By Quanthlete
A systematic review and meta-analysis, published in BMC Sports Science, Medicine and Rehabilitation, investigates the effectiveness of free-weight versus machine-based strength training on maximal strength, muscle hypertrophy, and jump performance. Conducted by Haugen and colleagues, this study compiles data from 13 studies, with a sample size of 1016 participants, aiming to provide evidence-based insights for trainers, athletes, and fitness enthusiasts on which method yields optimal training outcomes.
Background
Resistance training variables like volume, intensity, frequency, and exercise selection are well-established drivers of strength and hypertrophy. Yet the role of exercise stability—specifically, the trade-offs between free weights and machines—remains understudied. Free weights (barbells, dumbbells) demand greater stabilization and coordination, potentially engaging more synergist muscles, while machines reduce stability requirements, enabling heavier loads via fixed movement paths.
Proponents argue free weights offer superior functional transfer and muscle activation, though evidence linking this to hypertrophy remains inconclusive. Conversely, machines may enhance strength through higher mechanical tension but show no clear hypertrophy advantage over free weights when training close to failure. While machines are often considered safer due to lower skill requirements, the evidence on injury risk is inconclusive, with most reported free-weight injuries resulting from accidents, not execution.
Meta-analytic research shows that strength gains are modality-specific (free-weight strength improves most with free weights, machine strength with machines), with no significant differences in overall hypertrophy or power. However, limitations in existing research—such as variability in training experience and under explored interactions between stability and load—highlight the need for updated, rigorous comparisons.
Methods
This systematic review and meta-analysis followed PRISMA guidelines and was preregistered in PROSPERO (CRD42021270740). The literature search was conducted across MEDLINE, Embase, and SPORTDiscus up to January 1, 2023, using a combination of keywords related to resistance training, free weights, machines, and outcomes like strength, power, and hypertrophy. Secondary searches included citation tracking and reviewing personal libraries.
Studies were selected based on predefined inclusion criteria: experimental design, peer-reviewed English publication, adult participants (18–60 years), minimum 6-week intervention comparing free weights and machines, and outcomes on strength, hypertrophy, or jump performance. Machines/equipment without a fixed-path resistance were excluded.
Training status was classified based on strength thresholds relative to body weight, distinguishing trained from untrained individuals. Methodological quality of studies was assessed using the 12-point TESTEX scale, covering study design and reporting. Independent blinded ratings were done by two authors, with disagreements resolved by a third and verified for consistency using Kappa analysis and consensus discussions.
Exercise Comparison
In the analysis of maximal dynamic strength, comparisons focused on exercises that were most frequently examined across studies. For lower-body strength, the free-weight squat (or lunges in one case) was compared with machine-based exercises including the leg press, hack squat, or Smith machine squat. Other lower-body machine exercises, such as knee extensions, were excluded.
For upper-body strength, the free-weight bench press was compared with machine equivalents such as the Smith machine bench press, lying chest press, or seated chest press. In studies where both upper- and lower-body exercises were tested, results from the two were averaged into a single standardized effect size to maintain statistical validity. Adjustments were made to align variations in exercise format—for example, substituting seated machine press data where lying versions were unavailable. This standardized approach ensured comparability between modalities while minimizing potential biases in effect size calculations.
Isometric strength was assessed by measuring lower-body force output in Newtons. Most studies included in the analysis tested under stable conditions to ensure comparability. For consistency, only the most stable test conditions were used. In cases where exact data were not reported, numerical values were extracted from graphs using digital tools. Additionally, one study used the sum of isometric strength from both legs to represent performance. These standardized test conditions allowed for pooled analysis of isometric strength across studies, with pre-post differences calculated using a consistent correlation coefficient and adjusted where necessary through sensitivity analyses.
To assess the effectiveness of each training modality in isolation, an additional analysis was conducted comparing strength changes within the same training groups. Specifically, the free-weight training group’s strength gains were evaluated using free-weight exercises, while the machine training group’s strength improvements were assessed using machine-based exercises. Because the absolute loads and movement mechanics differ between modalities (e.g., a 35 kg gain in squats vs. 66 kg in leg press may reflect similar relative improvements), the analysis used standardized mean differences (SMDs) rather than raw values. These pre-post SMDs served as normalized effect sizes, which were then used to create a “synthetic effect size” by subtracting the machine group’s SMD from the free-weight group’s SMD, combining variances from both. This approach allowed for a direct, modality-specific strength comparison without conflating differing load scales. Sub-analyses were also conducted for upper vs. lower body exercises and by training status (trained vs. untrained). Additionally, within-group meta-analyses determined whether each group significantly improved in maximal dynamic and isometric strength, jump performance, and hypertrophy from baseline.
Key Findings
1. Maximal Strength:
- Free-Weight Exercises: In tests measuring maximal strength using free-weight exercises (bench press, squat or lunges), participants who trained with free-weights showed significantly greater gains compared to those who used machines. The standardized mean difference (SMD) was -0.210 (95% CI: -0.391 to -0.029, PI: -0.484 to 0.064, p-value = 0.023).
- Explanation: The negative SMD indicates that free-weight training had a stronger impact on strength gains when tested using free-weight exercises. The relatively low p-value (< 0.05) supports that this result is statistically significant, suggesting that free-weights allow for more strength adaptations, likely due to the activation of stabilizing muscles and engagement of multiple muscle groups. This supports the principle of specificity, where training in the same mode as testing yields better outcomes.
- Machine Exercises: For machine-based strength tests, machine training showed a slight advantage, with an SMD of 0.291 (95% CI: -0.017 to 0.600, PI: -0.147 to 0.729, p = 0.064). Although this result suggests that machine-based training was more effective when tested on machine exercises, the p-value (> 0.05) indicates that this difference was not statistically significant.
- Explanation: The findings highlight that, although machine training might improve strength on machine-based tests, it may not translate well to free-weight tasks. This outcome further supports that specific adaptations occur in the mode of training. While machine exercises may not demand the stabilization required in free-weight exercises, they allow for isolated muscle focus, which may explain the improvement on machine-based tests.
2. Hypertrophy:
- The analysis found no significant difference in muscle hypertrophy between free-weight and machine-based training. The SMD was -0.055 (95% CI: -0.397 to 0.287, PI: -0.611 to 0.500, p = 0.751).
- Explanation: With an SMD near zero and a high p-value (p > 0.05), the findings suggest that both training modalities are equally effective in promoting muscle growth. This is likely because hypertrophy depends on factors like load, volume, and time under tension rather than the type of equipment. Free-weights activate stabilizing muscles, potentially benefiting overall muscle balance, while machines may allow for a more focused activation of target muscles. For hypertrophy goals, combining both free-weight and machine exercises could optimize results by engaging both primary and stabilizing muscle groups.
3. Jump Performance:
- Jump performance, assessed through countermovement jump (CMJ) height, showed no significant advantage for either training modality. The SMD was -0.209 (95% CI: -0.597 to 0.179, PI: -1.208 to 0.790, p = 0.290).
- Explanation: Jump performance is often used as a measure of explosive lower-body strength, particularly useful in sports performance contexts. Since neither free-weight nor machine-based training directly targets explosive power, it’s unsurprising that both methods showed similar results. The wide prediction interval suggests high variability, meaning some individuals may benefit more from one type of training than another, but overall, neither is superior. To specifically improve jump performance, incorporating plyometric or explosive training may be more effective than solely relying on traditional resistance training.
Practical Applications and Recommendations
The results highlight that the principle of specificity applies to strength training modality: to gain maximal strength in a specific exercise, one should train using that exercise. Free-weights are recommended for athletes who seek functional strength gains or are preparing for sport-specific movements, while machines may benefit beginners seeking controlled, isolated strength gains without high injury risk.
Conclusion
The study suggests that neither free-weight nor machine-based training holds a universal advantage across all outcomes. Trainers and athletes should tailor their approach based on individual goals, with the understanding that each method offers unique benefits suited to different training needs.
References:
- Haugen, M. E., Vårvik, F. T., Larsen, S., Haugen, A. S., van den Tillaar, R., & Bjørnsen, T. (2023). Effect of free-weight vs. machine-based strength training on maximal strength, hypertrophy and jump performance – a systematic review and meta-analysis. BMC Sports Science, Medicine and Rehabilitation, 15(103).