Introduction & background
Searching for evidence about ‘what works’ has been pervasive in the fields of relationships and sexuality education (RSE)1 (Grunseit et al. Citation1997; Jones et al. Citation2009; Kirby, Laris, and Rolleri Citation2007; Poobalan et al. Citation2009; Pound et al. Citation2017). Programme evaluations have contributed vital insights into RSE programme development, implementation, and evaluation (Denford et al. Citation2017; UNESCO Citation2018; UNFPA Citation2014). Some recent reviews suggest that experimental designs dominate in this field (Ivanova et al. Citation2020; Rocha, Silva, and Duarte Citation2022), while others highlight the diversity of methodologies and methods used (Goldfarb and Lieberman Citation2021). Despite varied evaluation efforts, there is a gap in evidence concerning RSE programmes for younger children in primary (elementary) schools,2 with most programme evaluations having been undertaken with older children in secondary (high) schools (Roien, Graugaard, and Simovska Citation2018). This is important because writers have stressed that RSE should be a lifelong process (Ketting, Friele, and Michielsen Citation2016). Therefore, programme evaluations must be time- and context-sensitive (Michielsen et al. Citation2016).
For many years, researchers have emphasised that preparing students to manage the physical, social and emotional changes associated with puberty is fundamental (Crockett et al. Citation2019; Goldman Citation2010). At the very least, it is important because students have the right to education about their bodies and relationships consistent with their evolving capacities (Ketting and Winkelmann Citation2013). Given that the onset of puberty occurs at an earlier age than in past decades and behavioural norms change, providing developmentally and culturally appropriate RSE in primary schools is crucial (Collier-Harris and Goldman Citation2017; Gilbert and Lamb Citation2019; Suleiman et al. Citation2017). Negative socio-emotional outcomes for students who begin puberty before their peers may be prevented via the provision of early and universal puberty education (Warren and Yu Citation2016). Similarly, RSE programme effectiveness can be maximised when students experience RSE before they engage in ‘sexual activity’ (Goldfarb and Lieberman Citation2021). Yet relatively little is known about the features of RSE programmes in primary schools and how programme evaluations can best be conducted.
Against this background, we conducted a review of RSE programmes implemented in primary schools using a systematic search and screening process (Moher et al. Citation2009). The review was informed by principles of realist evaluation (Pawson and Tilley Citation1997), which involves identifying contexts, mechanisms, and outcomes to generate programme theories that can explain why, how and for whom programmes may work (Greenhalgh et al. Citation2015). The purpose of the review was to generate from existing research, programme theories about the functioning of RSE in primary school. By programme theories we mean ideas about the conditions and ways in which programmes may influence change in participant behaviour or reasoning at the individual student level (Marchal, Kegels, and van Belle Citation2018; Pawson and Tilley Citation1997). We searched for contexts in which school-based RSE programmes might trigger (i.e., influence) mechanisms that can change outcomes for primary school students: a chain of effects that is central to realist evaluation. We focused on upper primary school (Years or Grades 4 to 6, or ages 9–12.5 years) because the review was conducted to guide the research design for an evaluation of an established Australian RSE programme for upper primary school children known as the All School programme, developed and delivered by True Relationships and Reproductive Health in Queensland, Australia (see https://www.true.org.au/all-school). The research was conducted in partial fulfilment of a doctoral degree study by the first author and supervised by the co-authors.
We used an existing RSE reference database established for a systematic review of universal school-based RSE for children aged 4–18 years, led by this paper’s third author and conducted by the paper’s co-authors. This database houses search results from six academic databases: ERIC (EBSCOhost), Education Source (EBSCOhost), Academic Source Elite (EBSCOhost), APA PsycInfo (EBSCOhost), A+ (Informit), and Education Database (ProQuest). The original search strategy included the keywords ‘respectful relationships’ as well as ‘sex education’ and ‘school*’ not ‘higher education’. Searches were conducted between August and September 2019.
Materials and methods
Search strategy and item selection
The RSE database (N = 7,952 references) was searched using the key terms ‘elementary’ and ‘primary’ used in the title, abstract, or keywords of records published since 2005. This time period was used to mark the introduction of the first of several technical guidelines for RSE in the new millennium (FoSE Citation2020; SIECUS Citation2004; UNESCO Citation2018; WHO and BZgA Citation2010). An EndNote library was created for new results (n = 618). After duplicates had been removed, 469 records were uploaded to Covidence, an online systematic review project management tool.
Titles, abstracts and the full text of records were screened and selected by the first author in two iterations. This process focused on identifying records in which contexts, mechanisms, and outcomes could be identified, congruent with the type of evidence synthesis recommended for realist evaluation (Wong et al. Citation2013). Inclusion and exclusion criteria shown in were applied to the titles and abstracts resulting in 112 included records. Relevance was prioritised over rigour for records that could provide rich information on RSE programme theories. For example, RSE programme manuals were included (DEECD (Department of Education and Early Childhood Development) Citation2011) as well as commentaries and journal editorials.
Next, the inclusion and exclusion criteria shown in were applied to full text screening. Figure 1 shows the PRISMA flowchart. Nine studies were included in the review.
Data extraction and synthesis
Deductive thematic analysis (Boyatzis Citation1998; Terry et al. Citation2017) was used. Data extraction and synthesis was undertaken by the first author and reviewed by the co-authors. Records were manually coded in EndNote and extracted to a standardised template in MS Excel. The spreadsheet column headers comprised record number, extract and code. Codes were assigned for contexts, mechanisms and outcomes as defined in realist evaluation (Pawson and Tilley Citation1997):
Contexts refer to the conditions in which programmes are embedded and were coded as individual, cultural, or structural contexts (De Souza Citation2013; Pawson and Tilley Citation1997). Individual and cultural contexts were open coded. Structural contexts were coded at one of three levels reflecting social-ecological frameworks (Bronfenbrenner Citation1996; De Souza Citation2015; Lawson Citation2017). The macro-level involved broader patterns organising systems at international and national level (e.g., regulations, institutions, policies and infrastructure); the meso-level involved the links connecting individuals with their environment rather than the most immediate setting (e.g., partnerships); and the micro-level involved relationships between individuals and the organisation in which they are mostly involved (e.g., student-school) (Lawson Citation2017).
Mechanisms (hereafter ‘programme mechanisms’) refer to interactions at the interface between programme resources and programme participants (e.g., a programme element that triggers a change in behaviour) (Pawson and Tilley Citation1997; Westhorp Citation2018). Mechanisms were open coded.
Outcomes refer to programme effects that enabled assessment of the extent to which programme objectives are achieved (Pawson and Tilley Citation1997). Outcomes were open coded.
Theme development involved clustering contexts, mechanisms and outcomes, reviewing text extracts, comparing these to each other for clear differentiation, and assigning descriptive labels. Theme definition involved drafting and re-drafting narrative descriptions of each theme and comparing these against each other to achieve clarity and internal consistency.
provides details of the included studies.
A total of 20 contextual factors were derived from 71 text extracts as shown in . Individual, cultural and structural contextual factors were further classified as being relevant to programme implementation, and/or programme efficacy and effectiveness.
Individual contextual factors were defined as the specific characteristics of students involved in RSE programmes (Pawson Citation2006). Six factors were found. Four of them, derived from two papers (DEECD (Department of Education and Early Childhood Development) Citation2011; Kamara Citation2020), detailed common attributes of children in upper primary school: puberty experience, interest in the subject, exposure to sexual themes in media, and age. One factor, maturity level (Newby and Mathieu-Chartier Citation2018), was related to programme implementation. The remaining factor, gender, related to programme effectiveness, was typically reported using a binary approach (female/male; girl/boy) which does not fully account for gender diversity (Newell, Britt, and Graham Citation2011; Mason Citation2010; Piercy and Haynes Citation2006; Roberts Citation2015).
Cultural contextual factors were defined as ideas, values, and beliefs shared by a social group in relation to place and time (De Souza Citation2013). Three cultural contextual factors were identified in the papers as connected to RSE programme implementation. A fourth factor was found connected to both programme implementation and programme effectiveness.
The first factor referred to cultural shifts within broader society (Cahill et al. Citation2019; Newby and Mathieu-Chartier Citation2018; Piercy and Haynes Citation2006). For example, increasing concerns about social issues can create environments in which RSE programmes may be considered favourably.
The second factor referred to ideas about children and sexuality (Piercy and Haynes Citation2006). For example, programme implementation may be facilitated in contexts in which children are seen as in need of learning about relationships and sexuality.
The third factor referred to the shared school ethos (e.g., policies and practices) which can be relative to geographical location or faith (Kamara Citation2020; Piercy and Haynes Citation2006). For example, preferences for specific types of RSE programmes may be influenced by this ethos.
The fourth factor, connected to both programme implementation (Cahill et al. Citation2019; DEECD (Department of Education and Early Childhood Development) Citation2011; Roberts Citation2015) and effectiveness (DEECD (Department of Education and Early Childhood Development) Citation2011; Roberts Citation2015), referred to families’ values. For example, parental decisions as to when, what, how and to what extent children should learn about RSE may affect the capacity of RSE programme to influence programme outcomes.
Structural contextual factors were defined as elements within institutions in which society is organised (De Souza Citation2013). A greater number of extracts belonged to this category, suggesting that structural factors are seen as highly relevant to RSE programme implementation and effectiveness.
Structural contextual factors have three levels macro– meso– and micro– and each also has sub-factors. Three macro-level factors were found.
The first factor, connected only to implementation, referred to systemic power to secure RSE provision which may be a key factor affecting the implementation of universal RSE programmes (Cahill et al. Citation2019; DEECD (Department of Education and Early Childhood Development) Citation2011; Newby and Mathieu-Chartier Citation2018; Piercy and Haynes Citation2006; Roberts Citation2015).
The second factor, connected to implementation (Mason Citation2010; DEECD (Department of Education and Early Childhood Development) Citation2011) and effectiveness (Kamara Citation2020; Piercy and Haynes Citation2006), referred to how decisions were enacted to guide practices. Without policy directives and resources, schools may struggle to deliver effective RSE.
The third factor, connected to implementation (DEECD (Department of Education and Early Childhood Development) Citation2011; Kamara Citation2020) and effectiveness (Newby and Mathieu-Chartier Citation2018; Walsh and Peters Citation2011), referred to established processes and hierarchies in education systems such as division into different sub-systems (e.g., private, public). RSE programmes could be thwarted by factors such as a crowded curriculum, lack of funding, and lack of programme evaluation.
Two meso-level structural factors were connected to RSE programme implementation.
The second factor referred to collaborative work to secure support services and funding necessary for sustained RSE programme delivery (Cahill et al. Citation2019; Roberts Citation2015; Walsh and Peters Citation2011).
Five micro-level structural factors were found.
The first factor, connected to implementation (Cahill et al. Citation2019; Walsh and Peters Citation2011) and effectiveness (Cahill et al. Citation2019; Mason Citation2010; Roberts Citation2015; Piercy and Haynes Citation2006), referred to supporting teachers and guiding teaching practices. This factor was the most frequently mentioned, suggesting that teachers are central to delivery of effective RSE.
The second factor, connected to implementation (Kamara Citation2020; Roberts Citation2015; Walsh and Peters Citation2011) and effectiveness (Cahill et al. Citation2019; Roberts Citation2015; Walsh and Peters Citation2011), referred to the availability of resources and funding. Limited funding was a barrier to implementation and programmes benefit from appropriate resourcing.
The third factor, connected to implementation (Kamara Citation2020; Walsh and Peters Citation2011) and effectiveness (Kamara Citation2020), referred to established processes for including RSE in the school curriculum. RSE programmes may be more likely to effect change in contexts in which RSE is a regular part of school life.
The fourth factor, connected only to implementation, referred to schools taking an active role to secure RSE provision (Newby and Mathieu-Chartier Citation2018; Roberts Citation2015). The uptake of RSE programmes could be promoted in contexts in which schools and families share the responsibility for RSE.
The fifth factor, connected to effectiveness, referred to ideal personnel to deliver RSE. On the one hand teachers’ knowledge about students may be helpful for RSE (Mason Citation2010). On the other hand, some students may feel more comfortable with external educators (Pound et al. Citation2017).
A total of nine mechanisms were derived from 103 text extracts as shown in . Identifying mechanisms was most challenging because a degree of inference was required as papers did not report information under the heading ‘mechanisms’.
Two mechanisms were identified in a high number of extracts:
‘children involved in interaction’ (Cahill et al. Citation2019; DEECD (Department of Education and Early Childhood Development) Citation2011; Kamara Citation2020; Mason Citation2010; Newby and Mathieu-Chartier Citation2018; Roberts Citation2015). From the text extracts, we identified that children enjoyed talking with their peers about their feelings and growing up. Participatory approaches were highly recommended. Visual aids, web-based programmes, and other technologies (e.g., videos, Smartboard) were used to engage students.
‘adaptation for appropriateness and acceptability’ (DEECD (Department of Education and Early Childhood Development) Citation2011; Mason Citation2010; Newby and Mathieu-Chartier Citation2018; Newell, Britt, and Graham Citation2011; Piercy and Haynes Citation2006; Roberts Citation2015). We inferred from text extracts that children valued clear and comprehensible explanations and were interested in learning about their bodies and puberty. However, not every child liked to talk about sexual matters. Programme resources included formative assessment, a question box, school community consultation, and enabling children to use familiar terms.
Two mechanisms were present in five papers:
‘children feel safe and secure’ (DEECD (Department of Education and Early Childhood Development) Citation2011, Kamara Citation2020; Mason Citation2010; Newby and Mathieu-Chartier Citation2018, Newell, Britt, and Graham Citation2011), which implied a certain level of comfort discussing relationships and sexuality. A safe and secure environment provided ground rules, overturned ideas that RSE topics should not be discussed, and allowed anonymity and passive participation.
‘relevance to children’s lives’ (Cahill et al. Citation2019; DEECD (Department of Education and Early Childhood Development) Citation2011; Newby and Mathieu-Chartier Citation2018; Newell, Britt, and Graham Citation2011; Piercy and Haynes Citation2006). Children could assess whether RSE content was timely and relevant and discarded content they did not agree with. Relevant content was characterised as being useful, helpful, and having educational value. A spiral curriculum in which topics were revisited was the most recommended resource to enable children to expand on prior knowledge at different ages.
We identified further mechanisms related to:
‘legitimisation of content through multiple sources’ (DEECD (Department of Education and Early Childhood Development) Citation2011; Kamara Citation2020; Newell, Britt, and Graham Citation2011; Piercy and Haynes Citation2006). Children perceiving RSE as a valuable aspect of the adult world could legitimise RSE content, and this could be achieved by using scientifically-accurate language, parents/caregivers reinforcing key messages, and inviting external experts as educators.
‘modelling communication about relationships and sexuality’ (DEECD (Department of Education and Early Childhood Development) Citation2011; Kamara Citation2020; Newell, Britt, and Graham Citation2011; Piercy and Haynes Citation2006). Teaching RSE from an early age, addressing children’s questions comfortably and sensibly, and using a whole-school approach to RSE can help to overcome feelings of embarrassment and, in turn, generate positive responses to RSE programmes.
‘validation of children’s knowledge and capacities to guide their learning’ (Kamara Citation2020; Piercy and Haynes Citation2006; Roberts Citation2015). Children appreciated being recognised as knowledgeable subjects in relation to RSE. Two examples from papers as to how this was achieved were requesting feedback from children and including them in a school-based RSE committee.
Extracts were also used to derive the mechanisms:
‘children feel included and represented’ (DEECD (Department of Education and Early Childhood Development) Citation2011; Kamara Citation2020). RSE programmes can use language, images and curriculum content for children to see their lives represented in relation to different ‘masculinities and femininities, ethnicities, faiths, relationships, sexual orientation, and family composition’ (DEECD (Department of Education and Early Childhood Development) Citation2011, 21).
‘active, embodied, and fun’ (DEECD (Department of Education and Early Childhood Development) Citation2011; Roberts Citation2015). RSE programmes can use play and games to promote children’s learning while they focus on the fun aspect of an activity, as has been suggested regarding RSE for adolescents (Taylor Citation2020).
A total of 59 thematic extracts were gathered from 34 text extracts and clustered into 26 outcomes of RSE programmes. In conducting the review, we found that outcomes of respectful relationships education programmes were sharply distinct from RSE programme outcomes3. Hence, we did not include the outcomes identified in papers on respectful relationships education (Cahill et al. Citation2019; Walsh and Peters Citation2011). shows RSE programme outcomes further classified in individual, cultural and structural terms. Outcomes at the individual level are presented at five hierarchical levels from Level 1 (immediate knowledge outcomes) to Level 5 (future-oriented conceptual outcomes). Individual outcomes were organised as outcome chains (Funnell and Rogers Citation2011), on the assumption that each level leads to the next highest level increasing gradually in complexity.
Individual level outcomes
We found 22 outcomes, which suggests that programme evaluations to date have focused significantly more on individual change than changes at other levels.
Level 1. Outcomes involved recalling information or developing knowledge and were derived from 30 thematic extracts identified in three papers (DEECD (Department of Education and Early Childhood Development) Citation2011; Newell, Britt, and Graham Citation2011; Roberts Citation2015). Almost half of these extracts focused on understanding about body parts and functions (n = 7) and puberty changes (n = 9). Understanding human reproduction and personal safety were derived from four thematic extracts respectively. Understanding about feelings and emotions was found once, as well as understanding about gender stereotypes (e.g., facts and fictions of being boy/girl).
Level 2. Outcomes referred to skills and action verbs (e.g., to seek help). Four outcomes were derived from 11 thematic extracts identified in the same source (DEECD (Department of Education and Early Childhood Development) Citation2011).
Level 3. Outcomes referred to the impact of RSE programmes and may only be corroborated in the long term. Four outcomes were derived from five thematic extracts (Kamara Citation2020; Newby and Mathieu-Chartier Citation2018; Newell, Britt, and Graham Citation2011; Roberts Citation2015). Positive impact on skills, attitudes, values and beliefs about relationships and sexual health was identified in two papers, while the remainder were derived from one paper each. The notion of gender was not found in Level 3.
Level 4. Outcomes referred to broad aspects of the whole individual and should be understood as long-term goals. Four outcomes were derived from five thematic extracts identified in two papers (DEECD (Department of Education and Early Childhood Development) Citation2011; Roberts Citation2015).
Level 5. Outcomes referred to the future of individuals in relation to self and as part of society. Three outcomes were derived from one extract each (DEECD (Department of Education and Early Childhood Development) Citation2011; Kamara Citation2020; Piercy and Haynes Citation2006).
Cultural level outcomes
Three outcomes were derived from five extracts identified in four papers (DEECD (Department of Education and Early Childhood Development) Citation2011; Kamara Citation2020; Roberts Citation2015; Newby and Mathieu-Chartier Citation2018). RSE programmes were associated with fostering change in cultural contexts (e.g., culture of prevention, parental endorsement of RSE programmes).
Structural level outcomes
One outcome was found (Roberts Citation2015). RSE programmes could change teachers’ practices and promote the uptake of training courses, which is a structural contextual factor.
We conducted a literature review and identified nine relevant studies of relationships and sexuality education (RSE) programmes for primary school students. Adopting a theory-generating approach informed by realist evaluation enabled us to identify contexts, mechanisms and outcomes that could be formulated into programme theories to explain the functioning of RSE programmes. Our findings will interest a range of audiences wishing to use to utilise RSE programme theories in research and practice, including evaluators, researchers, RSE programme designers and practitioners.
We identified several contextual factors that influence RSE programmes and that could be incorporated into programme theories. At a structural level, findings suggested that decisions made at the highest level in educational bureaucracies and government policies are crucial to RSE programme implementation in many ways (e.g., by promoting specific RSE approaches, allocating funding, and creating teaching materials). In school, supporting teachers and guiding teaching practices were identified as the most relevant structural factors, consistent with the existing literature (O’Brien, Hendriks, and Burns Citation2021). At a cultural level, decisions regarding when, how and how much to discuss with children about relationships and sexuality were identified as values associated with perceived programme effectiveness, consistent with previous reviews (Mullis et al. Citation2021). At an individual level, we noted that only gender tended to be used as variable in data analysis (as evident in reporting of study data) despite the myriad of other individual factors that mightinfluence RSE programme outcomes (e.g., puberty experience, children’s interest in RSE, exposure to sexual themes in media). Additionally, all included studies reporting on gender used a binary approach thereby failing to capture diverse gender identities (e.g., non-binary gender identities) which may be out of step with international guidelines on sexuality education (UNESCO Citation2018) and human rights conventions which promote non-discrimination on the basis of gender identity (UN Citation1989). This is especially important because the time approaching puberty is recognised as challenging for children who may be questioning their gender identity (UNESCO Citation2018). Overall, programme evaluations could do better by incorporating data collection modules to account for a wider range of cultural, structural and individual factors that may influence programme mechanisms and thereby affect programme outcomes.
The approach to mechanisms that we adopted revealed the need to seek children’s opinions about how different mechanisms influence the effectiveness of RSE (i.e., the reasons why children think RSE programmes may work). One example of this type of research is provided by Robinson and Davies (Citation2014), who used image-guided interviews and focus groups to elicit children’s perspectives. Despite sparse data on children’s opinions in the included studies, we were able to derive nine mechanisms. Our findings add to recent research on mechanisms within RSE programmes (e.g., Sell, Oliver, and Meiksin Citation2021; Taylor Citation2020). These mechanisms should be tested in further theory-driven evaluations to inform programme logic models. A challenge for future evaluators is to find ways to elicit children’s thoughts on the reasons why RSE programmes can work. Congruent with the United Nations Convention on the Rights of the Child (UN Citation1989), children should be facilitated to give their opinions on issues that affect them, and adults must take these opinions seriously.
We found in papers that RSE programme evaluations focused largely on individual and short-term outcomes. This has also been reported in previous reviews (e.g., Levy et al. Citation2020; Rocha, Silva, and Duarte Citation2022). Outcomes were frequently limited to improving students’ understandings of the biological aspects of sexuality (e.g., puberty, body parts, and human reproduction). Investigating children’s understandings of gender and gender equality appears neglected in evaluations of RSE programmes, even though these outcomes have been recognised as key to sexual health and wellbeing outcomes (Haberland and Rogow Citation2015; UNESCO Citation2018). While several reasons may explain why evaluations have focused on a narrow range of outcomes, this tendency may lead to missed opportunities to assess broader structural and cultural RSE outcomes. From the perspective of a health promoting school, for example, the whole school can be seen as an environment to reinforce individual outcomes through structural and cultural change (Ollis and Harrison Citation2016; WHO/UNESCO Citation2021). Therefore, programme evaluations should intentionally gather evidence on the influence of RSE on both structure and culture (e.g., capturing data on attitudes in school communities towards preparing children for puberty at a younger age and prior to the onset of puberty). Longitudinal designs are required to understand the long-term effects of RSE programmes on individuals, structures and cultures.
Strengths and weaknesses
Despite the relatively small sample of papers focused on here, there was ample opportunity to identify contexts, mechanisms and outcomes. Although study quality was not assessed, all papers were associated with best practice principles and successful programme outcomes. We cannot identify causal relationships between the identified contexts, mechanisms and outcomes. However, we have proposed a pool of potential contexts, mechanisms and outcomes upon which evaluators and researchers can draw.
We did not place language or location restrictions on our review search criteria. However, we included only studies published in English reporting on studies conducted in English-speaking countries. This must be acknowledged as a review limitation.
Review studies are by nature backward-looking, so our findings may not reflect the current reality and status of RSE programmes. Preventive measures for COVID-19 have altered students’ lives and health (Panchal et al. Citation2021) changing how students may choose to live their relationships and sexualities. Emergent educational contexts (e.g., emergency forms of remote learning) have also changed how educators and students engage with RSE programmes (Bowling et al. Citation2022; Rolleri, Bass, and Taverner Citation2021). A redefinition of RSE programme outcomes and mechanisms may be necessary in the light of this. Nevertheless, this review provides a platform on which to advance programme evaluations.
Our selection criteria did not permit us to include topic-specific programmes overlapping with RSE. We excluded a body image programme involving puberty education (Cousineau et al. Citation2010) and a socio-emotional learning programme including sexuality content (Schonfeld et al. Citation2012). In doing so, we excluded the validated outcome measures used in those studies which may have expanded the pool of outcomes identifiable.
Published studies tend to focus on evaluation of programmes for which teaching manuals exist, which are not the only source of RSE for upper primary school-aged children. For example, classroom teachers may deliver RSE to students by bringing together materials from a variety of sources and being innovative in their teaching. This review could not capture these aspects of RSE delivery.
Finally, review results were influenced by the choice of realist evaluation as a principle underpinning the style of review. As the realist evaluator, Westhorp (Citation2018) explained, categorising contexts, mechanisms and outcomes depends on a programme’s aims and characteristics. Hence, factors and variables that we interpreted as contexts, may be considered mechanisms or outcomes using a different style of programme evaluation.