A patient who books an appointment on their worst day, an athlete scolded after a terrible game, a supplement started on a day of peak exhaustion… and then, as if by magic, things get better. Beware: a quiet statistical law — regression to the mean — explains a large share of these apparent « miracles. »
Introduction
An athlete has a disastrous game. The following week, their coach delivers a harsh dressing-down. At the next match, performance improves.
Apparent conclusion: the coach’s criticism worked.
Another situation. A patient consults at the very worst moment of their illness. A treatment is started. A few days later, they feel much better.
Apparent conclusion: the treatment was effective.
And yet, in both situations, another explanation is possible: what we are seeing is simply a consequence of regression to the mean. This quiet statistical law shapes our perception of the world every single day — usually without us noticing.
What is regression to the mean?
Regression to the mean describes a natural tendency:
When an observation is exceptionally high or exceptionally low, the next observation tends, statistically, to be closer to the average.
This property is present in every system that contains a part of randomness. Put differently: extreme events tend to be followed by less extreme ones — not because some mysterious force is acting on them, but because part of their extreme character was due to chance.
The discovery story
In the late 19th century, Francis Galton studied the heights of parents and their children. He observed a striking pattern:
- very tall parents generally had tall children;
- but, on average, the children were slightly shorter than them.
Conversely:
- very short parents often had short children;
- but, on average, the children were slightly taller than them.
Heights seemed to « drift back » toward the population average. Galton called this phenomenon « regression toward mediocrity », later renamed regression to the mean.
Why does it happen?
Imagine that an athletic performance depends on two factors:
- real talent;
- an element of luck.
An athlete delivers an outstanding performance. It probably reflects genuine ability — but also an especially favourable set of circumstances.
Next time, their talent is the same. But that exceptional luck has little reason to repeat itself exactly. The result will therefore often be less spectacular: the performance regresses toward its usual level.
The patient who consults at their worst
Most patients don’t make an appointment when they feel fine. They consult:
- during a painful flare;
- when their condition is worsening;
- when symptoms hit a peak.
In other words, the starting point of the observation is already an extreme. Even without treatment, a portion of these patients would have improved spontaneously. That natural improvement can then be wrongly credited to whatever intervention was performed.
Why alternative therapies sometimes « seem » to work
This phenomenon explains part of the success of many unvalidated approaches. The patient consults when they are suffering the most, at the worst point of their illness. After the consultation:
- symptoms naturally subside;
- the acute episode resolves;
- the disease drifts back to its usual level.
The patient concludes: « the treatment worked. » When in fact part of the improvement may have been already under way before the intervention.
This is exactly why clinical trials use control groups.
Athletes and coaches
In sport, regression to the mean is everywhere. A coach often notices that:
- when they praise an athlete after an excellent performance, the next one is worse;
- when they harshly criticise an athlete after a bad performance, the next one is better.
They may conclude: criticism works better than compliments. But all they are observing is regression to the mean. An exceptional performance is statistically unlikely to be followed by an even more exceptional one. Likewise, a very poor performance is often followed by a return to the usual level.
Supplements and diets
The phenomenon is especially common in nutrition. A person starts a new supplement at a moment when:
- fatigue is at its peak;
- weight is at its highest;
- performance is degraded.
A few weeks later, things improve: weight drops, fatigue eases. The supplement takes all the credit. Yet part of the improvement may simply reflect a natural return to the average.
Regression to the mean in public health
Many evaluations are affected:
- prevention campaigns;
- audits;
- hospital performance indicators;
- healthcare-associated infections;
- hand-hygiene compliance.
A unit posting unusually bad results in a given year will often show an improvement the following year, even without any major intervention. Conversely, a unit with exceptionally good results may see a slight decline. These fluctuations do not necessarily reflect a real change in care quality.
Link with the adaptive view of health
Regression to the mean illustrates a fundamental property of biological systems:
Life constantly fluctuates around dynamic equilibria.
Health is not a straight line. It oscillates continuously under the influence of:
- sleep;
- stress;
- diet;
- infections;
- physical activity;
- biological randomness.
Observing an individual at a single point in time is often closer to photographing a wave than measuring a mean level.
How to avoid being fooled
Several strategies help limit interpretation errors.
- Repeat measurements — a single value is rarely representative.
- Look at long trends — sustained changes are far more informative than isolated fluctuations.
- Use control groups — they estimate what would have happened without intervention.
- Be wary of spectacular results — the more extreme a result, the more likely regression to the mean is at work.
Conclusion
Regression to the mean is a universal statistical law. It quietly operates in medicine, sport, economics, education and public health.
Its main danger is not statistical — it is psychological. We love to attach a cause to every change we observe. Yet some changes are not the consequence of an intervention: they are simply the natural expression of the world’s variability.
Understanding regression to the mean means accepting that an improvement is not always proof of effectiveness, and that a deterioration is not always a sign of failure. It also means embracing a deeper idea:
In complex systems, a return to the ordinary is often more frequent than miracles.
Key takeaways
- An extreme value is statistically followed by a value closer to the mean, with no intervention required.
- Patients consulting at their worst, athletes after a disastrous match, hospital units in a bad year: all regress toward their usual level.
- This mechanism fuels the illusion of effectiveness of many interventions (supplements, alternative therapies, coaches’ dressing-downs).
- Safeguards: repeated measurements, long-term trends, control groups, scepticism toward spectacular results.
- Understanding Galton means accepting that biology fluctuates — and that every return to normal is not a miracle.
This article is an educational summary for informational purposes only. It is not a substitute for individual medical advice or official guidelines.
Laisser un commentaire