A review of 13 studies and 1,016 people compared machines and free weights. In the five studies that measured muscle growth, the gap was too small to tell apart from chance.
Where muscle growth was actually measured, the hierarchy had no gap to point to.
Every gym has a ranking nobody posted on the wall. Barbells at the top. Dumbbells in the middle. Machines at the bottom. The hierarchy doesn't come with a source or a study. It comes with something more persuasive: a social code that rewards visible effort over invisible outcomes.
Chalk your hands, load your plates, ask for a spot, and you're a serious lifter. Sit down at the leg press, and something about your status shifts.
You know this ranking. You've followed it. You've felt that quiet pull when you chose the chest press instead of loading the bench, or walked toward the leg press when the squat rack was open. Most lifters have. The hierarchy doesn't just rank equipment. It ranks the people who use it.
Haugen and colleagues wanted to know whether any of that is earned. They pooled 13 studies covering 1,016 people (789 men, 219 women, ages 18 to 60, trained and untrained) that put free-weight training against machine training and measured strength, jump height and muscle growth. Five of those studies measured muscle growth, and they answer the question this page is about: does training with free weights produce more muscle growth than training with machines?
The equipment hierarchy every gym teaches has no muscle-growth result behind it. Whether the data holds for longer training periods and more diverse populations is the open question the evidence has not yet closed.
- Machines and free weights produced similar muscle growth in the five of 13 studies that measured it, with no significant difference between them.
- For upper-body strength, with each group tested on the equipment it trained with, a sub-analysis found machines actually outperformed free weights, though this exploratory finding needs more research to confirm.
- Training experience made no difference to the strength results: tested on the equipment they trained with, neither strength-trained nor untrained lifters showed a significant gap. The review did not split muscle growth by experience.
- For muscle growth, equipment choice is a preference, not a performance decision. Consistency with whatever you enjoy matters more than which equipment category you pick.
The Total Evidence for the Hierarchy
The difference was 0.055 standard deviations, a hair toward free weights. That was the whole gap in muscle growth between free weights and machines across the five studies that measured it, and it was not statistically significant (p = 0.751): a gap at least that large would turn up often by chance even if the two worked the same. The five studies also varied no more than chance would predict.
To put that number in context: on the scale the authors use, anything up to 0.2 standard deviations counts as small, and 0.055 is about a quarter of that. Where muscle growth was actually measured, the hierarchy had no gap to point to. The ranking that made you feel like a second-tier lifter for choosing the leg press found no support in those studies.
This is not one study's conclusion, but it is not thirteen studies' either. Five studies, with different populations, different exercises and different measurement tools (ultrasound, BodPod, circumferences, skinfolds), went into the muscle result. Pooled, they found machines and free weights produced similar muscle growth, the authors' own word.
Where the Hierarchy Might Run Backwards
If the story stopped at parity, it would be useful but quiet. Many lifters already suspected machines were "good enough." The data went further.
The review also compared strength gains with each group tested on the equipment it trained with: machine trainees on machines, free-weight trainees on free weights. Overall, that showed no significant difference. But in a sub-analysis of upper-body exercises, machine training produced significantly greater strength gains than free-weight training did on its own equipment. Not equal. Better. For the measurement where the hierarchy promised barbells would dominate, the equipment gym culture calls training wheels came out ahead.
That finding has honest limits. The overall comparison between upper and lower body was not statistically significant. This is an exploratory sub-analysis, not a settled conclusion. And strength runs both ways: when both groups were tested on free weights, the free-weight trainees gained more (p = 0.023), though that result stopped being significant when one large study (Lennon 2010) was left out.
But the data shows a direction: for upper-body strength measured on the equipment each group trained with, the numbers lean toward machines, not away from them. For that one measurement, the hierarchy might be pointing the wrong way.
The meta-analysis found machines and free weights build similar total muscle. But total muscle is not the whole story.
The researchers pointed to other studies suggesting that different exercises grow different parts of the same muscle: the barbell squat did not seem to grow the rectus femoris, one of the quad muscles, while the leg extension did. Two programmes could produce similar total growth while building different regions of a muscle. The study's authors speculated that combining both modalities could result in more complete muscle growth, because exercises differ in their biomechanics, range of motion and resistance curve.
This is the practical insight hiding behind the headline: using both machines and free weights is not a compromise. It may be the strategy that covers the most ground. The complete evidence — including exactly which exercises grow which regions and which training variables are worth zero — runs through 2,193 participants across six independent findings.
What the Data Cannot Tell You Yet
The parity finding is consistent. It is also limited in ways the researchers themselves named.
Only 5 of the 13 studies measured muscle growth, using different tools: ultrasound, BodPod, circumferences, skinfolds, which the authors say vary greatly in how well they detect change. Study durations averaged 8.9 weeks, with a range of 6 to 12. The researchers wrote that "one should be careful with the interpretations of the results, especially for muscle hypertrophy."
Regional muscle-growth differences might exist even when total hypertrophy is equal. The authors point to other studies in which the barbell squat did not seem to grow the rectus femoris, one of the quad muscles, while the leg extension did. Free weights may also transfer slightly better to jumping: the jump results leaned toward free weights but were not statistically significant.
These are real limits, not disclaimers. The five muscle studies did not disagree beyond chance, but the confidence interval still leaves room for a small-to-medium difference either way, and the authors call the current evidence tentative. What the data does show is that the hierarchy, as gym culture teaches it, has no muscle-growth result behind it: the studies that measured growth found no muscle advantage for free weights.
One of the five shows what the pooled number looks like up close. In Schwanbeck and colleagues' 2020 trial, 46 men and women were randomized and 36 finished eight weeks, one group exclusively on machines and the other exclusively with free weights. Biceps and quadriceps muscle thickness increased in both groups with no difference between them [1]. It is part of the pooled result, not a separate confirmation of it.
Not Even for Experienced Lifters
The hierarchy has one remaining defense: "Machines are fine for beginners, but serious lifters need free weights." Haugen's data tested that defense for strength, not for muscle: the review did not split its muscle result by training experience.
In the strength comparison where each group was tested on the equipment it trained with, the studies of strength-trained lifters showed a difference of 0.0005 between machine and free-weight training. The untrained subgroup showed a slightly larger but still non-significant effect (0.150). The comparison between subgroups was not significant either. "Trained" here means lifters who cleared a strength threshold relative to body mass, not a number of years under the bar.
Tested on the equipment they trained with, the stronger lifters in the dataset showed no sign of a machine penalty. The defense that machines are acceptable for newcomers but inferior for advanced trainees found no support in that strength data. One caution: when a single study (Augustsson 1998) was taken out, the trained result tipped toward machines, so the number is less fixed than its four decimal places suggest.
The studies that measured it found no muscle you gave up by using machines. What you lost was the peace of mind that should have been there all along.
What the Guilt Actually Cost
The data changes something more personal than a training program. It changes the feeling that went with it.
Every session where you chose the leg press and glanced at the squat rack, every chest day where you sat at the machine press and felt like you were cutting corners, every time someone told you barbells build "real" muscle and you half-believed it: none of that guilt had evidence beneath it.
The studies that measured it found no muscle you gave up by using machines. What you lost was the peace of mind that should have been there all along.
For muscle growth, equipment choice is a preference, not a performance decision. The machine section is not a lesser section: in the studies that measured it, it produced similar muscle growth. For strength on a particular lift, the authors' advice is specificity: train the exercise you want to be stronger in.
Train on whatever you enjoy, whatever is available, whatever your body responds to on any given day. The hierarchy said that choice mattered for muscle. The studies that measured growth found it did not.
For total muscle, the equipment question has a working answer, which the authors call tentative. But it opens a different one. If machines and free weights build similar total muscle, does the specific exercise matter? Does a leg press and a leg extension grow the same part of the quad, or do different movements build different regions of the same muscle?
For muscle growth, equipment choice is a preference, not a performance decision. In the five studies that measured it, machines and free weights showed no significant difference.
When machines make life easier: crowded gyms where the squat rack has a queue, sessions where you want faster transitions between exercises, and days when a fixed movement path feels better than balancing a barbell.
When free weights make sense: you enjoy compound movements, you train at home with limited equipment, you want the coordination challenge, or you want to get stronger on the free-weight lifts themselves, since strength gains follow the equipment you train on.
The variable that actually drives results is consistency. The equipment you enjoy enough to show up for four times a week builds more muscle over a year than the "optimal" equipment you dread and use twice.
What other research found
What this means for you
In the review's strength comparison, with each group tested on the equipment it trained with, the studies of strength-trained lifters showed a difference of 0.0005 between machine and free-weight training. The untrained showed no significant gap either.
The script that says machines are fine for beginners but insufficient for serious lifters found no support in that strength data. The review did not split muscle growth by training experience, but across the five studies that measured growth there was no significant difference. The leg press is not a step down from the squat rack. For strength on a particular lift, though, the authors' advice is to train that lift.
The review did not split its muscle result by experience, but across the five studies that measured growth there was no significant difference between machines and free weights. In the strength comparison, the untrained subgroup showed no significant difference either.
Machines have one practical advantage for newer lifters: they reduce the stability demand. A chest press guides the weight through a fixed path, so the muscles do the work instead of your coordination. That means more focus on the muscle you're trying to grow and less energy spent keeping the bar straight.
The evidence found machines build similar muscle to free weights. For someone building their first training habit, the simplicity may make it easier to show up consistently.
A sub-analysis of upper-body exercises, with each group tested on the equipment it trained with, found something the overall data hides: machine training produced significantly greater strength gains than free-weight training.
This is an exploratory finding from a sub-analysis, not a settled conclusion. The overall comparison between upper and lower body was not statistically significant. But the direction is clear: for upper-body pressing strength on the equipment you train with, the data leans toward machines, not away from them.
If the lift you care about is a machine press, the chest press may deserve a lead role, not a supporting one. If it is the barbell bench press, train the bench: the authors advise training the exercise you want to be stronger in, and when both groups were tested on free-weight lifts, free-weight training gained more.
If machines and free weights build similar muscle, the time-efficient option becomes the productive option.
Machines eliminate plate loading and unloading. No spotter needed. No waiting for the one squat rack. Transitions between exercises take seconds instead of minutes. A 30-minute machine circuit covers more total work than a 30-minute free-weight session where 10 minutes went to setup.
The parity finding turns time-efficiency into a genuine training advantage, not a lazy shortcut.
Before you change anything
Adults aged 18 to 60, both men and women, with and without prior training experience, free from chronic disease or injury. Recreational gym-goers seeking general muscle growth and strength.
Not studied: competitive athletes (none of the 13 studies included them), adults over 60, anyone under 18, and individuals with chronic disease or injury. Clinical rehabilitation populations were excluded.
One caution: 77% of the total participants were male. The meta-analysis may underrepresent female-specific responses, though Schwanbeck 2020, one of the included trials and majority-female at completion, found no difference in muscle thickness between the groups and no difference between men's and women's responses.
Only 5 of 13 studies measured muscle growth. The core muscle-growth finding rests on fewer studies than the full 13-study count implies.
Measurement methods varied across studies: ultrasound, BodPod, circumference tape, and skinfolds. These tools have different precision levels, which adds noise to the pooled result.
Study durations ranged from 6 to 12 weeks (average 8.9). Longer-term adaptations beyond 12 weeks remain unknown.
The free-weight strength advantage (F2) was fragile: removing a single study (Lennon 2010) shifted the p-value from 0.023 to 0.153. Any conclusion about free weights being stronger for free-weight exercises depends heavily on one dataset.
Regional muscle-growth differences between modalities are possible even when total muscle growth is equal. A leg press and a squat could grow different parts of the same muscle.
The muscle-growth finding carries moderate confidence. It rests on 5 of the 13 studies, measured with tools of differing precision. Those five varied no more than chance would predict (I² = 0), but the confidence interval (-0.397 to 0.287) still leaves room for a small-to-medium difference either way, and the authors call the evidence tentative.
The upper-body reversal carries low confidence. It comes from a sub-analysis of the comparison in which each group was tested on the equipment it trained with. The overall comparison between upper and lower body was not statistically significant (p = 0.104), and the parent analysis had significant heterogeneity. The direction is suggestive, not settled.
The training-status finding is about strength, not muscle. In that same comparison, the strength-trained subgroup showed an effect size of 0.0005 and the untrained 0.150, neither significant, but subgroup comparisons have lower statistical power than the main analysis, and leaving out one study (Augustsson 1998) tipped the trained result toward machines. The review did not split muscle growth by training status.
For total muscle, machines and free weights showed similar growth in the studies that measured it. But that finding opens something more specific: if a leg press and a leg extension both train the quad, do they build the same part of it?
Burke's team put one leg on the leg press and the other on the leg extension in the same person and measured where the quad actually grew. The answer reshapes how exercise selection works.
What This Study Found
All findings from this paper, in plain language.
- Machines and free weights produced similar muscle growth in the five studies that measured it, with a difference too small to tell apart from chance.
- People who trained with free weights got slightly stronger at free-weight exercises, but this advantage was fragile and depended on a single study.
- When each group was tested on the equipment it trained with, machine and free-weight training produced similar strength gains, with no significant difference.
- When strength was measured on neutral tests (not favoring either equipment type), there was no difference between groups.
- Jump height showed no significant difference between machine and free-weight training, though the results leaned toward free weights.
- Training experience did not change the strength results: tested on the equipment they trained with, neither strength-trained nor untrained lifters showed a significant difference between machines and free weights.
- For upper-body strength, with each group tested on the equipment it trained with, machine training significantly outperformed free-weight training in a sub-analysis, though the test of upper against lower body was not statistically significant.
- The researchers speculated that combining both machines and free weights could result in more complete muscle growth across different regions of the same muscle.
- The five muscle-growth studies varied no more than chance would predict (I² = 0), which with so few studies may also reflect low power; only the direct strength comparison varied significantly between studies.
- The included studies were rated as fair to good quality on a standardized assessment, with five scoring fair and eight scoring good.