Exercise Plus Volunteering: What a 2,856-Person GoodGym Study Found, and How Big the Effect Really Is

Who was studied, and how

The study, "Getting fit by doing good: Exercise, volunteering, and mental health at scale," was written by Maria Cotofan of King's College London and Christian Krekel and Ekaterina Oparina of LSE's Centre for Economic Performance. It appeared online in Social Science & Medicine on September 4, 2026, and LSE publicized it later that month in an announcement summarizing the headline results.

GoodGym itself is the interesting ingredient. Members walk, run or cycle to a task, then do it: sorting cans at a food bank, planting trees in a park, clearing a garden for an isolated older person, or paying a weekly social visit. According to the published paper, the charity was founded in 2009, has more than 25,000 members across 64 local areas, is open to anyone 18 or older, is free for both volunteers and the people they help, and has completed more than 383,000 tasks for over 58,000 people.

Krekel framed the appeal in LSE's release with a question that will make plenty of gym-goers smile: why spend time in a gym when you could exercise alongside neighbors while helping your community? It is a fair point, and it is also where the study's own framing ends and its data begin.

If the gym-versus-alternatives question is on your mind, our breakdown of home gym vs. membership costs over five years puts real numbers on the conventional options.

The researchers sent three online surveys to every registered GoodGym member, in February 2023, October 2023 and March 2024, and entered respondents into a prize draw as a thank-you. They then linked each answer to GoodGym's own records of who actually showed up for tasks, rather than relying on people's memories. After removing incomplete records, the analysis covers 3,590 survey responses from 2,856 people.

The headline results

The comparison is between two groups who both signed up: members who completed at least one task, and members who registered but never took part. Here is what the participants reported, on average, compared with the non-participants:

• Less psychological distress. Mental health was measured with the GHQ-12, a 12-item questionnaire about recent strain, such as losing sleep over worry. Participants scored 0.23 standard deviations better in the most cautious version of the analysis.
• Higher life satisfaction. On a 0-to-10 scale answering "Overall, how satisfied are you with your life nowadays?", participants scored 0.43 points higher on average.
• More connection. Participants reported less loneliness and stronger feelings of belonging, connectedness and trust in people in their local area, with differences between 0.12 and 0.27 standard deviations.
• A lingering association. Among people whose last task was more than six months earlier, life satisfaction was still 0.31 points higher.

Those four social measures together accounted for roughly 40% of the mental health association in an exploratory analysis. The authors suggest some of the rest may reflect the direct effects of exercise, which their surveys could not measure.

Standard deviations, translated

A standard deviation is a way of describing spread. If you lined up every survey respondent from least to most distressed, the standard deviation tells you how far apart typical scores sit. A difference of 0.23 standard deviations is modest by that yardstick. As a rough illustration (a Daily Vitality estimate that assumes a bell-shaped spread of scores), a person who would sit right in the middle of the non-participant group, at the 50th percentile, would land at about the 59th percentile if shifted by 0.23 standard deviations.

That is a real, noticeable shift at the level of a population. It is not a prediction of how any individual will feel, and it is not a clinical threshold. It is an average difference between two groups of people.

The number also depends on how much the researchers adjusted for. With only basic adjustments for region and survey timing, the gap was 0.31 standard deviations. It shrank to 0.29, then 0.24, then 0.23 as the researchers accounted for age, gender, ethnicity, marital status, education, work, income, children, and exercise or volunteering done elsewhere. That gentle shrinking matters: it suggests part of the raw gap reflects who joins and participates, not the activity itself, although most of it remained after those adjustments.

The two sets of benchmarks, reconciled

To make 0.43 points of life satisfaction feel concrete, LSE compared it with two differences that most people intuitively understand. Here is the part most coverage has missed: two different sets of percentages are circulating, and both are arithmetically correct because they describe two different estimates.

The researchers themselves call the 0.31-point figure the more conservative one, and its plausible range runs from 0.19 to 0.43. Either way you slice it, the association is meaningful on paper. On a 0-to-10 scale, though, it is still less than half a point, roughly the difference between rating your life a 7 and rating it a 7.4.

How the association varied with participation

The study also tracked how the mental health association shifted with how often people took part and how recently they had done so. The percentile column uses the same rough bell-curve illustration as above.

Two patterns stand out. The association tends to be larger among more frequent participants, though it levels off at the top, and even people with four or fewer tasks differed from non-participants. It is also strongest right after a session and fades over time, without disappearing completely by the six-month mark. For belonging, connectedness and trust, the associations became small beyond six months, while loneliness held up better.

Association, not cause and effect

This was not a randomized trial. Nobody was assigned to run to a food bank or to stay home. The researchers compared people who chose to participate with people who signed up but didn't, at a single point in time. Only five respondents switched from never participating to participating across the survey waves, too few to track change within the same person.

The authors took several steps to test whether self-selection was driving the result. They compared participants with a narrower control group of 709 people who didn't take part for reasons arguably outside their control, like a canceled session, bad weather or something unexpected coming up, and found similar results. Among people surveyed within 100 days of joining, the mental health difference was similar too, though less precise. Even so, the authors write plainly that they cannot rule out hidden differences, such as personality or recent life events, that might shape both who participates and how they feel. They note that bias could push the estimates in either direction.

That caution fits a long-running pattern in this field. A 2013 systematic review of volunteering and health found that cohort studies tended to link volunteering with better wellbeing and life satisfaction, but that experimental studies had not confirmed those findings.

We have seen this dynamic before on other topics, where tighter study designs tend to shrink headline effects, as in our look at why better studies often find smaller effects.

Who answered the survey

Each survey wave reached only about 5% to 7% of registered members. People who respond to a charity's research survey may be the most engaged, or the most motivated to share either a great or a disappointing experience. All the wellbeing measures were self-reported, too.

The sample also looks different from the general population. Compared with a nationally representative UK survey, GoodGym respondents were:

• Younger on average (43 vs. 47 years)
• More likely to be women (69% vs. 56%)
• Far more likely to hold a degree (63% vs. 21%)
• More likely to work full time (70% vs. 42%)

The researchers note their sample more closely resembles UK adults who already exercise and volunteer, and they caution against generalizing beyond that group. Their cost analysis, which estimates benefits at roughly 40 times GoodGym's spending of about £60 per member, comes with its own caveats: it is a partial, funder's-eye calculation, and it drops to about 7.8 times once volunteers' own time is counted at minimum wage. The paper also notes that GoodGym doesn't track how members were referred, so the study evaluates the activities themselves rather than "social prescribing," the practice of clinicians pointing people toward community activities, that LSE's announcement highlights.

The study split its results into three age groups, and the pattern is both encouraging and a reminder to read carefully. The mental health association was similar for people under 35 (0.21 standard deviations) and ages 35 to 54 (0.22). For people over 55, it was smaller at 0.16 standard deviations and only marginally significant, based on 608 responses. The researchers found no statistically meaningful difference between the age groups, but the oldest group's estimate is the least precise of the three.

The connection measures varied too. Feeling connected to people in the local area showed essentially no association for the over-55 group, while belonging and trust showed modest, marginally significant ones. Men showed a larger association with mental health than women, and women showed larger gains in belonging and trust. None of that tells you what an individual reader in their 40s, 50s or 60s would experience, but it does suggest the strongest evidence in this study comes from working-age adults.

For a different angle on how the type and mix of movement relate to long-term health, our coverage of a Harvard study linking exercise variety to longevity offers useful context.

How many Americans volunteer, by generation

The GoodGym model rests on two habits, and the US has solid data on both. On the service side, Census Bureau and AmeriCorps data show that 28.3% of Americans 16 and older, about 75.7 million people, formally volunteered through an organization between September 2022 and September 2023, while 54.2% informally helped neighbors with things like errands or lending tools. Time commitments are shrinking, though: average hours per volunteer fell from 96.5 in 2017 to 70 in 2023, which fits a broader shift toward short, episodic volunteering, the very kind of low-commitment task GoodGym is built around.

Volunteering rates also vary by generation, and the 40-plus crowd leads on several fronts. These figures come from AmeriCorps' 2023 survey data as compiled by USAFacts:

Generation X has posted the highest formal volunteering rate in each of the last four waves of the survey, and baby boomers lead in informal helping.

How many Americans meet the activity guidelines

On the movement side, the federal guidelines call for at least 150 minutes a week of moderate activity, like brisk walking, plus muscle-strengthening work on at least two days. According to the most recent CDC data brief on aerobic activity, 47.2% of US adults met the aerobic portion in 2024, a share that falls steadily with age, from 54.0% among adults 18 to 34 to 38.4% among those 65 and older. Meeting both parts is rarer still, with CDC's FastStats putting it at 24.2% of adults.

Among adults 65 and older, only 13.9% met both the aerobic and muscle-strengthening guidelines in 2022, according to a National Center for Health Statistics report on older adults.

The strength half of that recommendation is often the one that slips, which is why our report on the strength-training sweet spot for a longer life may be useful background.

GoodGym operates in the UK, but several familiar American formats already braid physical activity and helping others together. These are contextual parallels, not equivalent programs, and none of them was studied in the LSE paper.

• Weekly community 5Ks. parkrun USA hosts free, timed 5K events every weekend at 102 locations, and participants register once whether they plan to walk, jog, run, volunteer or mix it up.
• Trail crews. The National Park Service's volunteer program includes roles maintaining and rebuilding trails, available for a single day or year-round, alone or with family and friends.
• Home builds. Habitat for Humanity notes that volunteers don't need construction experience to help on a build site, just a willingness to learn and to get a little dirty.

Each of these shares at least two of GoodGym's three ingredients: physical effort, service to others and time with people. What none of them shares is GoodGym's specific design of moving to a good deed, often for an isolated older neighbor. Whether that particular combination matters, or whether any active, social, helpful routine produces similar associations, is a question this study can't answer.

Running as a social activity has its own momentum in the US, too, as our look at how run clubs are becoming the new dating apps shows.

The GoodGym study is a well-built, carefully hedged look at a genuinely appealing idea. Among nearly 3,000 members of one UK charity, people who ran, walked or cycled to help their communities reported less distress, more satisfaction with life and a stronger sense of belonging than people who signed up but never went. Sized honestly, the life-satisfaction difference is somewhere between about half and three-quarters of the gap between having a partner and being single, depending on which estimate you use, and it fades without vanishing after people stop.

What the study can't do is tell you that the activity caused those differences, or that the same pattern would hold for every age group, in every country, or for you personally. Its sample was self-selected, mostly working-age, and heavily degree-educated, and its strongest evidence comes from adults under 55. For American readers, the most solid takeaway may be the simplest one: the ingredients GoodGym combines, moving your body, doing something useful and doing it alongside other people, are already within reach in many communities, and the research on how much each one adds is still being written.