Group Training Culture

Group training and exercise adherence: what keeps members coming back

Exercise adherence is how well people keep to a training programme over time, usually measured as the share of planned sessions they attend. Training in a group does not, on its own, make people more consistent than a well-supported individual programme: in the largest recent meta-analysis, the average difference was small and not statistically significant. The advantage appears when a class works as a real group, with members who know each other, shared goals and time to interact, and when sessions happen at the same times each week. Most of this evidence comes from adults over 50 and from programmes lasting a few months.

How studies measure adherence

Trials usually count attendance: the number of sessions attended divided by the number offered. Some ask participants to report their weekly activity instead. Attendance shows that someone came, not how hard or how long they trained, and self-reports tend to overestimate activity. When this article says a group "adhered better", it means its members attended a larger share of sessions.

Effect sizes in these studies are reported as standardised mean differences (Cohen's d or Hedges' g): the gap between two groups' averages divided by the standard deviation of individual results. A value of 0.2 is conventionally called small and 0.5 moderate.

Group versus individual programmes

A 2026 meta-analysis by Kritz and colleagues in Nature Human Behaviour pooled 71 studies that compared group-based and individual physical activity programmes, in person and online. For physical activity outcomes, a category that included both attendance and self-reported activity, group programmes showed a small advantage that was not statistically significant (g = 0.086, about 22,000 participants). A formal equivalence test indicated that the two formats have similar effects on physical activity. Group programmes did better on functional outcomes such as strength and balance, but only after outliers were removed.

The authors also sorted groups by type. A true group is one in which the leader deliberately uses group dynamics principles, such as group goals, a shared identity and structured interaction between members. A standard class puts people in the same room without those strategies. Only 14 of the 71 studies used true groups, and they were linked to slightly larger effects on health outcomes, not on activity itself. An earlier 2006 meta-analysis by Burke and colleagues, summarised in the 2026 paper, had found a trend in the same direction: true groups did better than standard classes, and standard classes did no better than home exercise with regular contact from a professional.

More than half the studies targeted people aged 50 or older, heterogeneity between studies was high, and the authors rated the certainty of evidence from randomised trials as moderate. A group format is therefore not automatically better; what happens inside the group appears to matter more.

What makes a class work as a group

Group cohesion is the tendency of a group to stick together in pursuit of its goals. Researchers distinguish task cohesion, the commitment to the group's shared training goals, from social cohesion, the liking and connection between members.

A 2015 realist review by Harden and colleagues examined 52 group-dynamics-based interventions for adults. Of the 39 that outperformed an individual comparison, the most frequent strategies were:

Strategy Interventions using it (of 39)
Interaction and communication between members 25
Social support 22
Group goals 12
Distinctiveness or team identity 12
Competition 10

Most interventions combined several strategies, about seven on average, so the review cannot say which one does the work. Only four of the 39 tested whether cohesion explained the results. In one programme for women of colour in the United States, cohesion did mediate the effect on attendance. The review included no youth programmes, did not weight studies by size and noted likely publication bias.

A trial that changed the group, not the workout

The GOAL randomised controlled trial by Beauchamp and colleagues (2018) tested the group-composition idea directly. It assigned 627 adults aged 65 and over (mean age 72, 71% women) at YMCA centres in Vancouver to one of three class types, each meeting three times a week:

  1. Similar age, same gender, with group-identity strategies.
  2. Similar age, mixed gender, with the same strategies.
  3. Standard mixed-age, mixed-gender classes already running at the centres.

According to the trial protocol, the two intervention arms also used peer instructors of a similar age, matching T-shirts and organised time to socialise after class. Over 12 and 24 weeks, both intervention arms attended more sessions than the standard classes. The effect was moderate in the similar-age, mixed-gender arm (d = 0.51 at 12 weeks, 0.47 at 24 weeks) and small in the same-gender arm (d = 0.28 and 0.29). The two intervention arms did not differ significantly, so separating classes by gender added nothing.

Because composition and the social strategies were changed together, the trial cannot show which element mattered. Its participants were also older adults, and the result may not transfer to a 30-year-old training for a hybrid race.

Habit and the value of fixed times

A habit in this research means a behaviour that starts automatically in response to a familiar cue, such as a time of day or a place, without a deliberate decision each time. A 2015 study by Kaushal and Rhodes followed 111 new gym members (mean age 48, 70% women) for 12 weeks with questionnaires. Members who exercised at least four times a week for about six weeks reached a stable habit score; members who trained less often did not maintain one. Over time, consistency, exercising at the same time each day, was the strongest predictor of habit, followed by keeping the exercise simple, the training environment and enjoyment.

Both habit and training frequency were self-reported, and the four-sessions figure came from one sample of members, so it is a useful reference point rather than a threshold that applies to everyone.

Illustration: a six-week starter block

The following example combines these findings into a class plan for new members. It is an illustration, not a tested protocol.

Element Example choice Evidence behind it
Fixed slots Monday, Wednesday and Friday at 18:00, plus one open session Consistency predicted habit
Stable cohort The same 8–12 people for six weeks Interaction and social support
Group goal Total sessions attended by the cohort Group goals
Simple content A repeatable session structure each week Low complexity predicted habit

With 10 members and four sessions a week for six weeks, the cohort can attend at most 240 sessions. A goal of 200 corresponds to about 83% attendance and gives the group a shared number to track.

On a live heart-rate board such as the Pulz8 TV dashboard, the shared target should be attendance or time in each person's planned zone, not everyone showing the same colour, because the same effort produces different heart rates in different people.

Key points

  1. Treat the group format as a tool, not a guarantee, since standard classes and individual programmes produce similar activity levels on average.
  2. Build interaction into every class, such as partner work, using names and time to talk before or after the session.
  3. Give the group a shared goal, such as total attendance over a block, alongside individual targets.
  4. Keep class times fixed for new members, because exercising at a consistent time was the strongest predictor of habit.
  5. Keep the first weeks simple and repeatable, and check attendance after about six weeks.

Sources


Heart rate is a training signal, not medical advice. The studies above describe group averages, and a programme for someone with a health condition should be agreed with their doctor.

See your own zones live. Pulz8 is free for individuals.

← All articles Start free