriculum, Assessment and Reporting Authority (ACARA), 2021) and 400% higher than what may be considered above-average progression (Bloom et al., 2008). Although not as substantial, mean effect sizes for skills (1.01) and attitudes (0.83) can both be classified as large. In terms of control group post-test comparisons, all three domains displayed similarly moderate-to-large mean effect sizes. This relative uniformity is a clear contrast to the outlying within group content knowledge mean effect size. The strong trends for skills and attitudinal growth have positive implications for long-term science and scientific literacy of learners. Indeed, the overwhelming majority (96%) of longitudinal research analysed in this scoping review showed that student science gains were retained after the conclusion of interventions, ranging from small single class (e.g., Blanchard et al., 2010) or multiple class (e.g., Durmuş & Bayraktar, 2010) projects to large whole school (e.g., Amaral et al., 2002) and multi-school (e.g., Bethke Wendell & Rogers, 2013) programmes. The lone exception was a sample of 19 year 2, 3 and 4 students who did not retain their NoS knowledge gains the year after the intervention (Fouad, Masters, Akerson et al., 2015). The following paragraphs will present some of the most effective interventions in terms of student gains in science content knowledge, skills and attitudes.

Figure 5. Effect sizes for content knowledge, skills and attitudes data.

The most effective interventions in terms of pre-to-post test effect size growth within separate research contexts were identified for comparison. Table 7 presents the top 10 within group pre-to-post test effect sizes for content knowledge, skills and attitudes presented in the literature sample. 25 research outputs were included in these rankings because 5 articles occupied more than one position across the rankings (H. Akcay & Yager, 2010; Girod, Twyman, Wojcikiewicz et al., 2010; Panasan & Nuangchalerm, 2010; Prokop, Tuncer, Kvasničák et al., 2007; Safaruddin, Ibrahim, Juhaeni, Harmilawati, Qadrianti et al., 2020). There was considerable variation in terms of science approaches and nations represented across the three rankings. In descending order of frequency, the science approaches were: Inquiry Learning (11), Project/Problem-based Learning (8), Technology (8), Science Beyond the Classroom (7), Cooperative Learning (6), Nature of Science Instruction (5), Cross Curricular Integration (4), ‘Hands-on/Student-centred (NFS)’ (3), Constructivism/5Es (3), Literacy Support (1) and Narrative Focus (1). With 10 contributing nations, these effective interventions have shown to be viable globally. Similarly to the full sample (142), the USA, Turkey and Taiwan were the nations that were most represented in these top 10 rankings. Despite accounting for 41.5% of all literature included in this scoping review, only 20% of the top 10 outputs were from the USA. Conversely, Turkey was overrepresented in the top 10 rankings (40%) relative to the full sample (22.5%).

Table 7. The 10 highest reported content knowledge, skills and attitudes effect size increases.

Top 10 Content Knowledge Effect Sizes			Top 10 Skills Effect Sizes			Top 10 Attitudes Effect Sizes
Source – Nation	Teaching Approaches	Effect Size	Source – Nation	Teaching Approaches	Effect Size	Source – Nation	Teaching Approaches	Effect Size
(Prokop, Tuncer, Kvasničák et al., 2007)^a – Slovakia	Science Beyond the Classroom	14.39	(Safaruddin, Ibrahim, Juhaeni, Harmilawati, Qadrianti et al., 2020)^a – Indonesia	Problem/Project Based Learning, Technology	4.99	(Safaruddin, Ibrahim, Juhaeni, Harmilawati, Qadrianti et al., 2020)^a – Indonesia	Problem/Project Based Learning, Technology	8.33
(Ocak, 2010) – Turkey	Cooperative Learning, Cross Curricular Integration	6.91	(H. Akcay & Yager, 2010)^a – USA	Problem/Project Based Learning, Inquiry Learning, Science Beyond the Classroom, Nature of Science Instruction	2.07	(H. Akcay & Yager, 2010)^a – USA	Problem/Project Based Learning, Inquiry Learning, Science Beyond the Classroom, Nature of Science Instruction	2.95
(Panasan & Nuangchalerm, 2010)^a – Thailand	Inquiry Learning	6.60	(Cuevas et al., 2005) – USA	Inquiry Learning	1.97	(Prokop, Tuncer, Kvasničák et al., 2007)^a – Slovakia	Science Beyond the Classroom	2.25
(Liu, Peng, Wu, Lin et al., 2009) – Taiwan	Technology, Constructivism	6.10	(Taşkın-Can, 2013) – Turkey	Cross Curricular Integration	1.89	(Genc, 2015) – Turkey	Inquiry Learning, Science Beyond the Classroom	1.85
(Panasan & Nuangchalerm, 2010)^a – Thailand	Problem/Project Based Learning	5.97	(Brown et al., 2010) – USA	Technology, Literacy Support	1.81	(Girod, Twyman, Wojcikiewicz et al., 2010)^a – USA	Inquiry Learning, Nature of Science Instruction	1.17
(Liou, Yang, Chen, Tarng et al., 2017) – Taiwan	Inquiry Learning, Technology, Constructivism	5.23	(Malamitsa, Kasoutas, Kokkotas et al., 2009) – Greece	Nature of Science Instruction	1.48	(Yazicioğlu & Çavuş-Güngören, 2019) – Turkey	Technology	1.01
(Abdi et al., 2013) – Iran	Cross Curricular Integration	4.68	(Ting & Siew, 2014) – Malaysia	Science Beyond the Classroom	1.37	(B. Akcay & Akcay, 2015) – Turkey	Inquiry Learning, Cooperative Learning, Hands on/Student Centred	0.80
(Girod, Twyman, Wojcikiewicz et al., 2010)^a – USA	Inquiry Learning, Nature of Science Instruction	4.59	(Can et al., 2017) – Turkey	Problem/Project Based Learning, Cooperative Learning, Hands on/Student Centred	1.33	(Erdoğan, 2015) – Turkey	Inquiry Learning, Science Beyond the Classroom, Cross Curricular Integration	0.54
(Randler & Hulde, 2007) – Germany	Hands on/Student Centred	4.59	(Kaynar, Tekkaya, Çakiroğlu et al., 2009a) – Turkey	Constructivism	1.24	(Meluso, Zheng, Spires, Lester et al., 2012) – USA	Cooperative Learning, Technology, Narrative	0.53
(Karaçalli & Korur, 2014) – Turkey	Problem/Project Based Learning, Cooperative Learning	3.50	(Wu & Hsieh, 2006) – Taiwan	Problem/Project Based Learning, Inquiry Learning, Cooperative Learning	1.23	(Ercan, 2014) – Turkey	Technology	0.48

^aDenotes publications that appear in more than one list.

The top 10 highest content knowledge effect sizes were linked to interventions featuring an average of 1.6 science approaches, including fairly even representation of Inquiry Learning (3), Project/Problem-based Learning (2), Cooperative Learning (2), Technology (2), Cross Curricular Integration (2) and Constructivism/5Es (2). Perhaps due to the field’s focus on content knowledge or measurement differences, the average content knowledge effect size gain eclipsed the threshold for solving the 2-sigma problem (Bloom, 1984). The work of Prokop, Tuncer, Kvasničák et al. (2007) was a major outlier, as the students’ mean knowledge gains across a series multiple choice and drawing based assessments (Cohen’s d = 14.39) were more than double the second largest knowledge increases identified. Regardless of the impact of any antecedent variables, it is emblematic of the existing literature on the efficacy of Science Beyond the Classroom (e.g., Bathgate et al., 2015; Leblebicioglu, Metin, Yardimci, Cetin et al., 2011; White, Eberstein, Scott, Ito et al., 2018). Ocak (2010) found that a 25-day programme wherein students completed a variety of rich, cross curricular science activities in cooperative learning groups resulted in very large science knowledge increases (Cohen’s d = 6.91). Similarly large science content knowledge increases were observed in a sample of year 5 Thai students who participated in 8-lesson Project-based learning (Cohen’s d = 5.97) and Inquiry (Cohen’s d = 6.60) units. Virtual and Augmented reality technologies (Liou, Yang, Chen, Tarng et al., 2017) and mobile tablets (Liu, Peng, Wu, Lin et al., 2009) were innovations that proved effective when combined with established student centred practices. In terms of improving primary students’ science content knowledge, the field appears to be on an evolutionary trajectory as a myriad of student-centred approaches have proven to be extremely successful across different educational contexts.

The 10 most effective interventions in developing primary learners’ scientific skills featured an average of 1.9 science approaches, with Project/Problem-based (4) and Inquiry (3) learning approaches being most common. Safaruddin, Ibrahim, Juhaeni, Harmilawati, Qadrianti et al. (2020) used survey and observation checklists to investigate the impact of a media-enhanced project-based learning intervention on 29 year 5 students’ science motivation and process skills. The very large science process skills gains (Cohen’s d = 4.99) of these students were more than double those reported in H. Akcay and Yager’s (2010) evaluation of the Science/Technology/Science (STS) programme (Cohen’s d = 2.07), an exemplar of an authentic multifaceted intervention including: student agency, local relevance, student inquiry, science beyond the classroom, problem solving. Similarly rich Project/Problem-based interventions have also shown to substantially improve year 6 students’ capabilities in constructing scientific explanations (Wu & Hsieh, 2006) and kindergarten students’ science process skills (Can et al., 2017), suggesting a degree of flexibility in primary learning utility. In addition to further establishing more common approaches, such as cross curricular integration (Taşkın-Can, 2013), science beyond the classroom (Ting & Siew, 2014) and the 5Es model (Kaynar, Tekkaya, Çakiroğlu et al., 2009a), strong evidence for the impact of deeper epistemological treatments on primary students’ science skills has emerged. Brown et al. (2010) found that the use of everyday language as a bridge to science nomenclature correlated with substantial development in primary students’ written and oral science language skills (Cohen’s d = 1.81). Furthermore, explicit Nature of Science instruction, focussing on argumentation and the history of science, has shown to positively influence the critical thinking skills of year 6 students (Cohen’s d = 1.48; Malamitsa, Kasoutas, Kokkotas et al., 2009). The body of evidence indicates that explicit scaffolding to support learners in student-centred, authentic approaches leads to the largest increases in primary students’ science skills.

Inquiry (5), Science Beyond the Classroom (4) and Technologies (4) are the predominant approaches amongst the top 10 science attitudinal effect sizes reported within the literature sample. With an average of 2.2 approaches, it appears that multifaceted interventions are associated with improving primary students’ science attitudes. A noteworthy observation was that the four of the five highest ranked research outputs in terms of student attitudinal improvement were also included in the content knowledge and skills top 10 lists. This may speak to the overall effectiveness of project-based learning (Safaruddin, Ibrahim, Juhaeni, Harmilawati, Qadrianti et al., 2020), the STS programme (H. Akcay & Yager, 2010), science beyond the classroom (Prokop, Tuncer, Kvasničák et al., 2007) and inquiry learning enriched by nature of science instruction (Girod, Twyman, Wojcikiewicz et al., 2010): the overlap across these ranking lists may simply signal conceptual links between the development of knowledge, skills and attitudes in primary science learners. Moreover, additional research has echoed these studies. For example, students who participated in an extended, independent inquiry programme where they engaged in scientific processes, such as collecting data from community members, and deep reflection displayed far better attitudes towards science (Cohen’s d = 1.85; Genc, 2015). Other approaches that led to large increases in primary learners’ reported science attitudes were games-based learning activities (Yazicioğlu & Çavuş-Güngören, 2019) and teacher-adapted STS (B. Akcay & Akcay, 2015). An interesting observation was that the 8th, 9th and 10th largest effect sizes were moderate (Ercan, 2014; Erdoğan, 2015; Meluso, Zheng, Spires, Lester et al., 2012). This may be a sign that more attention to attitudinal change is warranted. Without making definitive statements about the efficacy of the ranked interventions, it is worth noting that the science approaches within the top 10 attitudinal ranking diverge from the relative similarity between the content knowledge and skill rankings. Indeed, Inquiry learning (5), Science Beyond the Classroom (4) and Technology (4) feature far more frequently in the top 10 research outputs as arranged by attitudinal effect sizes. However, due to the limited sample size of this scoping review, this observation should be treated with scepticism and not be considered beyond speculation until further research on this area has been completed. These nuanced points will be addressed further in the following discussion.

Discussion

This review analysed 142 research outputs published between 2001 and 2020. Our procedures did not yield any publications from 2021, thus we can only at best claim this scoping review to be representative but not exhaustive. Still, this scoping review consolidates a sample of research in a field requiring additional focus (Skamp, 2020) in a way that is both focused and replicable whilst still representing a selection of evaluative research designs across a variety of national contexts. In this discussion we will answer the two research questions, provide informed speculation on the trajectory of primary science education, outline the research limitations and make recommendations for stakeholders in primary science research. This will further clarify ‘best practice’ in primary science teaching by outlining the evidence based surrounding the multitude of different primary science approaches.

In addressing the first research question (‘What is the state of primary science intervention research in terms of demographics, pedagogical approaches, and research designs?’), the primary science intervention literature can be classified as an established field of evaluative research due to the preponderance of contributing nations (26) and the high quality of the research designs, with the majority quasi-experimental and experimental designs allowing researchers to make covariant and causal arguments respectively. Most importantly, the high percentage of research outputs with a control group (64.8%) helps to strengthen collective arguments for the effectiveness of a myriad of student-centred approaches relative to more traditional teacher-centred approaches. Another contributing factor to the strength of this literature sample was the high mean number of participants per publication (221). The established nature of the primary science intervention literature should not be taken for granted as more emergent fields lack the same coherence and developed evidence bases. For example, Dyment and Downing’s (2020) systematic review of 492 articles on online teaching practice in ITE programmes found much of the literature to be fragmented and siloed.

This is not to say that the primary science literature analysed for this scoping review is beyond reproach. As mentioned in the introduction, regular systematic analysis of bodies or literature can help to ensure more cohesive research trajectories within sub-discipline areas, such as Potvin and Hasni’s (2014) systematic review of research (n = 228) addressing science and technology intervest, motivation and attitudes of K-12 learners wherein they found an overabundance of research focusing on science attitudinal differences across genders. Despite the strength of the research in this scoping review, there are areas that could be further developed. Although there are 26 contributing nations, the USA, Turkey and Taiwan account for nearly three quarters (71.7%) of the research output, so it would behove all countries to continue focusing on primary science research to ensure their different cultural and educational characteristics are accounted for the body of literature. There is also a dearth of research focusing on the early years of primary school, with only 13.6% of the literature focused on Kindergarten to Year 2 (ages 6–8). This remains a clear gap in science education research relative to rapidly developing research programmes in early childhood (e.g., MacDonald et al., 2020) and middle school (e.g., McNeill & Krajcik, 2008; Sahin & Yilmaz, 2020) transitions. Living World (Biology & Environmental Sciences) and Working Scientifically are the most established science foci as they account for over 23.0% and 22.7% of the codes respectively, followed by Physical World (Physics) (15.5%), Earth and Space (Astronomy & Geology) (13.7%) and Nature of Science (13.7%). The limited focus on Material World (Chemistry) (7.2%%) and Design and production (3.2%) could represent gaps in the literature that relate to issues of time, resourcing and primary teachers’ preferences for natural sciences (Brigido et al., 2013). This could also be explained by researcher preferences and contextual factors that restrict teacher choice. Indeed, in national contexts outside of Australia the absence of Material World (Chemistry) research may reflect differences in the structure of science syllabi. Methodologically, arguments can also be made for more longitudinal research (18.3%) and increased focus on skills development (18.6%) to ensure the field remains aligned to the central goal of developing learners’ science and scientific literacies (Bybee, 1997; Roberts & Bybee, 2014).The established status of the primary science intervention evaluation literature, both in terms of research outputs and contributing nations, should strengthen discussions about ‘best practice’ in primary science education.

To answer question two (‘How effective are the different approaches employed in the primary science intervention research in terms of enhancing students’ science content knowledge, skills and attitudes within different contexts?’), effect sizes were calculated or collected for measures of science content knowledge, skills and attitudes which produced appropriate interval or ratio data. In another sign of the methodological rigour of the literature sample, effect sizes were extracted from 101 (71%) of the 142 research outputs. In terms of within group pre-to-post test effect sizes, the very large mean (Cohen’s d = 2.07) for content knowledge gains in primary learners not only exceeds what might be considered normal and above average growth (Australian Curriculum, Assessment and Reporting Authority (ACARA), 2021; Bloom et al., 2008), but solves the 2-sigma problem (Bloom, 1984); meaning that the collective array of student-centred approaches is comparable to one-to-one and small group tutoring in terms of reported student content knowledge increases. Most importantly, an effect size of 2 or greater would mean that 98% of students would score above the pre-test mean on the post-test (Bloom, 1984); meaning that many of the science approaches evaluated in this scoping review could help to address issues of equity at the heart of modern education (Ling & Nasri, 2019). Although the science skills (Cohen’s d = 1.01) and attitudes (Cohen’s d = 0.83) mean increases were not comparable to the content knowledge increases, for an array of potential reasons such as the greater emphasis on content knowledge and operational differences in research designs, the evidence still suggests that the array of student-centred approaches can be effective in improving science and scientific literacy by improving learners’ science dispositions and capabilities in applying science skills and knowledge beyond the classroom; areas of global concern as 90% of OECD nation fall below the high international benchmark in Year 4 science TIMSS scores (M. O. Martin et al., 1997; Thomson et al., 2020a). The control and intervention post-test effect sizes provide further evidence of the superiority of the student-centred science approaches in enhancing primary students’ science content knowledge (Cohen’s d = 0.80), skills (Cohen’s d = 0.73) and attitudes (Cohen’s d = 0.83) relative to the passive, teacher-centred practices of rote learning, lectures, textbook work and ‘cook-book’ investigations that have traditionally permeated primary science education (Goodrum et al., 2001). Perhaps most importantly, 96% of the longitudinal research showed that primary students’ improved science outcomes were retained in the absence of formal interventions; a signal that many of these approaches may hold utility in longer term goals of science and scientific literacy enhancement.

Closer analyses of the largest reported pre-to-post test effect size increases in primary students’ content knowledge, skills and attitudes reflects the breadth of student centred approaches observed in the full literature sample. Aside from the greater prevalence of Project/Problem-based approaches on the skills list, both the content knowledge and skills top 10 lists were of similar composition, with many of the approaches presented with the established primary science education frameworks (Deehan, 2022; Deehan et al., 2017), such as Project/Problem-based Learning, Inquiry Learning, Cooperative Learning, Science Beyond the Classroom, Nature of Science Instruction, Technology, Cross Curricular Integration, ‘Hands-on/Student-centred (NFS)’ Learning and Constructivism/5Es, appearing on both lists. Taken together, these approaches appear to embody primary educators’ potential in effectively de-siloing science education through outward facing pedagogies and cross curricular integration (NAS, 2021). It is also worth noting that Inquiry Learning, Science Beyond the Classroom, and Technology are more associated with the greatest attitudinal changes, thus suggesting that authenticity is a key to improving early science dispositions. The principles of authenticity and student engagement are also widespread in the delivery of preservice primary science education (Deehan, 2022, 2021), which indicates alignment of principles across levels of science education.

However, notions of authenticity and student engagement are far from simple, as such principles must be realised through complex pedagogical repertoires that factor in student characteristics, educator traits and curriculum requirements whilst addressing time and resourcing limitations. What is more, the research literature often represents classroom contexts with considerable external support that are atypical in more primary education systems. Without the robust external support that is often implicit in ‘researched classrooms’, the resilience and pedagogical pinballing (Mitchell et al., 2015) required to incorporate the primary science evidence base in classroom settings may prove challenging for primary teachers. Many of these teachers are still developing their content knowledge, confidence and advanced pedagogical repertoires (Toma & Greca, 2018) and can sometimes rely too heavily on transmission (Goodrum et al., 2001; Goodrum & Rennie, 2007) and wide-ranging conceptions of hands on learning (Kleickmann et al., 2016). Herein lies the central divide that explains the contradiction between the established research evidence base and the widespread challenges to the provision of science education: the research literature often represents classroom contexts with considerable external support, as a result of being the subject of academic research, that may not fully reflect the full systemic challenges to the provision of high-quality primary science education (i.e., time, resourcing, etc.).

In the pursuit of determining the effectiveness of primary science interventions, feasibility and scalability are not always considered overtly or fully, as research outputs often reflect more artificial, ‘best-case’ scenarios where teachers are afforded considerable support. Simply put, many of the systemic challenges that affect teachers’ science teaching practices, such as time (e.g., AITSL, 2021; Crump, 2005; Jenkinson & Benson, 2010), resourcing (e.g., Gonski, 2011; Rowe & Perry, 2020), socio-economic issues (Halsey, 2018; Sullivan et al., 2018) and curricular imperatives (e.g., Akar, 2018; Australian Primary Principals Association (APPA), 2014), are not or cannot be meaningfully considered, which limits the ecologically validity of much of the research, and, by extension, this scoping review. For example, the summer camps (e.g., Erdoğan, 2011; Larson, Castleberry, Green et al., 2010), university visits (e.g., Gluckman, Vlach, Sandhofer et al., 2014; Naizer, Hawthorne, Henley et al., 2014; Ozogul, Miller, Reisslein et al., 2019; Sengupta & Farris, 2012; Weinberg, Basile, Albright et al., 2011), museum visits (e.g., Martin, Durksen, Williamson, Kiss, Ginns et al., 2016; Potvin & Hasni, 2014), complex technological innovations (e.g., Evagorou, Korfiatis, Nicolaou, Constantinou et al., 2009; K. Y. Chin et al., 2014) and science community partnerships (Leblebicioglu, Metin, Yardimci, Cetin et al., 2011; Molina, Borror, Desir et al., 2016) are not always feasible in geographically isolated or resource-poor educational contexts, which somewhat limits their utility in affecting broad improvements in the quality of primary science education. To some extent, feasibility issues may be addressed by increasingly cost-effective and accessible technology tools but differing rates of uptake is likely to exacerbate socio-economic divides (Gonski, 2011; Rowe & Perry, 2020). Scalability also remains an issue as the academics and expert teachers who deliver the interventions within the science education literature are not necessarily reflective of the reflect attitudes, capacities, and circumstances of the majority of generalist primary teachers, many of whom already work considerably more than their contracted hours (AITSL, 2021). The outcomes associated with interventions delivered by academics (e.g., Brackmann et al., 2017; Dalacosta et al., 2009; Diakidoy & Kendeou, 2001; Vosniadou, Ioannides, Dimitrakopoulou, Papademetriou et al., 2001; Weinberg, Basile, Albright et al., 2011; Wu & Hsieh, 2006) may be particularly challenging to replicate at scale due to the shortage of specialist science teachers (Fraser et al., 2019; Hobbs, 2013). Another problem is that intervention descriptions in research literature often lack the detail necessary for teachers to translate theory into classroom teaching practice (Abdi et al., 2013; Panasan & Nuangchalerm, 2010; Weinberg, Basile, Albright et al., 2011); however, efforts have been put forward in this space as academics have supported educators with large scale professional development programmes (e.g., B. Akcay & Akcay, 2015; Amaral et al., 2002; Harris, Penuel, D’Angelo, DeBarger, Gallagher, Kennedy, Krajcik, Krajcik et al., 2015; Molina, Borror, Desir et al., 2016; Wendell & Rogers, 2013) and more formal school-university partnerships (e.g., Hobbs et al., 2015, 2018). The role of practitioner journals, such as ‘Science and Children (published by the National Science Teachers Association in the U.S.)’ and ‘Science Eduction News’ (published by the Science Teachers Association of New South Wales in Australia), in advocating and supporting enhanced science teaching practices should also be acknowledged. It is essential that intervention design and implementation support mechanisms are sufficiently dealt with in research to ameliorate the potential for issues of feasibility and scalability to unintentionally entrench educational disadvantage and inequity.

It is acknowledged that this review is subject to limitations. The review is dependent upon the availability of information on the topic, and it is possible that relevant sources of information have been omitted (Peters et al., 2020). Further, the utilisation of a scoping review methodology limits the ability to rate and fully quantify the quality of the evidence provided (Peters et al., 2020). This scoping review may serve as a basis for future systematic reviews that seek to more formally assess the quality of evidence in regards to more explicit questions. Further to this point, the inclusion of an array of research designs and data collection instruments means that definitive statements about the effectiveness of different science interventions across contexts cannot be made; rather, the analyses provide insight into contextually bound improvements in primary students’ science content knowledge, skills and attitudes. Therefore, this review provides a useful catalyst for primary science teachers to take a data-informed approach to their practice but should not be viewed as a replacement for their own professional insights and evidence. Like all academic reviews, this manuscript is limited by the biased conceptualisations of the researchers. The coding framework is defensible, as it is grounded in existing science education literature, yet it remains contestable, and the authors urge other researchers to pursue different analyses in this space. Relatedly, the quantitative approach affords little room for nuances and differences within each intervention. As approaches were simply coded as either being included or absent, no insight into the complex relationships amongst approaches could be offered. For example, although Inquiry Learning and Nature of Science Instruction may both be coded within a paper, the Inquiry Learning may be a supplementary part of a large suite of activities contributing to Nature of Science Instruction (e.g., Girod, Twyman, Wojcikiewicz et al., 2010). It is also challenging to compare science content knowledge, skills and attitudes both within and across research outputs due to differences in emphasis and operationalisation. For example, attitudinal Likert scale data (e.g., Hwang, Wu, Chen, Tu et al., 2016) may not allow for an absolutely zero score in the same way as a multiple choice science test (e.g., Kandlhofer & Steinbauer, 2016), potentially artificially diminishing reported attitudinal effect sizes relative to content knowledge. Finally, any interpretation of the findings and resultant speculation of this scoping review should be cautious as the authors can offer little insight into the unknown or unreported antecedent factors, such as teacher traits, student-traits, fidelity of implementation and/or methodological issues, that will have influenced findings.

This review has revealed several promising directions for future practice, research and strategy. Firstly, there are clear implications for the continuance of student-centred trajectories in primary science teaching. All primary science educators should now begin to or continue to explore the use of student-centred practices in their professional and pedagogical experience repertoires (Loughran et al., 2004), supporting stakeholders (such as academics, administrators, departments and governments) should work to remove barriers to student-centred classroom practice wherever practically possible. Secondly, and germanely, science education researchers should continue to focus on addressing the issues of feasibility and scalability by working with preservice teachers, inservice teachers and other primary stakeholders to help develop the develop the dispositions and skills, alongside advocating for the professional conditions, needed for long-term, reflexive student-centred primary science teaching practice. Professional Learning Networks (PLNs; e.g., Fentie, 2019), professional development programmes (e.g., Hume, 2012), large school programmes (e.g., B. Akcay & Akcay, 2015; Amaral et al., 2002; Harris, Penuel, D’Angelo, DeBarger, Gallagher, Kennedy, Krajcik, Krajcik et al., 2015; Molina, Borror, Desir et al., 2016; Wendell & Rogers, 2013) and school-university partnerships (e.g., Hobbs et al., 2015, 2018), are all proving to be fruitful contributions to this end. However, the focus should remain primarily on the practice and impact of regular primary science teachers, be they generalists or specialists. It would also be helpful for the editors and reviewers of academic journals to, where possible, place greater emphasis on the detail and quality of intervention descriptions in their evaluation of manuscripts to aid those who choose to draw on academic sources to inform practice. In lieu of clear procedural detail regarding the implementation of very broad approaches, such as inquiry learning and project/problem based learning, underdeveloped notions of hands on learning or unsophisticated pedagogical imitations may be unintended consequences of the sincere desires of educators to replicate the student outcomes reported in the science literature. Thirdly, although the primary science intervention evaluation literature can be considered established, areas for additional research were identified in this scoping review, including more: research into relatively underserviced domains such as Material World (chemistry), longitudinal research designs, and operationalisation of skills development within methodologies.

This review also helps to consolidate advancements in primary science education practice so that teachers, academics, and other education stakeholders can direct their limited time towards further advancing the collective understanding of effective science teaching practice, rather than reacting to possibly ill-informed and detrimental political (e.g., Hodson, 1994; Lall & Vickers, 2009) and social (e.g., Shine, 2018, 2020; Shine & Rogers, 2020) influences and imperatives. For example, this could enable more attention to communication practices with primary science classrooms, such as classroom dialogue (France, 2021) and Argument Driven Instruction (Çetin & Eymur, 2017; Choi et al., 2015), would also help with the translation of academic evidence to feasible and scalable student-centred classroom practice. Most importantly, this review clearly highlights the dearth of research focused on the early years of primary school (i.e. K-2) and, moreover, the scarcity of interventions specifically designed to support primary science education in the early years. This is concerning in light of research which indicates that K-2 teachers may experience higher levels of uncertainty around their science teaching practices (Sandholtz & Ringstaff, 2011, 2013, 2016), which threatens to further aggravate existing trends of early science disengagement (Ali et al., 2013; DeWitt et al., 2014; Lindahl, 2007; Potvin & Hasni, 2014; Said et al., 2016) and potentially bypasses opportunities for rich science learning in the early years (Eshach & Fried, 2005).

There remains a pressing need for consolidation of primary science education through literature reviews, meta-analyses and meta-syntheses because it is an area that remains underrepresented in scholarly output relative to secondary and post-secondary contexts (Skamp, 2020; Wright & Park, 2022). As with any systematic review process, the specific inclusion criteria will always exclude insightful and valuable research that should influence wider thinking and discourse related to science education. In this scoping review, the focus on articles published after 2001 ignored the influence of prior research. Rosebery et al. (1992) work is illustrative as they showed a small sample of bilingual students displayed employed improved scientific reasoning after engaging in a collaborative, inquiry approach to science. Additionally, the focus on defined, pre-to-post test data collection periods investigating growth in achievement and performance excluded a plethora of ecologically valid, rich qualitative research projects that explore the complex interactions that occur in science classrooms. One such example would be Wilmes and Siry’s (2021) case study of a pluralistic student’s experiences in science inquiry. Their multimodal interaction analyses revealed the iterative, diverse nature of how science knowledge and scientific literacy can be embodied in classroom contexts. Ideally, further primary science reviews should capture nuanced interactions within classroom settings. Further thematic reviews (e.g., Aubusson et al., 2015) and systematic reviews of specific teaching approaches, such as those produced for cross curricular integration (Gresnigt et al., 2014) and flipped classrooms (Wright & Park, 2022), and specific programmes, such as the review conducted for the Primary Connections programme (Aubusson et al., 2015 et al., 2019) would complement the broader scoping review of primary science interventions presented in this paper.

Conclusion

This scoping review has outlined and described the impact of a varied selection of student-centred approaches on primary students’ science content knowledge, skills and attitudes across different national contexts. Furthermore, the impacts of these student-centred approaches, relative to more traditional teacher-centred approaches, have been quantified. While there remain some areas for further development in the field, with additional focus on K-2 science education perhaps the most pressing need, the larger issues of feasibility and scalability will need to continue to be addressed if the last 20 years of research in primary science education is to be enable education systems to produce scientifically and science literate citizens. This scoping review should inform the work of educational researchers and teacher educators as they seek to enhance the quality of primary science education. In particular, the framing of the role of primary teachers in this review shows that they are worthy of support and advocacy and should be perceived through a deficit lens. We also hope this publication can empower primary educators, either directly through this article or indirectly through stakeholder influence and dissemination, in the selection and justification of their science teaching practices. We believe much of the literature presented in the review shows the high-quality science education which primary teachers are capable of facilitating, and therefore should reinvigorate efforts by other educational stakeholders to address systemic hindrances to the provision of high-quality science education. In sum, this body of literature provides compelling evidence for the effectiveness of student-centred approaches as ‘best practice’ in primary science education practice to empower educational stakeholders to inform and justify their educational choices when confronted by contradictory, unsupported public discourse and political imperatives.

Acknowledgments

We thank Benjamin Thompson for his contributions to the literature search process for this scoping review and Lauren Brumby for assisting in the conceptualisation of the methodology.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Charles Sturt University.

Notes on contributors

James Deehan

James Deehan is a Lecturer in Teacher Education in the School of Education at Charles Sturt University, Bathurst, Australia.

Amy MacDonald

Amy MacDonald is Associate Head of School (Research and External Engagement) and Associate Professor of Early Childhood Mathematics Education in the School of Education at Charles Sturt University, Albury-Wodonga, Australia.

Christopher Morris

Christopher Morris is a current classroom teacher and a Research Assistant in the School of Education at Charles Sturt University, Bathurst, Australia.