# A network meta-analysis of pedagogical models in physical education: evaluating multidimensional learning outcomes and instructional duration effects

## TL;DR

This study compares different teaching methods in physical education to see which ones best improve student outcomes like skill and motivation, and finds that the length of instruction matters.

## Contribution

The study introduces a Bayesian network meta-analysis to compare pedagogical models in physical education, highlighting instructional duration as a key moderator.

## Key findings

- Hybrid Pedagogical Models (HPMs) are most effective for tactical decision-making and skill execution.
- Teaching Games for Understanding (TGfU) is most effective for student motivation.
- Instructional duration of around 720 minutes improves consistency of pedagogical outcomes.

## Abstract

This study systematically evaluated the comparative effectiveness of pedagogical models (PMs) on multidimensional learning outcomes in physical education and examined the moderating role of instructional duration.

A systematic search was conducted across seven databases (e.g., Web of Science, PubMed) for studies published between 2000 and 2025. Methodological quality was assessed using MINORS. A Bayesian Network Meta-Analysis (NMA) was performed using standardized mean differences (SMDs) to enable comparison across diverse outcome measures.

Sixty-five studies met the inclusion criteria. Probabilistic rankings suggested that Hybrid Pedagogical Models (HPMs) showed comparatively high effectiveness for tactical decision-making (SUCRA = 80.1%) and skill execution (SUCRA = 77.2%), Teaching Games for Understanding (TGfU) ranked highest for student motivation (SUCRA = 82.1%), and Sport Education (SE) demonstrated favourable effects for enjoyment (SUCRA = 74.0%). However, pairwise comparisons revealed overlapping credible intervals, indicating uncertainty in the magnitude of differences between models. In addition, significant publication bias was detected for skill execution outcomes (p = 0.0077), suggesting that these effects should be interpreted cautiously. Subgroup analyses identified instructional duration as an important moderator, with interventions approaching or exceeding approximately 720 min—equivalent to a sustained instructional unit—associated with more consistent effects and reduced heterogeneity.

No single pedagogical model can be considered universally superior. Instead, the findings support a goal-oriented approach to model selection aligned with intended learning outcomes and implemented over sufficient instructional duration. Ensuring sustained exposure (≈720 min) appears critical for enabling student-centered pedagogies to realize their educational potential.