How weekly set volume relates to muscle growth
Across 67 studies, the posterior probability that the marginal slope exceeded zero — that gains rise as weekly set volume rises — was 100% for both muscle size and strength. The best-fit models suggest that increase flattens out as volume climbs, with the flattening considerably more pronounced for strength than for size. For weekly training frequency the picture splits by outcome: the posterior probability was 100% for strength, while for hypertrophy it was less than 100%, which the authors read as compatibility with negligible effects.
Joshua C Pelland (2026) · Sports Medicine (Auckland, N.Z.) · PMID 41343037
It does not name a number of weekly sets. No threshold, no optimum, no ceiling, no "X sets per muscle per week" — anyone quoting this paper for a specific set count is inventing that number. The abstract reports no effect size and no confidence interval anywhere, so the 100% figures are directional certainty about a slope, not a measure of how much volume delivers.
"Diminishing returns" is not "a point where more volume stops working" and not "a point where more volume becomes harmful." The abstract reports neither. Each added set buys less than the one before; the paper does not locate a place where it buys nothing.
It does not show that an indirect set produces half the growth of a direct one. The paper compared three ways of counting indirect sets (1, 0.5, 0) and found the 0.5 model had the strongest relative evidence, so it used that for the primary models. That is a counting convention selected by model comparison, not a measured exchange rate of stimulus.
It does not rank frequency against volume. Both hit 100% posterior probability for strength, both with diminishing returns, and the abstract compares their magnitudes nowhere.
It does not say frequency is irrelevant. The hypertrophy-frequency result is an absence of an identifiable effect in this dataset, not evidence of absence.
It does not establish that the pooled studies were randomised controlled trials — the word "randomised" appears nowhere in the abstract and the record carries no RCT publication type.
Finally, the sample is 79.1% male with an average age of 25.16 years, so it licenses no claims about older adults or about women specifically. A separate network meta-analysis in adults 60+ (Radaelli et al., PMID 39405023) reached a different practical conclusion for that population — a different population, not a contradiction, but an informed commenter will raise it within an hour if the card is worded as a universal rule.
Every figure below is shown with the sentence it was taken from, so you can check it against the paper rather than trusting us.
- 67studies
Number of studies pooled in the meta-regressions
A series of multi-level meta-regressions were performed for muscle hypertrophy and strength, utilizing 67 total studies of 2058 participants (79.1% male, 20.9% female; average age 25.16 ± 5.22 years).
- 2058participants
Total people across all pooled studies
A series of multi-level meta-regressions were performed for muscle hypertrophy and strength, utilizing 67 total studies of 2058 participants (79.1% male, 20.9% female; average age 25.16 ± 5.22 years).
- 79.1%percent of participants
Share of the pooled sample that was male
A series of multi-level meta-regressions were performed for muscle hypertrophy and strength, utilizing 67 total studies of 2058 participants (79.1% male, 20.9% female; average age 25.16 ± 5.22 years).
- 20.9%percent of participants
Share of the pooled sample that was female
A series of multi-level meta-regressions were performed for muscle hypertrophy and strength, utilizing 67 total studies of 2058 participants (79.1% male, 20.9% female; average age 25.16 ± 5.22 years).
- 25.16 ± 5.22years
Average age of the pooled sample, with its standard deviation. Carried across exactly as printed — not rounded, not restated as a range.
A series of multi-level meta-regressions were performed for muscle hypertrophy and strength, utilizing 67 total studies of 2058 participants (79.1% male, 20.9% female; average age 25.16 ± 5.22 years).
- 0.5sets counted per indirect set
The weight assigned to an indirect set under the 'fractional' counting method — one of three candidate quantifications the authors defined and tested (1, 0.5, 0). This token covers the definition only. That the fractional method had the strongest relative evidence is a separate finding in a separate sentence and must not be presented as part of this quotation.
Then, weekly set volume/frequency for indirect sets was quantified as 1 for 'total,' 0.5 for 'fractional,' and 0 for 'direct.'
- 100%posterior probability that the marginal slope exceeds zero
Bayesian probability that the volume slope is above zero for both hypertrophy and strength. This is directional certainty, not an effect size — the abstract reports no effect size or interval anywhere, so this must never be captioned as how much volume delivers.
The posterior probability of the marginal slope exceeding zero for the effect of volume on both hypertrophy and strength was 100%, indicating that gains in muscle size and strength increase as volume increases.
- 100%posterior probability that the marginal slope exceeds zero
Bayesian probability that strength gains rise with weekly training frequency — contrasted in the abstract with the frequency-on-hypertrophy model, which was less than 100%. CAUTION: this sentence opens with 'In contrast' and inherits its subject (frequency) from the preceding sentence. Standalone on a card a reader cannot tell which variable it refers to. Pair it with the hypertrophy-frequency sentence or rewrite the caption.
In contrast, the posterior probability for strength was 100%, suggesting strength gains increase with increasing frequency, albeit with diminishing returns.
2058 participants across 67 studies, 79.1% male and 20.9% female, average age 25.16 ± 5.22 years. Training status is not broken out in the abstract; it appears only as a model covariate ("All models were adjusted for the duration of the intervention and training status"). The abstract does not state that the pooled studies were randomised. This is a young, overwhelmingly male sample — not older adults, not a predominantly female sample.
Joshua C Pelland, et al. (2026)The Resistance Training Dose Response: Meta-Regressions Exploring the Effects of Weekly Volume and Frequency on Muscle Hypertrophy and Strength Gains.
Sports Medicine (Auckland, N.Z.) · PMID 41343037
Meta-regression · 67 total studies · 2058 participants (79.1% male, 20.9% female; average age 25.16 ± 5.22 years)
Who funded itAuthors declare no conflicts of interest.
Volume, rep ranges, rest, stretching, sleep — each with its numbers and its limits.
Every figure is machine-checked against the sentence it came from. Including what that does not guarantee.
Repstra handles the programming, check-ins and admin so the conversation you have with a client is the coaching one. Free for your first 5 clients, 0% commission on your own.
See how it works