Effective implementation of the Sport Education Model in physical education: A meta-analysis of participant and intervention characteristics

PLOS One | https://doi.org/10.1371/journal.pone.0331228 October 16, 2025 4 / 17 Exclusion criteria were as follows: (a) irrelevant populations, such as participants who are not currently enrolled stu- dents or those not involved in physical education courses; (b) not SEM intervention or if the hybrid pedagogical model was not explicitly based on the structural framework and elements of SEM, the study would be excluded; (c) studies with designs other than randomized controlled trials or quasi-experimental studies, and control groups that did not use stan- dard or traditional teaching methods were excluded, such as SEM compared with other teaching modes, or mixed teach- ing compared with SEM; In addition, non-equivalent control studies will be excluded. (d) books, conference proceedings, theses, reviews, meta-analyses, retracted studies, erratum, and studies without available abstracts were also excluded (e) studies lacking sufficient data for analysis. Literature screening Two researchers independently screened the literature and extracted data. Due to the large number of initially selected articles, only titles and abstracts were reviewed after duplicates were removed. Articles that did not meet the PICOS inclu- sion criteria were excluded. Subsequently, full-text readings were conducted for studies that potentially met the criteria, and studies that did not align with the research objectives and design were excluded. Finally, the two researchers (JLZ, GGY) cross-checked the screening results; in case of disagreements, a third researcher (WSX) was consulted.

Risk of bias in individual studies and certainty of evidence The RoB-2, the latest version of the Cochrane Risk of Bias tool, was used to evaluate the risk of bias for each randomized controlled trial included [24]. For quasi-experimental studies, the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool was used to assess the risk of bias [25]. The certainty of the evidence was analyzed and summarized according to the guidelines outlined in the GRADE Manual [26]. Two research team members (JLZ, GGY) independently evaluated the risk of bias for each included study. Eigenvalue coding and data extraction The characteristics of the included studies were systematically coded for subsequent analyses. In alignment with the focus of this research, the independent variable was coded as the Sport Education Model (SEM), while the depen- dent variable was coded as physical education learning efficiency. Modern developmental psychology indicates that the cognitive development of children and adolescents should be divided into different stages based on their growth processes. Students must progress through various cognitive development phases, ranging from concrete to abstract thinking and from perception and induction to reasoning [27]. Considering the patterns of student growth and devel- opment, an asynchronous relationship exists between their cognitive and non-cognitive abilities [28]. Therefore, the dependent variable was classified into cognitive factors (codified as A): academic performance, motor skills, physi-

cal fitness, theoretical knowledge, and practical skills; and non-cognitive factors (codified as B): learning motivation, learning interest, learning attitudes, and mental health [28]. Participant and intervention characteristics were also coded as moderator variables. Participant characteristics: number of students, learning stage, experience with SEM, and class size. Intervention characteristics were: the frequency of the intervention, the duration of the session, and the period of the intervention. The respective rules were: (1) school stage that was coded as the school of elemen- tary, secondary, and university; (2) previous SEM experience coded as “yes”, “no”, and “not mentioned”; (3) class size coded as the small class (0–30 students), medium class (31–50 students), and large class (>50 students); (4) fre- quency of the intervention coded as ≤1 lesson/week, 2 lessons/week, and >2 lessons/week; (5) session length coded as greater than 60 min/session, 30–60 min/session, and less than or equal to 30 min/session; (6) the intervention period coded by one of the following: < 18 sessions and ≥18 sessions. In addition, research sources, the number of students, and intervention program were extracted.