'Non-responders' to weight training may be a statistical illusion

Gym culture is full of talk about responders and non-responders: the idea that identical training produces large gains in some people and almost nothing in others because of who they are. That belief shapes how people interpret a disappointing few months and whether they abandon a programme. Yet the visible spread of results in any trial mixes several things together, including measurement error, day to day fluctuation and behaviour outside the gym. This analysis asked whether genuine person to person response differences to resistance training can actually be detected in older adults carrying extra weight.

The authors reused data from a previous meta-analysis covering 15 randomised controlled trials of resistance training lasting at least eight weeks. In total 587 men and women aged 60 and over were represented, 333 assigned to resistance training and 254 to control. Outcomes were body weight and body composition, specifically percent body fat, fat mass, body mass index and lean body mass. Rather than only comparing group averages, they used the change scores' standard deviations in the training and control groups from each study to estimate true inter-individual response differences, pooling both these estimates and conventional comparisons.

Resistance training produced statistically significant improvements in body weight and in every body composition outcome measured. However, no statistically significant inter-individual response difference was found for any outcome, and all the 95% prediction intervals overlapped. In other words, the scatter of individual results in the training groups was no greater than the scatter seen without training. The authors conclude that factors other than variation in training response, including random variation and physiological changes tied to behavioural shifts not caused by the training itself, account for the differences observed between people.

This complicates the popular framing of responders and non-responders by showing that, at least in this population and with these outcomes, the apparent variation does not survive statistical scrutiny. It does not contradict the well established finding that resistance training improves body composition in older adults; it strengthens the case that the average benefit is the fair expectation for most people. Open questions remain: the analysis covers adults aged 60 and over with overweight and obesity, over training periods of at least eight weeks, so younger or leaner people and much longer programmes are untested. Whether the same holds for strength or muscle size outcomes is also not addressed here.

The practical read is that a modest or slow change in body composition after starting resistance training is not strong evidence that your body is unusual. Measurement noise, weight fluctuation and everything happening around the training plausibly explain more of the difference between two people than their innate response does. That is a reason to judge a programme on whether it is being done consistently and long enough, rather than concluding early that it does not work for you. It also means the average result reported in trials is a reasonable expectation rather than an average of dramatic winners and losers.