The impact of video-based training on football referees' decision-making skills: a systematic review and meta analysis

Abstract

Objectives

Video-based decision-making training is considered a promising intervention to enhance the decision-making skills of football referees. This study conducted a systematic review and meta-analysis to validate the effectiveness of video-based training and evaluate its overall impact on improving referees’ decision-making skills, providing a scientific basis for the optimization and innovation of referee training methods.

Method

A systematic search was performed across four electronic databases (EBSCO, PubMed, Scopus, and Web of Science). Inclusion and exclusion criteria were defined using the PICOS framework. Relevant literature was independently screened, and key information was extracted. The revised Cochrane risk-of-bias tool (RoB 2) was employed to assess the risk of bias in the included studies, and statistical analyses were conducted using CMA 3.0 software.

Result

Six randomized controlled trials (RCTs) involving a total of 163 participants were included. The meta-analysis revealed that video-based training significantly improved referees’ decision-making skills (Hedges’ s g = 1.718, 95% CI [1.058, 2.377], P < 0.001, τ² = 0.464). The overall risk of bias across the included studies was assessed as low to moderate, indicating a generally reliable methodological quality. Sensitivity analysis confirmed the robustness of the overall effect size.

Conclusion

This meta-analysis demonstrates that video-based decision-making training is an effective and practical intervention for significantly enhancing the decision-making skills of football referees. Its convenience and cost-effectiveness make it an essential supplementary training tool for referees. However, the findings are limited by factors such as the small sample size of included studies, incomplete descriptions of participant characteristics, and insufficient gender representation. Future high-quality research is needed to comprehensively evaluate the effectiveness of video-based training across referees of different levels and genders.

introduction

Undoubtedly, an exciting and high-quality football match relies on the precise officiating of referees, who must ensure that both teams compete fairly and safely at all times [1, 2]. Outstanding decision-making skill is considered the cornerstone for referees to make accurate judgments during matches, which is evident in team sports such as football, basketball, and rugby.

As one of the most critical skills in sports, decision-making is typically defined as the ability to perceive information, interpret it correctly, and respond appropriately [3, 4]. For referees, this skill is paramount because each decision they make directly impacts the fairness and flow of the game. Reports indicate that referees make approximately 200–250 decisions during a professional football match, and any form of misjudgment can adversely affect player performance and even the outcome of the game [5, 6]. A high-level football referee can make the most reasonable decisions swiftly based on the current situation amidst the ever-changing game scenes [7, 8]. They maintain high levels of concentration at all times, accurately capturing both static and dynamic information from every corner of the field (including the ball's position, players' actions, and the field boundary lines), and make reliable and convincing judgments within a very short period (such as awarding free kicks or penalties) [9]. Therefore, referees must possess quick and accurate decision-making skills to adapt to various complex situations in the game, and recognizing and enhancing referees' decision-making skills is crucial for improving the quality of officiating.

The decision-making of football referees during a match is influenced by numerous factors. These include on-field environmental factors such as stadium noise [10], team or player reputation [11], player height [12], the referee's position on the field [13], the timing of the match [14], and even the players' native language [15]. Additionally, referees' own officiating experience [16], physical exertion during the match [17], different visual search strategies [18], and emotional control ability [19] can interfere with their ability to capture cues and understand the game situation, thereby affecting their decisions. Currently, as the pace of transitions in football matches is accelerating, and the competition for time and space on the field is intensifying, referees are required to make accurate and reasonable judgments under higher physical and psychological pressures, posing new challenges to their decision-making skills. Although FIFA has recently introduced technologies such as Video Assistant Referee (VAR) and Semi-Automated Offside Technology (SAOT) to assist referees in making calls, which have reduced the number of incorrect or missed calls to some extent and positively impacted the fairness of the game, the use of these technologies has also negatively affected the continuity of the game, the mentality of players and fans, and even the credibility of referees [20]. We must clarify that the purpose of technology is to assist referees in making judgments; in other words, referees themselves are still the main actors in officiating the game, and their decision-making skill remains one of the crucial factors influencing the course of the match [21]. Therefore, effectively enhancing referees' decision-making skills remains a pressing issue to be addressed.

Compared to players, referees in the field of football face a significant challenge: the lack of practice environments. This leads to a general feeling among referees of insufficient practice, making it difficult to fully improve their professional skills and judgment levels [22]. Most previous studies on referee training have focused on how to effectively enhance the physical fitness of referees, rather than honing the crucial perceptual and decision-making skills [23]. However, with the increasing importance of decision-making skills, some training efforts have shifted towards enhancing referees' decision-making skills [24]. Traditional training methods for developing decision-making skills include continuing education seminars, which focus on declarative knowledge, and match experience accumulation, which focuses on procedural knowledge [25]. However, for referees looking to improve their decision-making skills through training, frequently organizing training seminars and repeatedly recreating real match scenarios in practice is often impractical due to the inefficiencies and organizational difficulties involved. Therefore, there has been a growing call for the development of new training methods aimed at allowing referees to practice decision-making skills multiple times within a limited time frame, thereby improving their officiating levels.

In recent years, sports scientists and football organizations have been actively working on optimizing and expanding referee training systems. A video-based training method, which can both overcome existing limitations and effectively improve decision-making skills, has emerged. This method primarily simulates or recreates match scenarios in video format, allowing referees to make decisions based on the video content, thus achieving the goal of training decision-making skills [26]. This more convenient method has been proven to improve athletes' decision-making skills in competitions [27] and has gradually developed into a common way to improve the decision-making skills of referees in sports such as football, rugby, basketball, and Australian football [28]. In video-based decision training, the content is usually presented as video clips of various match scenarios. However, some researchers have found that this method seems to only bring a slight improvement in referees' decision-making skills. They attribute this to the lack of consideration for the real match environment (such as match noise, crowd characteristics, physical load, and match timing) [29]. Additionally, the video clips used in training are mostly from a third-person perspective, which researchers believe is inconsistent with what referees observe from a first-person perspective during matches, potentially reducing the ecological validity of the decision training [30].

To address the shortcomings of ordinary video training in reproducing the real match context, some researchers have attempted to develop a video decision simulator that is more closely aligned with the actual match situation. This simulator retains fan noise and requires referees to shout out their decisions while on a treadmill, aiming to simulate the real match context as closely as possible [31]. However, due to the lack of a corresponding control group in the study, it remains to be considered whether this method is more effective than ordinary video training. To make decision training more ecologically valid, researchers have expanded and updated video training using virtual reality (VR) technology. Using head-mounted displays (HMDs) to present first-person video content can enhance the interaction and immersion between individuals and the environment [32]. Currently, wearing HMDs to watch 360° videos has been proven to be an effective method for assessing referees' decision-making skills in Australian football, but its effectiveness in the field of football refereeing still needs to be verified [33].

It is evident that how to conveniently and effectively train football referees' decision-making skills remains a hot topic among researchers and practitioners. Although video-based decision training, as a new method for enhancing referees' decision-making skills, has been proven to be effective in numerous studies, it still faces some controversies and lacks robust evidence from high-standard systematic reviews and meta-analyses [34]. As football rules may change, it is also crucial to consider how decision training for referees can adapt to the development of modern football. Therefore, further analyses are needed to validate the effectiveness of video-based decision-making training before its continued use or improvement.

This study aims to quantify the impact of video-based training on football referees’ decision-making skills through a systematic review and meta-analysis by synthesizing the results of existing independent studies. The specific objectives are as follows: (1) to systematically evaluate whether video-based training significantly improves referees’ decision-making accuracy; (2) to summarize and consolidate evidence regarding the effectiveness of video-based training, exploring its applicability and potential limitations; and (3) to provide scientific insights for the optimization and innovation of referee decision-making training methods.

Materials and methods

This systematic review and meta-analysis adhered to the standards outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement, including the use of a 27-item checklist, and followed the methodological guidance provided by the Cochrane Handbook [35].

Search strategy

A systematic search was performed across four electronic databases (EBSCO, PubMed, Scopus, and Web of Science) up to January 2024. The search algorithm incorporated all possible keyword combinations from the following categories: (“video” OR “video training” OR “video intervention”) AND (“football referees” OR “soccer referees”) AND (“decision” OR “decision making” OR “judgment”). Additionally, a manual search of the reference lists of the included articles was conducted to identify any further relevant studies. Only peer-reviewed research articles published in English were considered for inclusion.

Inclusion and exclusion criteria

The inclusion criteria for this study were defined using the PICOS framework as follows: (P) Population: Participants were healthy football referees aged 18 years or older; (I) Intervention: Video-based decision-making training with clearly defined intervention frequency and duration; (C) Comparison: A control group receiving no decision-making training; (O) Outcome: The primary outcome was the accuracy of decisions made by participants in video-based decision-making tasks, specifically regarding foul situations, measured as the number or proportion of correct judgments; (S) Study design: Only randomized controlled trials (RCTs) were included, encompassing studies with either two groups (intervention and control groups) or more than two groups (e.g., two intervention groups and one control group).

Studies meeting any of the following criteria were excluded: (1) Studies involving interventions other than video-based training, such as learning football rules or physical training. (2) Studies not published in English or Chinese. (3) Systematic reviews, interviews, meta-analyses, or conference abstracts. (4) Studies not employing an RCT design. (5) Studies with no available data or duplicate publications. (6) Studies for which the full text could not be obtained.

Initially, all retrieved studies were imported into Endnote software to remove duplicates. Subsequently, two independent reviewers screened the titles and abstracts of the remaining records against the pre-defined inclusion and exclusion criteria. Full texts of studies meeting the inclusion criteria were then reviewed. In cases of disagreement between the two reviewers, discussions were conducted to justify the inclusion or exclusion of specific studies. If disagreements persisted, a third reviewer was consulted to reach a final decision researcher [36].

Data extraction

Two researchers independently extracted relevant data from the published studies using a specially designed Excel sheet. Discrepancies were resolved through discussion with a third researcher. If study data were incomplete or unclear, the researchers contacted the original authors by various methods to obtain complete information.

The primary data extracted included: (1) Basic information about the study, such as title, authors, publication time, etc.; (2) Experimental characteristics of the study, such as sample size, participant characteristics (age, gender, and skill level), intervention measures, intervention frequency, intervention duration, and outcome measures (related tests of football referees' decision-making accuracy).

Risk of bias assessment

Two researchers independently assessed the risk of bias in the included studies using the Cochrane Collaboration’s Risk of Bias 2.0 (RoB 2.0) tool. This tool evaluates the risk of bias across five domains: the randomization process, deviations from intended interventions, missing outcome data, measurement of outcomes, and selective reporting. Any disagreements during the assessment process were resolved through consultation and discussion with a third researcher to ensure the objectivity and accuracy of the evaluation results.

Data synthesis

The data format for studies included in the meta-analysis primarily consisted of pretest–posttest-control design data, including mean scores and standard deviations for pre-test and post-test in both experimental and control groups.

For pretest–posttest-control design data, a reasonable and effective method to estimate the effect size (ES) was proposed by Morris [37]. This method calculates the effect size by subtracting the mean change from pre-test to post-test in the control group from the mean change in the experimental group, and then dividing by the pooled pre-test standard deviation (see Eqs. 1 and 2). Then, as suggested by Hedges, Cohen’s d is converted to Hedge's g (see Eqs. 3 and 4) [37]. For the correlation between pre-test and post-test measurements, this study followed the recommendations of Follmann [38] and Higgins [39], using an estimated correlation of 0.5.

$$\left[\text{Pooled Pre},-,\text{test SD},=,\sqrt{\frac{(n_{\mathit{Exp}}-1)S{D}_{Pre,Exp}^2+(n_{\mathit{Con}}-1)S{D}_{Pre,Con}^2}{n_{\mathit{Exp}}+n_{\mathit{Con}}-2}}\right]$$ 1

$$\left[\text{Cohen},_{}^{\prime },\text{s},,d,=,\frac{(M_{Post,Exp}-M_{Pre,Exp})-(M_{Post,Con}-M_{Pre,Con})}{Pooled\text{Pre}-\text{test SD}}\right]$$ 2

$$\left[\mathrm{Correction},,\mathrm{factor},,J,=,,1,-,,\frac{3}{4\left(n_{\mathit{E}\mathit{x}\mathit{p}},,+,,n_{\mathit{C}\mathit{o}\mathit{n}},-,2\right)-1}\right]$$ 3

$$[\text{Hedge}_{}^{\prime }\text{s}g=J\times {\text{Cohen}}^{\prime }\text{s}d]$$ 4

Data analysis

Considering the intervention methods, participant characteristics, and decision accuracy measurement methods, this study employed a random-effects model-based meta-analysis to estimate the overall impact of video training interventions on the accuracy of football referees' decision-making tasks. Hedge's g was used to estimate the effect size (ES) with 95% CIs, determined as follows: 0.2–0.5 as small, 0.5–0.8 as medium, and greater than 0.8 as large [40].

The I² index was used to determine the degree of heterogeneity among the studies, with thresholds of low, moderate, and high heterogeneity being 25%, 50%, and 75%, respectively [41]. Additionally, τ² statistics were provided to facilitate understanding of heterogeneity among studies beyond those included in the current analysis [42].

Meta-analysis and subgroup analyses were conducted using Comprehensive Meta-Analysis (CMA) 3.0 software [43]. All analyses employed two-tailed tests, with a p-value less than 0.05 considered statistically significant.

Additionally, a leave-one-out sensitivity analysis was conducted to evaluate the influence of each individual study on the overall effect size and to assess the robustness of the analytical results.

Results

Study selection

Initially, a total of 846 studies related to this topic were retrieved from various databases. After excluding 394 duplicate studies, 452 articles remained. Upon careful reading of the titles and abstracts, 439 articles were removed. The remaining 13 articles were then subjected to full-text review, from which 5 were excluded due to irrelevant results and 2 were excluded due to inaccurate or unusable outcome measures. Ultimately, a total of 5 articles, comprising 6 studies, were included in the meta-analysis. The specific selection of studies followed the PRISMA guidelines. The overall study selection process is illustrated in Fig. 1.

Fig. 1.

Study selection flow diagram

Study characteristics

Table 1 summarizes the basic characteristics of the six included studies. This meta-analysis incorporated six randomized controlled trials (RCTs) involving a total of 163 participants, with an average age range of 23 to 40 years. These studies were conducted in Germany, the Netherlands, and Belgium. Two studies included female participants, while the remaining four involved only male participants. Except for one study that recruited university football players with refereeing experience, the participants in the other five studies were professional referees from professional or semi-professional football leagues, with refereeing experience ranging from 3 to 13 years.

Table 1.

Basic characteristics of the included studies

Study	Sample (n)	Age (years)	Gender	Skill level	Intervention	Control	Frequency (per week)	Duration (weeks)	Outcome indexes	statistical data (Pre-test/Post-test)
Schweizer & Plessner [2011]Exp 1 [26]	EG = 16 CG = 18	23.7 ± 3.2	M/FM	No-elite	Video-based training	No training	3	3	RA	EG: 25.69 ± 2.47/28.06 ± 2.62 CG: 25.61 ± 3.24/24.61 ± 2.77
Schweizer & Plessner [2011]Exp 2 [26]	EG = 19 CG = 15	25.7 ± 6.4	M/FM	Elite	Video-based training	No training	3	3	RA	EG: 25.26 ± 2.31/28.79 ± 2.49 CG: 24.87 ± 3.72/23.4 ± 3.42
van Biemen et al. (2018) [29]	EG = 11 CG = 11	31.3 ± 8.1	M	Elite	Video-based training	No training	1	1	RA(%)	EG: 75.9 ± 4.3/79.7 ± 8.5 CG: 77.3 ± 8.9/72.7 ± 7.0
PutKoen et al. [2013] [44]	EG = 10 CG = 8	29.2 ± 5.8	M	Elite	Video-based training	No training	2	3	RA	EG: 25.2 ± 3.12/31.8 ± 3.88 CG: 27.0 ± 3.21/25.5 ± 5.37
Peter Catteeuw et al. [2010a] [45]	EG = 9 CG = 22	38.8 ± 4.6	M	Elite	Video-based training	No training	1	4	RA	EG: 25.22 ± 1.01/26.94 ± 0.89 CG: 26.41 ± 0.89/25.34 ± 0.61
Peter Catteeuw et al. [2010b] [22]	EG = 10 CG = 14	39.98 ± 3.16	M	Elite	Video-based training	No training	1	4	RA	EG: 28.5 ± 3.3/31.2 ± 3 CG: 26.4 ± 5.0/24.9 ± 2.8

EG experimental group, CG control group, F female, M male, RA reaction accurate

All included studies employed RCT designs, with the intervention type predominantly consisting of video simulation training. The intervention frequency ranged from once to three times per week, and the intervention duration spanned 1 to 4 weeks. The study designs primarily compared the decision-making performance of an experimental group (receiving video simulation training) with a control group (receiving no intervention).

The assessment of referees’ decision-making abilities in all studies was conducted using decision-making tasks presented in video format. The primary outcome measure was the number (or rate) of correct responses. These tasks were based on real match scenarios that depicted key events such as fouls and offside situations, requiring referees to make judgments. This approach recreated the complex conditions of actual matches, providing referees with a diverse range of decision-making environments and making it an effective tool for evaluating referees’ decision-making abilities. The number (or rate) of correct responses was determined by the agreement between the referees’ decisions and the predetermined correct answers established by an expert panel. Higher scores typically indicated that referees could quickly and accurately extract key cues (e.g., player movements or ball trajectories) from the video scenarios and integrate this information to make reliable decisions. Conversely, lower scores reflected potential deficiencies in decision-making abilities.

Finally, all six studies demonstrated that the decision-making performance of the experimental group improved significantly in post-test evaluations, indicating that video-based training is an effective method for enhancing referees’ decision-making skills.

Risk of bias assessment

The risk of bias in the included studies was assessed using the RoB 2 tool. The results (Figs. 2 and 3) indicate that most studies exhibited a low risk of bias in the domains of selective reporting and outcome measurement, reflecting a generally high level of methodological quality. Although a few studies were rated as having “moderate risk” or “high risk” in the domains of randomization processes and missing outcome data, the overall risk of bias across the included studies was assessed as low to moderate. This suggests that the findings of the included studies have a reasonable degree of robustness.

Fig. 2.

Risk of bias graph for the included studies

Fig. 3.

Summary of risk of bias across the included studies

Meta-analysis

Figure 4 depicts the overall estimated effect size (ES) of the impact of video training on football referees' decision-making skills. The results indicate that video-based decision-making training for football referees has led to significant improvements (Hedge’s g = 1.718, 95% CI [1.058, 2.377], P < 0.001, τ² = 0.464). Additionally, an I² value of 69.579% suggests a high degree of heterogeneity among the studies.

Fig. 4.

Forest Plot of the Effect of Video Training on Referees' decision-making skills

Sensitivity analysis

The results of the sensitivity analysis indicate that no single study significantly altered the overall impact of video training on referees' decision-making skills. When one study was removed at a time, the pooled estimates ranged from g = 1.420 (95% CI [1.036, 1.805]) to g = 1.906 (95% CI [1.065, 2.747]).

Discussion

This systematic review and meta-analysis aimed to evaluate the effectiveness of video-based training in enhancing referees’ decision-making abilities. The primary analysis (g = 0.979, p < 0.001) demonstrated a significant positive impact of video-based training on referees’ decision-making performance. However, the analysis also revealed a considerable degree of heterogeneity among the included studies (I² = 69.579%). Overall, these findings suggest that video-based training is a practical and effective approach for improving referees’ decision-making skills.

Mechanisms of the impact of video training on referees' decision-making skills

Judgment and decision-making have been interpreted as processes involving a sequence of social information handling, which includes perception, classification, memory processing, and information integration [46]. Based on this framework, the decision-making process of football referees has been analyzed to follow this sequence: (1) stimulus event; (2) perception; (3) classification; (4) memory process; (5) information integration; (6) behavioral response [47]. Important steps within this sequence are considered starting points for intervention in referees' decision training, explaining the positive effects of video training on improving decision-making skills.

First, video training enhances referees' ability to accurately identify stimulus events visually. Determining whether an action is a foul is a highly complex task [48], and success in this task initially depends on the referees' visual experience (also known as visual expertise) [49]. The appearance of stimulus events (e.g., fouls) is first observed visually by referees, and visual experience plays a crucial role in decision-making performance in football refereeing [50, 51]. It represents specific collective knowledge obtained visually during matches, helping referees know when and where to focus their attention and gather visual information from the environment, thereby improving their decision-making process [47]. We hypothesize that video-based training increases referees' exposure to foul scenarios to some extent, teaching them to ignore superficial non-kinematic information (e.g., jersey colors, player body types) and focus on fundamental kinematic features representing fouls, which benefits their decision-making skill [52].

Secondly, video training significantly enhances referees' perceptual-cognitive skills. Improvements in perceptual-cognitive skills are primarily achieved by constructing and refining sport-specific knowledge structures. Compared to general physical training for the visual system, specialized training for perceptual-cognitive skills tends to yield more significant results [53]. In football, assistant referees might make incorrect offside decisions (flagging errors or non-flagging errors) due to perceptual-cognitive skill deficiencies [54]. Human visual information processing has limitations, and assistant referees may experience a phenomenon called the "flash-lag effect" when making offside decisions [55]. This perceptual error makes referees perceive the receiving player as being further forward than their actual position at the precise moment of the pass, leading to decision errors [56]. Video-based training equips referees with a compensatory strategy to cognitively correct their visual perception errors (forward memory displacement of the receiver's position), better addressing the negative effects of the flash-lag effect, thus improving offside decision accuracy [22]. It is noteworthy that video training itself may not necessarily alter visual perception; illusions might still exist, but assistant referees now consider the consequences of these illusions before making offside judgments and adjust accordingly.

Next, video training improves referees' ability to classify single events. In certain situations, even good perceptual-cognitive abilities might not guide referees to make the most accurate decisions. Judging whether a player has committed a foul falls into this category, as referees (especially the main referee) must categorize a series of different characteristic scenarios into two discrete categories (foul or no foul). Therefore, training referees' ability to classify on-field situations appears beneficial for improving decision accuracy. Since this stage is an early one in the social information processing sequence, subsequent stages also benefit from improvements here [57].

As previously discussed, referees' foul classification cues are diverse and complex. They must predict and judge distal standards (feedback or not) by considering visible cues. The more closely referees' use of cues reflects ecological validity (i.e., considering relevant cues and only these cues), the more accurate their decisions are [58–62]. Studies have shown that under time pressure in football matches, referees often rely on intuition to make quick judgments in multi-cue environments [62–64]. Intuition is believed to be based on extensive knowledge in long-term memory, which can be acquired through associative learning [65, 66]. Video-based training effectively improves referees' intuition by creating a representative learning environment with immediate, accurate feedback [67]. Under video-based training, referees’ ability to process multiple cue features was also enhanced. However, there are also experts who believe that other decision-making tasks in football refereeing (e.g., decisions regarding the management of player dissent) may require thoughtful rather than intuitive processing with reference to specific incidents, on-field situations, game moments, game management needs, etc., and therefore the results of this part of the discussion need to be viewed with caution [7].

Finally, improved memory and information integration abilities may also be key factors in enhancing referees' decision-making. As part of the decision sequence, referees must process complex information through the continuous interaction between working memory and long-term working memory [68]. They also need to integrate new or constantly changing information with existing knowledge stored in long-term memory to make decisions. With continuous practice through video-based presentations, referees can acquire and encode more representative domain-specific knowledge. Many crucial action features or cues, used as the basis for referees' decisions, can form representations in long-term memory, allowing referees to quickly search, accurately identify, and process information that matches certain memory scenarios during decision-making, thereby making accurate foul judgments [45]. Additionally, video practice develops referees' domain-specific working memory abilities, including long-term working memory. Through practice, referees can improve their ability to temporarily store and process important relevant information, such as player identities, field positions, and on-field events. Leveraging excellent working memory capabilities, referees can effectively search for relevant information for key tasks from long-term memory, contributing to improved decision-making skills [21].

Practical applications

The main findings of this study indicate that video-based training is highly effective in enhancing referees’ decision-making abilities. This effectiveness is primarily achieved by strengthening referees’ perceptual-cognitive skills, optimizing the classification and integration of key cues, and improving their ability to respond rapidly in dynamic scenarios. By leveraging its flexibility and repeatability, video-based training allows for the simulation of diverse match situations, addressing the limitation of insufficient decision-making practice opportunities for referees. Therefore, in addition to traditional training focused on physical fitness, strength, and endurance, video-based decision-making exercises should be incorporated into referees’ regular training programs.

It is important to note that continually improving the quality and validity of video training remains a challenge. Traditional video training methods, which rely on match broadcast perspectives, are still somewhat controversial. While convenient to produce and play, many researchers believe that the overhead television broadcast view does not fully represent the first-person perspective observed by referees during matches. Traditional video training also lacks factors such as subjective fatigue, on-site noise, and fan influence, thus diverging from the actual match conditions [69].

In response, researchers have developed new decision simulators that require referees to watch videos on a tablet while running on a treadmill and verbally make decisions [31]. Additionally, 360-degree video training methods have been attempted in Australian football for referees' decision-making [70]. Exploring more effective and ecologically valid video training methods is currently a trend. Recent studies have also proposed a new decision-making framework for football referees, dividing their decisions into rule application and match management aspects, reconsidering various factors influencing referees' decisions [71].

Future efforts should utilize VR technology and integrate new referee decision-making frameworks to develop targeted and ecologically valid video training programs that closely mimic real match environments, achieving more comprehensive and effective training outcomes. Moreover, it is worth exploring how much of the training benefits from video training can be transferred to referees' match decisions.

Strengths and limitations

This study holds notable academic value and practical significance. First, the research strictly adhered to PRISMA principles and included only randomized controlled trials (RCTs), ensuring the quality and reliability of the findings. Second, a random-effects model was employed to analyze heterogeneity, providing a robust basis for evaluating the overall effectiveness of video-based training. Furthermore, the study specifically focused on the impact of video training on referees’ decision-making skills, addressing a gap in the field to some extent and offering a referential framework for future research. Lastly, the findings provide direct evidence and forward-looking recommendations for the application and optimization of video-based training methods.

Nevertheless, several limitations should be considered when interpreting and applying the results of this study. First, the limited number of included studies and the small overall sample size may affect the accuracy of effect estimates and pose a risk of overestimating the effectiveness of video training. Future research should aim to include a larger number of studies and expand sample sizes to verify the robustness of these conclusions. Second, the included studies provided insufficient descriptions of participants’ skill levels, making it challenging to explore how variations in skill levels may influence training outcomes. Future studies should clearly report and classify participants’ professional backgrounds. Finally, the low proportion of female referees in the current research highlights a lack of gender representation, potentially limiting the generalizability of the findings. Future research should include more female referees to evaluate whether gender differences influence the effectiveness of video-based training.

Conclusion

This study provides compelling RCT evidence that video-based decision-making training is an effective method for enhancing the decision-making skills of football referees. By simulating real match scenarios, this training approach strengthens referees’ ability to identify key events and make rapid judgments in dynamic game environments, thereby improving their overall officiating performance. Although the positive effects of video-based training may be influenced by small sample sizes, limited study effects, and low ecological validity, this promising tool can serve as a valuable supplementary training method for referees off the field. Future research should focus on refining and optimizing this approach to achieve even greater training efficacy.

Authors’ contributions

The authors contributed to this article as follows: Zhou and Zhang conceived the idea of the paper; Zhou, Zhang, and Li searched, screened articles, and collected data; Deng and Song performed quality assessment; and the paper was written, directed, and edited by Zhou and Zhang. All authors contributed to the article and approved the submitted version.

Funding

2023 Beijing Sport University "Philosophy and Social Sciences Research High Quality Development Special Project" (No. 2023992981).

Data availability

The corresponding author can provide the datasets used and/or analyzed for this work upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.