Physical Education (PE) is recognized by its crucial role in students’ acquisition of values and competencies that contribute to their motor, cognitive, emotional, personal and social development and facilitate the inclusion in current society, preparing them for the future (Giraldéz, 2006; Mesquita, 2012; Rosado and Mesquita, 2011; Taggart, 1988).

The explicit and formal character of the instructional process, evident in the traditional pedagogical approaches (called teacher-centered approaches) have tended to dominate PE instruction for much of the 20^th century (Gubacs-Collins, 2015; Gubacs-Collins and Olsen, 2010; Lee, 1993). The teaching/learning process is one in which the role of the teacher is primarily as instructional leader and students are expected to demonstrate more compliance than initiative (Metzler, 1989; Rosado and Mesquita, 2009; Rosenshine, 1979). This direct style has the potential to compromise students’ capabilities of building their own learning, and reduce the role of their cognitive and social processes, decision-making and autonomy (Bruning et al., 2004; Ennis, 2014; Metzler, 2011; Siedentop et al., 2011). As a result, a number of student-centered approaches (SCA) have been developed which are based on the constructive and social theories of learning (Hastie and Siedentop, 1999; Mesquita et al., 2012; Putnam et al., 1990). SCA are designed to stimulate the ability of students to make decisions, to reflect, and to solve problems (Dyson et al., 2004; Hastie and Mesquita, 2016). This then allows students to have a more proactive than reactive role in their own learning and changes the role of the teacher to one more of a facilitator (Ennis, 2014; Hattie, 2012; Jones, 2007; Mesquita, 2013).

In PE, one of the most widely implemented and researched “second generation of models that build on strong statements of democratic, student-centered practice” (Ennis, 2014, p.63) is SE (Siedentop et al., 2011). SE is a pedagogical model that incorporates the tenets of socio-constructivist learning theories recognized by its valences for "learning focused, provide measurable student outcomes, and assist students to become engaged in positive, learning-oriented sport environments" (Ennis, 2014, p.67). It is characterized by prioritizing more implicit and informal teaching strategies (prevails questioning) allowing students to make decisions during the learning process, and encouraging them to learn autonomously and responsibly (Mesquita et al., 2012). The features that underpin SE (seasons, affiliation, formal competition, culminating events, record keeping and festivity) aim to fulfil Siedentop’s (2002) goal of educating students to be “athletes in the fullest sense and to help them develop as competent, literate and enthusiastic sportspersons”. To do this, students are given opportunities to engage in a variety of roles, beyond that of simply as player. These can include coaches, referees, score keepers, statisticians, members of the sports organizing board or sports director.

According to the Cochrane Collaboration (Higgins et al., 2019), systematic reviews are important because they provide a high-level summary of primary research on a specific research question that attempts to identify, select, synthesize, and appraise all high-quality evidence relevant to that question to answer it. Further, systematic reviews collate all evidence pertinent to a priori selected criteria for eligibility to address the specific research question. Without them, researchers lack an understanding of the subject, of what has already been examined, how it has been researched, and what key concerns have been identified.

Previous systematic reviews of SE research (Araújo et al., 2014; Evangelio et al., 2018; Hastie et al., 2011; Hastie and Wallhead, 2016; Wallhead and O'Sullivan, 2005) have been structured with different focus. For example, Wallhead and O’Sullivan (2005) and Hastie, Martínez de Ojeda and Calderón (2011) focused on the achievement of the “big 5” aims of PE, namely students’ attitudes and values, personal and social skills, fitness, as well as motor skills and tactical knowledge. Later, Araújo, Mesquita and Hastie (2014) reviewed studies where there was a specific examination of students’ learning outcomes, while Hastie and Wallhead (2016) focused on the extent to which the competent, literate and enthusiastic goals of SE were achievable. Most recently, Evangelio et al. (2018) organized their review around SE’s impact on cognitive, social, affective and physical outcomes.

During recent years, PE programs have been challenged by the needs of children and youth in a changing environment (Chin and Edginton, 2014; Ennis, 2014; O’Sullivan, 2013) and the development of personal and social skills have become particularly valued. Hence, it is necessary to perform this review including the most prominent research, indicating which are the most studied and valued personal and social variables within the SE model, in order to answer specific research questions and indicate directions that future research and practice might follow.

The purpose of this study was to systematically review and synthesize the SE studies that have examined the development of students’ personal and social skills. The five research questions which guided the review of these studies were:

The systematic review followed the PRISMA protocol for reporting systematic reviews (Moher et al., 2009), and was conducted through electronic searches on eight databases. These include Web of Science, SCOPUS, Academic Search Complete, ERIC, SPORTDiscus with Full Text, Education Source, PsycINFO and PsycARTICLES. The search included all papers published up until March 2018, using the Boolean operators (AND, OR) to concatenate the search terms “Sport Education”, “physical education”. A secondary search was performed by screening the reference lists of the included studies and relevant review articles. The study selection was carried out independently by two authors to minimize potential selection bias. Both these authors have experience in this methodology and are knowledgeable of instructional models in PE, and any discrepancies were resolved by consensus.

Studies for this review were included according to the following criteria: (i) were published in peer-reviewed international journals indexed in JCR or SJR; (ii) were available in full-text; (iii) examined personal and social variables as either included or measured as main outcomes within the SE model.

Studies were excluded if they: (i) were review or opinion articles; (ii) were articles without full-text availability; or (iii) addressed issues related with SE other than the development of personal and social skills. Duplicate documents, opinion articles, books, book chapters, review articles, conference papers or theses were also excluded from this review.

Articles’ titles and abstracts generated from this process were read and included or excluded on the basis of the above criteria.

Figure 1 presents the study selection processes. The initial search, from the wide range of articles that identified “Sport Education” AND “physical education” in either the title, abstract or keywords (n = 1644), only peer-reviewed articles with impact factor related to SE research were selected for reading (n = 99). From this number, only those related to social and personal development were selected (n = 59). Review articles (n = 3) and articles without full text (n = 5) were excluded for this review. After a review of titles and abstracts, 49 articles were selected, and their full text was analyzed. Therefore, only peer review journal articles, published in journals with impact factor, that specifically studied the development of personal and social skills with SE were included to the present review (n = 51).

In order to extract all relevant data from the 51 studies included in this review, content analysis was performed. Studies are presented in Table 1, in alphabetical order according to the first last name of the principal author. The review categories used were defined a priori (Harris et al., 2014) seeking to answer the research questions. The categories are listed below (Table 1) with the legends used for each one in Table 2 appearing in brackets.

To assess the methodological quality of the included 51 studies we used the validated Downs and Black checklist (1998). Items that were not applied to the design of the analyzed studies were removed from the 27-item checklist. The modified version consisted of items 1-4, 6, 10-13, and 18-24, with the highest possible score of 16 (Table 2). Two main evaluators independently performed the assessment of the selected studies. Both researchers discussed and agreed upon the reconciliation of observed differences. In the present systematic review, no study was excluded due to a significantly low quality assessment score.

Table 3 provides an overview of each of the 51 studies included in this review.

The results of this study are presented in response to the research questions mentioned above.

Countries: According to the country where the research on the development of personal and social skills using the SE took place, the United States (29%) and Spain (29%) represented over half of all publications (58%), followed by Australia (12%) and the United Kingdom (12%), Russia (6%), Portugal (4%), Finland (2%), Ireland (2%), New Zealand (2%) and Singapore (2%).

Context: Regarding the studies’ context, the majority of research (94%) took place in school contexts and within PE lessons. Only three studies (6%) were conducted within sport club settings. Studies with involving data from only students were more frequent (67%). Ten studies (20%) used student and teacher data, whereas only 4 studies (8%) focused exclusively on teachers. Two studies (4%) were focused on athletes as participants while one study (2%) involved at same time coaches, athletes and parents.

Classification of sport type: With respect to the type of sport/activity studied, the predominance of team sports in 37 studies (72%) was noticeable, whereas only 7 studies (14%) incorporated individual sports (such as kickboxing, badminton, swimming) in their seasons. The remaining 7 studies (14%) were developed with multiactivities (individual and team sports).

Participants: The total sample of the included studies was 2949 students (83%) (1301 boys and 1118 girls, considering that in eight studies the gender of 353 participants was not specified), 496 teachers (15%) (107 with experience in teaching SE), 68 athletes (2%) (22 boys and 23 girls, considering that in two studies the participants’ gender was not specified), 8 coaches (0.2%) (without previous experience in SE) and 4 parents (0.1%).

School level: The most frequent school level studied (40%) was high school (considered as grades 9-12), followed by the middle school (31%), which is from 6^th until 8^th grade. Fewer studies (18%) took place in elementary schools (1^st to 5^th grade), and of those, the most common grade level were 4^th and 5^th.

Class composition: 36 studies (70%) were in a co-educational PE context, two examined only girls (4%), and 6 examined only boys (12%) in a single-sex PE context. Class composition was not reported in 7 (14%) of the studies.

Students with special needs or “at risk”: In 49 studies (96%) participants were not students with special needs or “at-risk” of failure due to undisciplinary behaviors or in process of early dropout. Of interest, only 2 studies (4%) included participants who had a disability (intellectual or motor) or were considered as “at risk” of failure or early dropout.

Figure 2 provides an illustration of the range of variables examined in studies in this review. As can be seen in the figure enjoyment/satisfaction, enthusiasm and engagement are the predominant outcome measures.

Study approach: Almost half of all studies adopted a qualitative approach (23; 44%), of the remainder, 14 adopted a quantitative approach (27%), while 14 assumed a mixed approach combining both quantitative and qualitative approaches (27%).

Study design: More than half of the studies (26) included in this review utilized a non-experimental design (51%) to describe the development of students’ personal and social skills within SE (without control, manipulation or alteration of the variables, focused on teachers and students’ perceptions).

However, is that of the 24 studies that involved a quasi-experimental design (49%) (being able to manipulate and control the variables with pre-posttest designs), only two included a control group. Further, there was no randomization of participants within these studies.

Instruments: The use of questionnaires was reported in all quantitative studies (27) and focused on students’ perceptions, and two collected data with direct observation. The analyzed qualitative studies have focused on students and teachers’ perceptions, and adopted semi-structured interviews and drawings to collect data. Data collection with mixed methods studies was mostly conducted using individual or group interviews and questionnaires, as well as through diaries, interviews and questionnaires.

The most widely used instruments to collect data in studies with a qualitative multimethod strategy were interviews (12/18), followed by diaries and/or field notes (8/18), focus group interviews (6/18), participant observation (5/18), journals and/or reflective logs (3/18), videotape (3/18), drawings (2/18), photovoice (1/18) and document analysis (1/18). One study adopted a quantitative multimethod strategy (with multiple forms of quantitative data) using the questionnaires and the field work.

Length of the season: In 18 studies, the SE season extended for more than 20 lessons (35%), 15 studies for between 15 and 19 lessons (29%), 13 between 10 and 14 (6%) and 4 with less than 9 lessons (4%). Each lesson lasted 45-60 minutes.

No information was given about the length of the SE season implementation in 2 studies (4%).

In 19 studies (37%) the fidelity of the SE implementation was confirmed, that is, the authors performed the validation of the model implementation and presented a detailed description of the program and curricular elements of the unit (Hastie and Casey, 2014). In 25 studies (49%), the authors presented only a description of the program or curricular elements of the unit, failing to carry out the validation of the model, that is, did not assure that the instruction was indeed consistent within the accepted standards for the SE.

The aim of this systematic review was to describe and examine what is currently known concerning students’ development of personal and social competencies when participating in PE classes with SE, in order to give directions for future research and practice.

The SE research included in this systematic review was published between 1992 and 2018 with an increasing number of publications over the years. The fact that the majority of studies took place in USA and Spain (58%) is in line with previous reviews. Although studies have been developed in other countries, namely in Australia and elsewhere in Europe (such as United Kingdom, Russia, Portugal), it would be important to expand the impact of the SE on those and other contexts and cultures.

The findings revealed that studies regarding the development of personal and social competencies focused mostly on students’ perceptions (83%), and were located in school contexts (94%), and involved co-educational classes (70%). However, in the main, there was minimal research that involved students with disabilities, or those considered “at-risk” of failure or who were in the process of early dropout (96%).

One of the long-term purposes of the SE is to make sport more widely accessible so that race, disability, or socioeconomic status are not barriers to participation, thereby promoting inclusion and equity (Siedentop, 1994). By consequence, samples that consider participants with those particular characteristics must be taken into account in SE season implementations. Moreover, given that it is recognized that sport is a privileged space for the development of personal and social skills (Fraser-Thomas et al., 2005; Wright and Côté, 2003) the implementation of SE should not be limited only to the area of PE. It should focus on addressing the level of transfer to other contexts, namely sport training settings, seeking to reinforce the positive impact of SE on the personal and social skills of athletes found in studies of Meroño et al. (2015; 2016).

Results showed that the development of personal and social competencies was more studied in high-school (40%) and 9^th grade (30%), which is in contrast to other reviews which identified middle (Araújo et al., 2014; Hastie et al., 2011) and elementary schools (Evangelio et al., 2018) as the most frequent levels studied. These findings are supported by studies (Ntoumanis and Standage, 2009; Van den Berghe et al., 2014) which report that older students, when participating in a SE program, are likely to develop personal and social skills. Furthermore, Layne and Hastie (2016) reported that in very young students the implementation of an SE season requires greater teacher preparation, presenting itself as a potential limitation when studying early grade levels.

Consistent with the previous reviews of (Araújo et al., 2014; Chu and Zhang, 2018; Hastie et al., 2011), the majority of studies in this review investigated team sports in SE programs (70%). This has the potential of undervaluing the possibility of different results in individual sports. As such, it is imperative to conduct further studies in individual and performance sports given that research has shown that students engaging in SE seasons of involving these activities are more likely to participate in these sports more regularly, and that participation may extend to adulthood (Tammelin et al., 2003).

From this systematic review, the personal and social variables that appeared more often were the same as those considered to be crucial for learning in PE irrespective of the teaching approaches. These included enjoyment and satisfaction (n=26), enthusiasm and engagement (n=25) and motivation (n=15) as the most prevalent. Findings suggested high levels of enjoyment and satisfaction in SE (e.g. Alexander and Luckman, 2001; Curtner-Smith, 2004; Menéndez and Fernandez-Rio, 2017; Meroño et al., 2015; Sinelnikov and Hastie, 2008), increases in enthusiasm and engaged participation (e.g. García-López et al., 2012; Grant, 1992; Hastie, 1996; Meroño et al., 2016; Mesquita et al., 2016; Smither and Xihe, 2011), and enhanced feelings of motivation (e.g. Burgueño et al., 2017; Hastie et al., 2014; MacPhail et al., 2008; Pill, 2010; Sinelnikov and Hastie, 2010). These results could be associated to the structural features of SE such as longer seasons, consistent team membership, and a significant amount of time allocated to game play, as well as features such as competition, festivity, and the presence of a culminant event. Further, the diversity of students’ roles within the team (playing and non-playing roles) as well their opportunity to make decisions may have a strong influence on enthusiasm and engagement. With regard to the more specific variables strongly foregrounded with SE, due to its own structure and pedagogical principles, the personal and social variables mostly studied were responsibility (n=18), affiliation and ownership (n=16), inclusion, peer support and equity (n=16), teamwork, cooperation and compliance (n=14), autonomy (n=13), empathy and friendship (n=12). Fair-play (n=8), empowerment, problem-solving and decision-making (n=7), leadership (n=4), trust and confidence (n=4), self-determination (n=3), assertiveness (n=2) were also present.

The theme of responsibility was found in a number of studies (e.g. Browne et al., 2004; Brunton, 2003; Fernandez-Rio and Menéndez-Santurio, 2017; Hastie, 1996; Hastie and Buchanan, 2000; Sinelnikov and Hastie, 2008). Students’ ability within SE to take on roles (e.g. referee, coach, and statistician) and the opportunities to solve problems were identified as key points to the perceptions of greater levels of responsibility developed by students. Nevertheless, in the study of Mowling et al. (2006) involving fourth grade students, a minimal representation of roles and responsibilities were noted. The author attributed this finding to the early age of the students who sought victory as their primary agenda. However, it should be noted that this study did find that students placed a strong focus on affiliation and festivity.

Studies gathered to this review highlighted the value of team affiliation in developing students’ ownership within a SE experience (e.g. Curtner-Smith and Sofo, 2004; Farias et al., 2018; Gutierrez Diaz del Campo et al., 2014; Hastie, 1998; Hastie and Sinelnikov, 2006; MacPhail et al., 2008). In a similar vein, a sense of cooperation, teamwork and compliance within the teams were all reported as important outcomes of participation in a season of SE. This feature was seen as crucial to ensuring the success and maintenance of team affiliation during the season (e.g. Alexander et al., 1996; Brunton, 2003; Fernandez-Rio and Menéndez-Santurio, 2017; Mesquita et al., 2016).

The focus of SE in promoting inclusion, equity and peer support was also suggested in a number of studies (Alexander and Luckman, 2001; Browne et al., 2004; Curtner-Smith and Sofo, 2004; Gutierrez Diaz del Campo et al., 2014; Hastie, 1998; Menéndez and Fernandez-Rio, 2017; O'Donovan, 2003; Pill, 2010). Here, the importance given to all team members (regardless of a student’s gender or skill level), the opportunities provided for inclusion participation and the emphasis on “doing your best” were highlighted. Nevertheless, in the Alexander et al. (1996) study, the analysis of female students’ journals indicated that they did not perceive such equitable participation as players in coeducational SE learning environments. However, in subsequent studies of gender inequity and marginalization (e.g. Alexander and Luckman, 2001; Hastie, 1998; Hastie and Sinelnikov, 2006) girls did not consider these inequities as problematic as they continued to feel a useful part of their teams, and continued to prefer SE over traditional models.

This review also showed that an enhanced level of autonomy was perceived by students, teachers and athletes as a result of their participation in SE (e.g. MacPhail et al., 2008; Méndez-Gimenez et al., 2015; Meroño et al., 2015; Romar et al., 2016; Smither and Xihe, 2011). Autonomy was seen as deriving from allowing students to select their teams, choosing the roles they wished to take within their team, as well as establishing and managing their own practices and games. The study of Cuevas(2015), however, provided evidence on only minor (but not significant) improvements in students’ autonomy after experiencing a SE season. The main argument for this finding was that students with higher social status tended to restrain more introverted students’ behaviors, thereby limiting their perceptions of autonomy.

Some studies also reported the development of perceived empowerment by students mostly due to the opportunities of SE to solve problems, make decisions and take control over their learning environment (e.g. Gil-Arias et al., 2017; Hastie and Buchanan, 2000; Romar et al., 2016; Sinelnikov and Hastie, 2008). Furthermore, students and teachers recognized that SE provides an excellent training for leadership capacity given the students’ ability to take on roles within the various activities in a season (e.g. Alexander et al., 1996; Clarke and Quill, 2003; Hastie and Sharpe, 1999). Similarly, there is empirical evidence with respect to the impact of pedagogical strategies used in SE seasons (e.g., particular roles students such as coach or reporter) on the development of students’ trust, resilience and self-confidence (e.g. Ang and Penney, 2013; Carlson and Hastie, 1997; MacPhail et al., 2004). The increases of students’ self-determination were also reported in some studies due to the features of SE (e.g. team affiliation and an affective game play rubric) (Cuevas et al., 2016; Perlman, 2011; Perlman and Goc Karp, 2010).

The findings also reported enhanced feelings of empathy and friendship among students in their experience with SE, producing positive changes in classmates’ perceptions (Fernandez-Rio and Menéndez-Santurio, 2017; Hastie and Sinelnikov, 2006; Menéndez and Fernandez-Rio, 2017; Wallhead and Ntoumanis, 2004). Nonetheless, the different interests and motivations among students that occur throughout the SE season can lead some students to adopt more egocentric positions, and not to put themselves in the place of the other. This was highlighted in the studies of e García-López et al. (2015; 2012) where they reported that empathy has decreased maybe due to large number of situations that occur within a SE season in which there is a clash of interests between students.

Regarding fair-play, studies with SE mentioned improvements including respect for oneself, others, adults and rules (e.g. Calderón et al., 2016; Clarke and Quill, 2003; Méndez-Gimenez et al., 2015), and decreases in the number of negative sporting behaviors (Perlman and Goc Karp, 2010; Vidoni and Ward, 2009).

The development of students’ assertiveness was only examined in two studies García-López et al. (2015; 2012). In the earlier study (García-López et al., 2012), students’ assertiveness did not increase. It was suggested that for this to occur, specific strategies related with assertiveness need to be deliberately implemented within the SE season design. Following this recommendation, in the García-Lopez (2015) study, findings suggested that SE proved to be a useful instructional model for improving students’ assertiveness.

Concerning the designs of the reviewed studies, almost half used a qualitative approach (47%) and a non-experimental design (53%) using multiple qualitative tools (35%). These tools included interviews, diaries and/or field notes, focus group interviews and participant observation. These findings are consistent with the reviews of Hastie et al. (2011) and Pozo et al. (2016). However, the most recent reviews of SE (Chu and Zhang, 2018; Evangelio et al., 2018) have indicated that significantly more studies in SE are following a quantitative (Chu and Zhang, 2018) or mixed method research approach (Evangelio et al., 2018). This divergence can be explained with the fact that Chu and Zhang’s (2018) review focused specifically on motivation. Nevertheless, due to the preponderance of qualitative studies in SE focusing on the development of personal and social competencies, new studies might begin to consider including mixed and quantitative methods, as these might provide objective and controlled measures and allow for their findings to be more widely generalized.

In quasi-experimental designs (47%), previously created class groups have always been used because it is very difficult to randomly distribute students in a school setting. However, when multiple classes are used it is important that the appropriate unit of analysis is used. Research has shown that usually the articles disregarded the unit of analysis and most of the articles applied the interventions to classes/groups, but used individual students as unit of analysis (Li et al., 2017).

Although the recommended length for a SE season is a minimum of 20 lessons (Siedentop, 1994) most studies (61%) did not comply with this principle. According to Siedentop (1994), seasons need to be long enough to allow for meaningful experiences, particularly since SE has more to accomplish. Specifically, when sport is taught more completely and authentically, it takes more time for students to develop the different roles and capabilities promoted by the model. Therefore, considering the main assumptions of the model, and the findings that development of social skills needs time (Ang and Penney, 2013; Farias et al., 2017; Hastie and Mesquita, 2016), in order to succeed and ensure more reliable results, future research must prioritize appropriate planning and design of the SE seasons themselves before any investigation of dependent measures is considered.

Fidelity of the implementation refers to the degree to which an intervention is delivered as intended and it is critical to successful translation of evidence-based interventions into practice (Carroll et al., 2007). Hastie and Casey (2014) consider that for an accurate and complete understanding of a study’s results, the methods section should include a rich description of the curricular elements of the unit, a detailed validation of model implementation, and a detailed description of the program context. Even though the research on SE highlights the importance of reporting the fidelity of the model implementation (Hastie and Casey, 2014; Ko et al., 2006), only 37% of the studies were in compliance with this aspect of design. This lack of model fidelity is consistent by with those of O’Donnell and Carol (2008) who state that fidelity of a model implementation is rarely reported in educational studies. The evaluation of the model implementation fidelity is essential because (a) it allows readers to moderate the relationship between an intervention and its outcomes, and (b) its assessment may also prevent potentially false conclusions.

Research concerning the impact of SE on students’ personal and social development has shown unequivocal results. In particular, the most examined personal and social variables within SE tend to be related with more general variables, which are crucial for learning in PE in all teaching approaches (p.e. enjoyment, satisfaction, enthusiasm and engagement). However, the interest of knowing the effect on variables strongly foregrounded with SE, due to its own structure and pedagogical principles (p.e. affiliation, ownership, peer support and fair-play) has been growing and becoming more specific (p.e. assertiveness, self-determination, compliance). In order to reinforce the positive impact of SE on the personal and social competencies it would be important that research consider other cultures, samples (e.g. coaches, athletes, disabled students), contexts (sport club setting) and types of sports (e.g. individual sports). A more equitable balance of research designs (mixed and quantitative methods), longer units with an effective planning of the SE season itself, as well as report of model fidelity is critical in future studies, as they might provide more robust and objective findings that can possibly be generalized.