Meta-analysis and evidence synthesis for sports psychology

Evidence synthesis in sports psychology

Sports psychology asks how mental processes affect performance and how interventions can help athletes. Topics include imagery, self-talk, goal setting, relaxation, attentional focus, pre-performance routines, motivation, anxiety and arousal, team cohesion, coach behavior, burnout and well-being. Meta-analyses have been published on most, often reporting small to moderate positive effects of psychological interventions on performance.

The evidence has features that call for care. Experiments are small, often with 20 to 40 participants, and many use students or recreational athletes on laboratory tasks. Performance outcomes are varied, from accuracy in a dart throw to times in a race. Survey studies rely on self-reported scales, and links between psychological variables and performance are often weak. Our methods follow meta-analysis and systematic review practice, adapted to these features. This page builds on the general guidance for sports and exercise sciences and the neighboring psychology area.

Performance tasks and ecological validity

Reviews of psychological skills training commonly find larger effects on laboratory tasks than on competitive performance. A review should code outcome type and separate it in analysis, and should present competition outcomes where available because they are closest to what athletes care about. Skill level also matters: effects in novices are often larger than in elite athletes, who have less room to improve and already use many mental strategies of their own.

Blinding, expectation and placebo effects

Participants who know they are receiving a psychological intervention expect it to help, and expectation can itself change performance. Active controls, such as attention-matched conditions, address this better than no-treatment controls. Blinding participants is often impossible, but outcome assessors can be blinded. A review should code the control condition and blinding of assessors and examine whether effects are smaller with active controls. Effects measured by athletes' own ratings are the most open to expectation, so they are analyzed separately from objective performance.

This issue is not unique to sport, and the research on placebo effects in performance has shown that belief about a treatment can change measured outcomes. It is a feature of the evidence to be reported, not a reason to ignore the studies.

Self-report scales and constructs

Anxiety, confidence, motivation and flow are measured with questionnaires, such as sport-specific anxiety scales. Different instruments define the constructs differently, and versions have been revised. Correlations between psychological scales and performance are usually small, and measures taken on the day of competition are rarer than general ones. A review should record the instrument, its version, the timing relative to competition and whether the scale is a trait or state measure. It should also note that studies often report reliability for their own sample, and it can use those values for psychometric corrections, presenting corrected and uncorrected estimates separately.

Small samples and statistical power

Small samples limit the power of studies and make estimates imprecise and prone to exaggeration when they pass the significance filter. Sports psychology has been described as having low average power, so a literature of significant findings may include many overestimates. We use bias-corrected standardized differences, random-effects models with small-sample adjustments and publication-bias methods, and we report how many studies back each analysis. Preregistration and registered reports are still uncommon in the field, but they are becoming more frequent, and a review can compare registered with unregistered studies as they appear.

Teams, coaches and group-level evidence

Team cohesion, leadership and motivational climate are studied with surveys of athletes in teams. Data are clustered, since players belong to teams, and analyses that ignore this overstate precision. Reviews should record the level of analysis, the number of teams and whether authors used multilevel methods. Meta-analyses have found positive associations between cohesion and team performance, with moderators such as team type (interacting or coacting sports) and the type of cohesion. The direction of influence is unclear because success may also increase cohesion, and cross-sectional designs cannot sort this out.

Athlete well-being and mental health

Research on burnout, stress, injury-related distress and mental health in athletes uses screening questionnaires and, sometimes, clinical interviews. Prevalence estimates depend on the instrument and cut-offs, and studies are often cross-sectional with convenience samples. A review should state these limits, present prevalence with the methods described on the prevalence meta-analysis page, and avoid presenting screening results as diagnoses. This service does not provide mental health advice or treatment. Athletes who need support should consult a qualified professional.

Publication bias and researcher flexibility

Studies with larger effects are more likely to be published, journals favor novel positive findings, and flexible choices about outcomes and analyses can produce significant results. We use funnel plots, selection models and p-curve analysis, compare published with unpublished work and check whether effects are larger in studies that did not report a sample size justification. As with other small-sample fields, these methods have limits and are reported together with their assumptions, so that readers can judge how much weight to give the corrected figures.

Imagery, self-talk and attentional focus

Three intervention families dominate performance studies. Imagery involves mentally rehearsing a skill or situation; self-talk uses cue words or instructional statements; attentional focus instructions direct attention to the movement or to its effect on the environment. Meta-analyses have found small to moderate average benefits for each, with larger effects for instructional self-talk on fine motor tasks and for external focus on motor learning tasks. Findings come largely from short laboratory experiments with novices, and there are few studies of long-term use in competitive athletes.

The definition of the intervention matters. Imagery can be visual or kinesthetic, internal or external, and delivered once or rehearsed for weeks. A review should code content, delivery, dose and whether participants practiced the skill physically as well, since the added value of mental practice is judged against physical practice alone. Studies often compare mental plus physical practice with no practice, which is a weaker test than comparing with physical practice alone.

Motivation, climate and coaching

Research based on self-determination theory and achievement goal theory examines how coach behavior and the motivational climate relate to athletes' satisfaction, persistence, effort and well-being. Surveys of athletes dominate, and associations are moderate for climates that support autonomy and mastery. The studies are cross-sectional more often than experimental, and few include objective performance. A review should be careful to describe these as associations, and should note that athletes' perceptions of the climate may reflect their own mood and success. Intervention studies that train coaches are growing in number and offer stronger evidence; they usually measure athlete reports as outcomes, which means that the people who receive the intervention also judge its effect, and a review should flag that. Differences between male and female athletes, between age groups and between individual and team sports are examined as moderators when enough studies report them, and the review states plainly when too few do.

Reviews should present the countries and sports covered, since coaching cultures differ, and be open about the dominance of studies from a few countries and from team sports.

Common pitfalls we look for

• Treating laboratory task results as competition results.
• Using no-treatment controls without noting expectation effects.
• Pooling different scales for the same named construct.
• Ignoring clustering of players within teams.
• Interpreting cross-sectional links as causes.
• Presenting screening scores as diagnoses.

Planning a sports psychology synthesis

We help define the intervention or construct, the athlete population and the outcome, plan searches in PsycINFO, SPORTDiscus, MEDLINE and Web of Science, and set up coding of control type, task, skill level, instruments and level of analysis. See the meta-analysis service for scope and process.

An invented example of interpreting an effect

Suppose a review of imagery training finds an average standardized difference of 0.55 for performance on laboratory motor tasks, 0.25 on sport-specific skill tests and 0.10 on competition outcomes, all with wide intervals. The pattern tells a story: effects shrink as the outcome moves closer to real competition. If studies with active control groups show smaller effects than those with no-treatment controls, part of the effect may be expectation. Readers interested in competition would reasonably read the review as showing small and uncertain benefits for performance, while noting that athletes may still value the practices.

Coding and transparency

Coding frames record sport, level (novice, collegiate, elite), age, sex, task, outcome type, intervention content and length, control type, assessor blinding, instruments, level of analysis and funding. Two coders work independently on a sample, agreement is reported and the data and code are shared. Where interventions are described briefly, we use templates for intervention description so that content can be compared across studies, and we contact authors for missing information.

How we support research projects in this area

Boundaries of this service

A sports psychology synthesis describes average effects across published studies. It does not provide psychological treatment, mental health advice, athlete selection or performance coaching. Most studies are small and use laboratory tasks or recreational athletes, so results may not carry over to elite competition.

Frequently asked questions