FitProven
Evidence Breakdown
Based on 6 studies
THE BIGGER QUESTION
Whether this one statement holds is settled above. What to actually do about it is a wider question, weighed across every claim that bears on it.
- Should I stretch before a workout? — this claim is one of 3 weighed there
Evidence map
For & against, at a glance
Meta-Analysis
Pro
Simic L et al. · 2013Scandinavian Journal of Medicine & Science in SportsMeta-analysis of 104 studies (61 data points for strength, 12 for power, 57 for explosive performance) pooling the acute effect of pre-exercise static stretching. The paper reports each outcome twice, in standardized units and in percentages, and the two do not agree on significance:
- **Strength**: ES −0.10 (95% CI −0.15 to −0.04), or −5.4% (95% CI −6.6% to −4.2%) — significantly reduced on both metrics.
- **Power**: ES −0.04 (95% CI −0.16 to 0.08), or −1.9% (95% CI −4.0% to 0.2%) — not significant on either.
- **Explosive performance**: ES −0.03 (95% CI −0.07 to 0.01) — *not* significant standardized; but −2.0% (95% CI −2.8% to −1.3%) in percentage terms, where the interval excludes zero and the reduction *is* significant.
Effects were unrelated to age, sex or fitness level, were more pronounced in isometric than dynamic tests, and were related to total stretch duration, with the smallest negative effects at durations ≤45 s. The authors concluded static stretching as the sole warm-up activity should generally be avoided.
0.90
Meta-analysis of 104 studies (61 data points for strength, 12 for power, 57 for explosive performance) pooling the acute effect of pre-exercise static stretching. The paper reports each outcome twice, in standardized units and in percentages, and the two do not agree on significance: - **Strength**: ES −0.10 (95% CI −0.15 to −0.04), or −5.4% (95% CI −6.6% to −4.2%) — significantly reduced on both metrics. - **Power**: ES −0.04 (95% CI −0.16 to 0.08), or −1.9% (95% CI −4.0% to 0.2%) — not significant on either. - **Explosive performance**: ES −0.03 (95% CI −0.07 to 0.01) — *not* significant standardized; but −2.0% (95% CI −2.8% to −1.3%) in percentage terms, where the interval excludes zero and the reduction *is* significant. Effects were unrelated to age, sex or fitness level, were more pronounced in isometric than dynamic tests, and were related to total stretch duration, with the smallest negative effects at durations ≤45 s. The authors concluded static stretching as the sole warm-up activity should generally be avoided.
Design Meta-Analysis (1.0) × quality 0.90 = impact 0.90
View sourceSystematic Review
Pro
Behm DG et al. · 2016Applied Physiology, Nutrition, and MetabolismSystematic review of the acute effects of static, dynamic and PNF stretching on performance, range of motion and injury. Static stretching produced an average 3.7% performance decrement and PNF 4.4%, while dynamic stretching produced a 1.3% improvement, with the impairments greatest when testing occurred immediately after stretching and when stretch duration exceeded 60 s per muscle group. The authors still recommended static stretching can be retained in a warm-up provided it is followed by dynamic activity.
0.72
Systematic review of the acute effects of static, dynamic and PNF stretching on performance, range of motion and injury. Static stretching produced an average 3.7% performance decrement and PNF 4.4%, while dynamic stretching produced a 1.3% improvement, with the impairments greatest when testing occurred immediately after stretching and when stretch duration exceeded 60 s per muscle group. The authors still recommended static stretching can be retained in a warm-up provided it is followed by dynamic activity.
Design Systematic Review (0.9) × quality 0.80 = impact 0.72
View sourceMeta-Analysis
Con
Warneke K & Lohmann LH · 2024Journal of Sport and Health ScienceMultilevel meta-analysis restricted to controlled pre–post designs: 83 studies, 2,012 participants, >400 effect sizes. Maximal strength showed only a small overall reduction (ES −0.21, p = 0.003), driven almost entirely by long holds — bouts ≥60 s produced a large deficit (ES −0.84, 95% CI −1.32 to −0.37, p = 0.004), while bouts <60 s produced no significant deficit at all (ES −0.13, 95% CI −0.32 to 0.07, p = 0.20). Crucially, athletic performance measures (jumping, sprinting, throwing) showed no impairment and jumping was even trivially improved (ES 0.15, p = 0.006); the authors state the data do not support excluding static stretching from warm-ups.
0.90
Multilevel meta-analysis restricted to controlled pre–post designs: 83 studies, 2,012 participants, >400 effect sizes. Maximal strength showed only a small overall reduction (ES −0.21, p = 0.003), driven almost entirely by long holds — bouts ≥60 s produced a large deficit (ES −0.84, 95% CI −1.32 to −0.37, p = 0.004), while bouts <60 s produced no significant deficit at all (ES −0.13, 95% CI −0.32 to 0.07, p = 0.20). Crucially, athletic performance measures (jumping, sprinting, throwing) showed no impairment and jumping was even trivially improved (ES 0.15, p = 0.006); the authors state the data do not support excluding static stretching from warm-ups.
Design Meta-Analysis (1.0) × quality 0.90 = impact 0.90
View sourceSystematic Review
Con
Kay AD & Blazevich AJ · 2012Medicine & Science in Sports & ExerciseSystematic review of 106 studies (yielding 149 findings) examining acute static stretch and maximal muscle performance, analysed by stretch duration. The authors explicitly declined to meta-analyse — "meta-analysis was deemed to be neither feasible nor appropriate" — and instead pooled raw percentage changes by duration band.
In the duration-stratified analysis (Table 1, 104 findings), stretches <30 s had no meaningful effect (pooled −1.1 ± 1.8%) and 30–45 s stretches were also non-significant (−1.9 ± 3.4%), whereas 60–120 s produced −4.2 ± 5.0% and >120 s produced −7.0 ± 5.7%. The abstract states there is "overwhelming evidence that stretch durations of 30–45 s also imparted no significant effect," with "a significant reduction likely to occur with stretches ≥60 s," and that this "strong evidence for a dose–response effect was independent of performance task, contraction mode, or muscle group." The authors concluded that "shorter durations of stretch (<60 s) can be performed in a preexercise routine without compromising maximal muscle performance" — i.e. the deficit is real but duration-dependent, not a property of static stretching per se.
(An earlier version of this record cited "up to about −7.5%" as the ≥60 s effect. That figure is the plantar-flexor subgroup at ≥60 s only (−7.5 ± 7.8%), not the pooled duration effect across muscle groups.)
0.72
Systematic review of 106 studies (yielding 149 findings) examining acute static stretch and maximal muscle performance, analysed by stretch duration. The authors explicitly declined to meta-analyse — "meta-analysis was deemed to be neither feasible nor appropriate" — and instead pooled raw percentage changes by duration band. In the duration-stratified analysis (Table 1, 104 findings), stretches <30 s had no meaningful effect (pooled −1.1 ± 1.8%) and 30–45 s stretches were also non-significant (−1.9 ± 3.4%), whereas 60–120 s produced −4.2 ± 5.0% and >120 s produced −7.0 ± 5.7%. The abstract states there is "overwhelming evidence that stretch durations of 30–45 s also imparted no significant effect," with "a significant reduction likely to occur with stretches ≥60 s," and that this "strong evidence for a dose–response effect was independent of performance task, contraction mode, or muscle group." The authors concluded that "shorter durations of stretch (<60 s) can be performed in a preexercise routine without compromising maximal muscle performance" — i.e. the deficit is real but duration-dependent, not a property of static stretching per se. (An earlier version of this record cited "up to about −7.5%" as the ≥60 s effect. That figure is the plantar-flexor subgroup at ≥60 s only (−7.5 ± 7.8%), not the pooled duration effect across muscle groups.)
Design Systematic Review (0.9) × quality 0.80 = impact 0.72
View sourceMeta-Analysis
Pro
Simic L et al. · 2013Scandinavian Journal of Medicine & Science in SportsMeta-analysis of 104 studies (61 data points for strength, 12 for power, 57 for explosive performance) pooling the acute effect of pre-exercise static stretching. The paper reports each outcome twice, in standardized units and in percentages, and the two do not agree on significance:
- **Strength**: ES −0.10 (95% CI −0.15 to −0.04), or −5.4% (95% CI −6.6% to −4.2%) — significantly reduced on both metrics.
- **Power**: ES −0.04 (95% CI −0.16 to 0.08), or −1.9% (95% CI −4.0% to 0.2%) — not significant on either.
- **Explosive performance**: ES −0.03 (95% CI −0.07 to 0.01) — *not* significant standardized; but −2.0% (95% CI −2.8% to −1.3%) in percentage terms, where the interval excludes zero and the reduction *is* significant.
Effects were unrelated to age, sex or fitness level, were more pronounced in isometric than dynamic tests, and were related to total stretch duration, with the smallest negative effects at durations ≤45 s. The authors concluded static stretching as the sole warm-up activity should generally be avoided.
0.90
Meta-analysis of 104 studies (61 data points for strength, 12 for power, 57 for explosive performance) pooling the acute effect of pre-exercise static stretching. The paper reports each outcome twice, in standardized units and in percentages, and the two do not agree on significance: - **Strength**: ES −0.10 (95% CI −0.15 to −0.04), or −5.4% (95% CI −6.6% to −4.2%) — significantly reduced on both metrics. - **Power**: ES −0.04 (95% CI −0.16 to 0.08), or −1.9% (95% CI −4.0% to 0.2%) — not significant on either. - **Explosive performance**: ES −0.03 (95% CI −0.07 to 0.01) — *not* significant standardized; but −2.0% (95% CI −2.8% to −1.3%) in percentage terms, where the interval excludes zero and the reduction *is* significant. Effects were unrelated to age, sex or fitness level, were more pronounced in isometric than dynamic tests, and were related to total stretch duration, with the smallest negative effects at durations ≤45 s. The authors concluded static stretching as the sole warm-up activity should generally be avoided.
Design Meta-Analysis (1.0) × quality 0.90 = impact 0.90
View sourceSystematic Review
Pro
Behm DG et al. · 2016Applied Physiology, Nutrition, and MetabolismSystematic review of the acute effects of static, dynamic and PNF stretching on performance, range of motion and injury. Static stretching produced an average 3.7% performance decrement and PNF 4.4%, while dynamic stretching produced a 1.3% improvement, with the impairments greatest when testing occurred immediately after stretching and when stretch duration exceeded 60 s per muscle group. The authors still recommended static stretching can be retained in a warm-up provided it is followed by dynamic activity.
0.72
Systematic review of the acute effects of static, dynamic and PNF stretching on performance, range of motion and injury. Static stretching produced an average 3.7% performance decrement and PNF 4.4%, while dynamic stretching produced a 1.3% improvement, with the impairments greatest when testing occurred immediately after stretching and when stretch duration exceeded 60 s per muscle group. The authors still recommended static stretching can be retained in a warm-up provided it is followed by dynamic activity.
Design Systematic Review (0.9) × quality 0.80 = impact 0.72
View sourceMeta-Analysis
Con
Warneke K & Lohmann LH · 2024Journal of Sport and Health ScienceMultilevel meta-analysis restricted to controlled pre–post designs: 83 studies, 2,012 participants, >400 effect sizes. Maximal strength showed only a small overall reduction (ES −0.21, p = 0.003), driven almost entirely by long holds — bouts ≥60 s produced a large deficit (ES −0.84, 95% CI −1.32 to −0.37, p = 0.004), while bouts <60 s produced no significant deficit at all (ES −0.13, 95% CI −0.32 to 0.07, p = 0.20). Crucially, athletic performance measures (jumping, sprinting, throwing) showed no impairment and jumping was even trivially improved (ES 0.15, p = 0.006); the authors state the data do not support excluding static stretching from warm-ups.
0.90
Multilevel meta-analysis restricted to controlled pre–post designs: 83 studies, 2,012 participants, >400 effect sizes. Maximal strength showed only a small overall reduction (ES −0.21, p = 0.003), driven almost entirely by long holds — bouts ≥60 s produced a large deficit (ES −0.84, 95% CI −1.32 to −0.37, p = 0.004), while bouts <60 s produced no significant deficit at all (ES −0.13, 95% CI −0.32 to 0.07, p = 0.20). Crucially, athletic performance measures (jumping, sprinting, throwing) showed no impairment and jumping was even trivially improved (ES 0.15, p = 0.006); the authors state the data do not support excluding static stretching from warm-ups.
Design Meta-Analysis (1.0) × quality 0.90 = impact 0.90
View sourceSystematic Review
Con
Kay AD & Blazevich AJ · 2012Medicine & Science in Sports & ExerciseSystematic review of 106 studies (yielding 149 findings) examining acute static stretch and maximal muscle performance, analysed by stretch duration. The authors explicitly declined to meta-analyse — "meta-analysis was deemed to be neither feasible nor appropriate" — and instead pooled raw percentage changes by duration band.
In the duration-stratified analysis (Table 1, 104 findings), stretches <30 s had no meaningful effect (pooled −1.1 ± 1.8%) and 30–45 s stretches were also non-significant (−1.9 ± 3.4%), whereas 60–120 s produced −4.2 ± 5.0% and >120 s produced −7.0 ± 5.7%. The abstract states there is "overwhelming evidence that stretch durations of 30–45 s also imparted no significant effect," with "a significant reduction likely to occur with stretches ≥60 s," and that this "strong evidence for a dose–response effect was independent of performance task, contraction mode, or muscle group." The authors concluded that "shorter durations of stretch (<60 s) can be performed in a preexercise routine without compromising maximal muscle performance" — i.e. the deficit is real but duration-dependent, not a property of static stretching per se.
(An earlier version of this record cited "up to about −7.5%" as the ≥60 s effect. That figure is the plantar-flexor subgroup at ≥60 s only (−7.5 ± 7.8%), not the pooled duration effect across muscle groups.)
0.72
Systematic review of 106 studies (yielding 149 findings) examining acute static stretch and maximal muscle performance, analysed by stretch duration. The authors explicitly declined to meta-analyse — "meta-analysis was deemed to be neither feasible nor appropriate" — and instead pooled raw percentage changes by duration band. In the duration-stratified analysis (Table 1, 104 findings), stretches <30 s had no meaningful effect (pooled −1.1 ± 1.8%) and 30–45 s stretches were also non-significant (−1.9 ± 3.4%), whereas 60–120 s produced −4.2 ± 5.0% and >120 s produced −7.0 ± 5.7%. The abstract states there is "overwhelming evidence that stretch durations of 30–45 s also imparted no significant effect," with "a significant reduction likely to occur with stretches ≥60 s," and that this "strong evidence for a dose–response effect was independent of performance task, contraction mode, or muscle group." The authors concluded that "shorter durations of stretch (<60 s) can be performed in a preexercise routine without compromising maximal muscle performance" — i.e. the deficit is real but duration-dependent, not a property of static stretching per se. (An earlier version of this record cited "up to about −7.5%" as the ≥60 s effect. That figure is the plantar-flexor subgroup at ≥60 s only (−7.5 ± 7.8%), not the pooled duration effect across muscle groups.)
Design Systematic Review (0.9) × quality 0.80 = impact 0.72
View sourceShowing the 4 strongest of 6 studies. Tap any node to expand its detail.
Evidence
PRO (2)
PRO Meta-Analysis0.90 Simic L, Sarabon N et al. (2013)
Acute pre-exercise static stretching reduced strength (ES -0.10, 95% CI -0.15 to -0.04; about -5.4%); pooled effects on power (ES -0.04, 95% CI -0.16 to 0.08) and explosive performance (ES -0.03, 95% CI -0.07 to 0.01) were not significant
Meta-analysis of 104 studies (61 data points for strength, 12 for power, 57 for explosive performance) pooling the acute effect of pre-exercise static stretching. The paper reports each outcome twice, in standardized units and in percentages, and the two do not agree on significance:
- Strength: ES −0.10 (95% CI −0.15 to −0.04), or −5.4% (95% CI −6.6% to −4.2%) — significantly reduced on both metrics.
- Power: ES −0.04 (95% CI −0.16 to 0.08), or −1.9% (95% CI −4.0% to 0.2%) — not significant on either.
- Explosive performance: ES −0.03 (95% CI −0.07 to 0.01) — not significant standardized; but −2.0% (95% CI −2.8% to −1.3%) in percentage terms, where the interval excludes zero and the reduction is significant.
Effects were unrelated to age, sex or fitness level, were more pronounced in isometric than dynamic tests, and were related to total stretch duration, with the smallest negative effects at durations ≤45 s. The authors concluded static stretching as the sole warm-up activity should generally be avoided.
Weighted 0.90 — k=104 studies (61 data points for strength, 12 for power, 57 for explosive performance). No pooled participant total is reported -- the unit of analysis is data points, not people -- so sample_size is omitted rather than invented. A large, well-conducted meta-analysis with tight confidence intervals and a dose-response gradient (smallest decrements at stretch durations <=45 s), which is why it sits near the top of its class. Funding statement not retrievable (paywalled).
Scandinavian Journal of Medicine & Science in Sports
PRO Systematic Review0.80 Behm DG, Blazevich AJ et al. (2016)
Acute performance change: static stretching -3.7%, PNF -4.4%, dynamic +1.3%; dose-dependent, with -4.6% for 60 s or more of static stretch per muscle group vs -1.1% below 60 s; no clear effect of static or PNF stretching on all-cause or overuse injury
Systematic review of the acute effects of static, dynamic and PNF stretching on performance, range of motion and injury. Static stretching produced an average 3.7% performance decrement and PNF 4.4%, while dynamic stretching produced a 1.3% improvement, with the impairments greatest when testing occurred immediately after stretching and when stretch duration exceeded 60 s per muscle group. The authors still recommended static stretching can be retained in a warm-up provided it is followed by dynamic activity.
Weighted 0.80 — Very large evidence base (well over 100 acute-stretching studies across performance, ROM and injury), by the four authors who define this literature, and the conflict-of-interest statement declares none. sample_size is omitted because the review pools percentage changes per outcome rather than participants, and no total is reported. Held just below the top band because the synthesis is descriptive rather than a formal meta-analysis with GRADE or risk-of-bias scoring. No funding statement appears in the paper.
Applied Physiology, Nutrition, and Metabolism
AGAINST (4)
AGAINST Meta-Analysisn=20120.90 Warneke K, Lohmann LH (2024)
Maximal strength ES -0.21 (p = 0.003) overall, driven by long holds: bouts >=60 s ES -0.84 (95% CI -1.32 to -0.37, p = 0.004) vs bouts <60 s ES -0.13 (95% CI -0.32 to 0.07, p = 0.20, non-significant). Jumping, sprinting and throwing unimpaired (jumping trivially improved, ES 0.15, p = 0.006)
Multilevel meta-analysis restricted to controlled pre–post designs: 83 studies, 2,012 participants, >400 effect sizes. Maximal strength showed only a small overall reduction (ES −0.21, p = 0.003), driven almost entirely by long holds — bouts ≥60 s produced a large deficit (ES −0.84, 95% CI −1.32 to −0.37, p = 0.004), while bouts <60 s produced no significant deficit at all (ES −0.13, 95% CI −0.32 to 0.07, p = 0.20). Crucially, athletic performance measures (jumping, sprinting, throwing) showed no impairment and jumping was even trivially improved (ES 0.15, p = 0.006); the authors state the data do not support excluding static stretching from warm-ups.
Weighted 0.90 — 83 studies, 2,012 participants and over 400 effect sizes in a multilevel model, with the dose-response subgrouping that actually resolves the field's contradiction — the deficit lives in long holds only. No funding and no competing interests declared. Its ceiling is the quality of the underlying controlled pre-post trials, not its own methods.
Journal of Sport and Health Science
AGAINST Systematic Review0.80 Kay AD, Blazevich AJ (2012)
Stretches under 30 s had no meaningful effect (pooled -1.1 +/- 1.8%) and 30-45 s were also non-significant (-1.9 +/- 3.4%), whereas 60-120 s impaired performance by -4.2 +/- 5.0% and stretches over 120 s by -7.0 +/- 5.7% -- a sigmoidal dose-response that the authors report is independent of performance task, contraction mode and muscle group
Systematic review of 106 studies (yielding 149 findings) examining acute static stretch and maximal muscle performance, analysed by stretch duration. The authors explicitly declined to meta-analyse — "meta-analysis was deemed to be neither feasible nor appropriate" — and instead pooled raw percentage changes by duration band.
In the duration-stratified analysis (Table 1, 104 findings), stretches <30 s had no meaningful effect (pooled −1.1 ± 1.8%) and 30–45 s stretches were also non-significant (−1.9 ± 3.4%), whereas 60–120 s produced −4.2 ± 5.0% and >120 s produced −7.0 ± 5.7%. The abstract states there is "overwhelming evidence that stretch durations of 30–45 s also imparted no significant effect," with "a significant reduction likely to occur with stretches ≥60 s," and that this "strong evidence for a dose–response effect was independent of performance task, contraction mode, or muscle group." The authors concluded that "shorter durations of stretch (<60 s) can be performed in a preexercise routine without compromising maximal muscle performance" — i.e. the deficit is real but duration-dependent, not a property of static stretching per se.
(An earlier version of this record cited "up to about −7.5%" as the ≥60 s effect. That figure is the plantar-flexor subgroup at ≥60 s only (−7.5 ± 7.8%), not the pooled duration effect across muscle groups.)
Weighted 0.80 — k=106 studies yielding 149 findings -- a large, independent evidence base with an explicit no-funding statement. But it pools raw percentage changes rather than participant-weighted effect sizes, states no total participant count (so sample_size is omitted), and applies no formal risk-of-bias or GRADE assessment, which makes the 60-second boundary approximate rather than precise.
Funding: none declared
Medicine & Science in Sports & Exercise
AGAINST Crossover Trialn=200.65 Blazevich AJ, Gill ND et al. (2018)
No effect of any stretch condition (5 s static, 30 s static as 3 x 10 s, 5-rep dynamic, or none) embedded in an otherwise identical full warm-up on 20 m sprint, vertical jump or change-of-direction performance; 18 of 20 athletes expected dynamic stretching to help most, yet nothing changed
Randomised, experimenter-blinded crossover trial in 20 male team-sport athletes comparing a 5 s static stretch, a 30 s static stretch (3 × 10 s), a 5-repetition dynamic stretch and no stretch, each embedded in an otherwise identical full warm-up, across 9 muscle regions. There were no effects of stretch condition on 20 m sprint, vertical jump or change-of-direction performance. Notably, 18 of 20 athletes expected dynamic stretching to help most, yet no condition changed measured performance — evidence that short static stretches inside a complete warm-up do not cost strength or power.
Weighted 0.65 — n=20 male team-sport athletes, randomised experimenter-blinded crossover across 9 muscle regions with the stretch embedded in a realistic complete warm-up — the ecological validity is the whole point of the study and is excellent. Discounted for small n, male-only, and the fact that a null in 20 athletes cannot exclude a small effect. No funding statement could be verified, so the field is omitted.
Medicine & Science in Sports & Exercise
AGAINST Narrative Review0.55 Chaabene H, Behm DG et al. (2019)
Static stretching of <=60 s per muscle group causes only trivial (1-2%) strength/power impairment; >60 s produces 4.0-7.5% declines; the effect is negligible when short holds sit inside a full warm-up
Review re-examining the caveats behind the static-stretching performance literature. It concludes that ≤60 s of static stretching per muscle group causes only trivial impairment (1–2%) of strength and power, while >60 s produces substantial declines (4.0–7.5%), and that when short static stretching is embedded in a full warm-up containing aerobic and sport-specific work the negative effect is negligible. The authors attribute this to short holds not measurably altering neuromuscular activation or musculotendinous stiffness.
Weighted 0.55 — Narrative review, so no pooled sample size and no systematic search or risk-of-bias appraisal — selection of the cited trials is at the authors' discretion. Rated at the top of the narrative-review band because it is independent, written by domain specialists, and its dose-response conclusion agrees with the subsequent meta-analytic literature.
Funding: Deutsche Forschungsgemeinschaft (DFG) and the University of Potsdam Open Access Publishing Fund
Frontiers in Physiology