It is widely believed that regular exercise helps people become fitter, stronger and healthier. For example, from Physical Activity Guidelines for Americans:

Regular physical activity is one of the most important things you can do for your health. It can help: [...] reduce your risk of cardiovascular disease; reduce your risk for type 2 diabetes and metabolic syndrome; reduce your risk of some cancers; strengthen your bones and muscles; improve your mental health and mood; improve your ability to do daily activities and prevent falls, if you're an older adult; increase your chances of living longer.

Indeed, there is plenty of evidence that exercise is correlated with lower incidence of various diseases.

However, it could be that people who are born naturally healthy (and have a genetically lower risk of these diseases) also are genetically coded to have more energy, and therefore exercise more.

Put another way, somebody who is sick and tired all the time is less likely to exercise, but this doesn't mean the lack of exercise caused their illness.

Is there any convincing evidence that this is a causal relationship from exercise to health, e.g. from randomized controlled trials? I am not interested in observational studies that show people who exercise are healthier - that much is very clear. I am interested in studies that prove that exercise directly causes health benefits such as those claimed above.

  • 2
    > Does exercise make you healthier, or do healthy people just exercise more? Both could be true. – fals dikotame Mar 27 '15 at 5:42
  • 3
    This has been flagged as containing no notable claim, but I think it is a widely-believed claim that doing exercise makes you healthier. Should be easy to prove with some randomised trials. (I remember hearing an experimenter reporting difficulties because members of a control group, having been enrolled in an exercise experiment, and then left alone with no treatment, became self-motivated by the neglect to increase their daily exercise themselves!) – Oddthinking Mar 27 '15 at 9:30
  • 2
    @georgechalhoub: I see the notable claim as "exercise makes you healthy", and John being skeptical about it, arguing correlation doesn't equal causation, and showing an inverse relationship is plausible. I am confident someone will post a randomised controlled trial which will demonstrate causality. – Oddthinking Mar 27 '15 at 10:28
  • 1
    @GeorgeChalhoub: Done. – Oddthinking Mar 27 '15 at 11:23
  • 2
    How would one measure health exactly? – Pharap Mar 28 '15 at 7:07


taking away your stress

exercise is proven to increase concentrations of norepinephrine, a chemical that can moderate the brain’s response to stress.

and boosting happy chemicals

usually within five minutes after moderate exercise you get a mood-enhancement effect.

and alleviating long-term depression

data suggests that active people are less depressed than inactive people, those who were active and stopped tend to be more depressed than those who maintain or initiate an exercise program

and improving your self-confidence

regardless of weight, sexual orientation, gender, or age, exercise can quickly elevate a person's perception of his or her attractiveness: self-confidence and self-worth.

and preventing cognitive decline

exercising boosts the chemicals in the brain that support and prevent degeneration of the hippocampus, an important part of the brain for memory and learning, which would prevent Alzheimer.

and reducing anxiety

the warm and fuzzy chemicals that are released after exercise can help people with anxiety disorders calm down

and help controlling addiction

we all are addicted to something, right? Whether it is drugs, alcohol, internet, food or sex, exercise can help in addiction recovery

and unleashing your creativity

It is reported that one gym session can boost creativity for up to two hours afterwards.

and reducing the chances of developing heart disease, stroke, and diabetes

It is reported by over 11 scientific articles that bicycling and walking reduce your chance of getting all those diseases.

and making you lose weight

you don't need evidence for that? do you?

and promoting better sleep

regular physical activity can help you fall asleep faster and sharpen your sleep.

and lowering risk of all the following

up to a 35% lower risk of coronary heart disease and stroke and up to a 50% lower risk of type 2 diabetes and up to a 50% lower risk of colon cancer and up to a 20% lower risk of breast cancer and a 30% lower risk of early death and up to an 83% lower risk of osteoarthritis and up to a 68% lower risk of hip fracture and a 30% lower risk of falls (among older adults) and up to a 30% lower risk of depression and up to a 30% lower risk of dementia

and preventing osteoporosis

Weight-bearing exercises, like running, walking and weight-lifting, help lower your odds of getting osteoporosis as you grow older

and reducing the severity of asthma

swimming is one of the best exercises for people with asthma, exercise helps in fewer or milder asthma attacks overall and a need for less medication.

and promoting a healthy pregnancy

Relaxation exercises and Kegel exercise that strengthen the pelvic muscles and back exercises are all important for pregnant women.

and having anti-ageing effects

exercise might be the best anti-ageing pill according to many scientists.

and improving your sex life

The medical research points towards it: the fitter you are, the better your sex life is.

and ignoring the fact that physical inactivity is twice as deadly as obesity

from recent report by the American Journal of Clinical Nutrition.

there is no evidence that exercise makes you healthier at all.


The OP updated his question saying he's asking about causation and not correlation.

It seems:

1) there is a lack of evidence that a causal link between physical exercise and good mental health exists.

This book called Workplace Health: Employee Fitness And Exercise provides a great summary to a series of studies which concluded that there is a lack of causal relationship between physical exercise and mental health.

enter image description here

In another large cross-sectional research(1), scientists and researchers were able to prove that physical exercise is correlated with lower risk of depression, however, fortunately for people against physical exercise, they couldn't prove a causal link between the two, researchers said they:

are not able to make any firm conclusions on the direction of causation in any of the associations described

NHS blogged about the study and said:

In general, this study adds to the evidence that exercise is good for mental health, although in isolation it does not prove a causal link between the two. It should be discussed in the context of what else is known about the benefits of exercise.

Digging deeper, I found that, according to a governmental report:

2) there is a causal link between physical exercise and reduction in risk in all-cause mortality, all CVDs combined, CHD, hypertension, colon cancer, and NIDDM.


The studies reviewed in this chapter indicate that physical activity is associated with a reduction in risk of all-cause mortality, all CVDs combined, CHD, hypertension, colon cancer, and NIDDM.

To evaluate whether the information presented is sufficient to infer that these associations are causal in nature, it is useful to review the evidence according to Hill’s classic criteria for causality (Hill 1965; Paffenbarger 1988).

Strength of Association. The numerous estimated measures of association for cardiovascular outcomes presented in this chapter generally fall within the range of a 1.5- to 2.0-fold increase in risk of adverse health outcomes associated with inactivity. This range represents a moderately strong association, similar in magnitude to the relationship between CHD and smoking, hypertension, or elevated cholesterol. The associations with NIDDM, hypertension, and colon cancer have been somewhat smaller in magnitude. The difficulty in measuring physical activity may lead to substantial misclassification, which in turn would bias studies toward finding less of an effect of activity than may actually exist. On the other hand, not controlling for all potential confounders could bias studies toward finding more of an effect than may actually exist. Efforts to stratify studies of physical activity and CHD by the quality of measurement have found that the methodologically better studies showed larger associations than those with lower quality scores (Powell et al. 1987; Berlin and Colditz 1990). In addition, cardiorespiratory fitness, which is more objectively and precisely measured than the reported level of physical activity, often is also more strongly related to CVD and mortality. Measures of association between physical activity and health outcomes thus might be stronger if physical activity measurements were more accurate. Consistency of Findings. Although the epidemiologic studies of physical activity have varied greatly in methodology, in ways of classifying physical activity, and in populations studied, the findings have been remarkably consistent in supporting a reduction in risk as a function of greater amounts of physical activity, or conversely, an increase in risk as a function of inactivity.

Temporality. For most of the health conditions included in this chapter (all-cause mortality, CVD, CHD, hypertension, NIDDM), longitudinal data from cohort studies have been available and have confirmed a temporal sequence in which physical activity patterns are determined prior to development of disease. For obesity and mental health, fewer longitudinal studies have been conducted, and findings have been more equivocal. Perhaps the strongest evidence for temporality comes from two studies of the effect of changes in activity or fitness level. Men who became more active or more fit had a lower mortality rate during follow-up than men who remained inactive or unfit (Paffenbarger et al. 1993; Blair et al. 1995).

Biological Gradient. Studies of all-cause mortality, CVD, CHD, and NIDDM have shown a gradient of greater benefit associated with higher amounts of physical activity. Most studies that included more than two categories of amount of physical activity and were therefore able to evaluate a dose-response relationship found a gradient of decreasing risk of disease with increasing amounts of physical activity (see Tables 4-1 through 4-8).

Biologic Plausibility. Evidence that physiologic effects of physical activity have beneficial consequences for CHD, NIDDM, and obesity is abundant (see Chapter 3, as well as the biologic plausibility sections of this chapter). Such evidence includes beneficial effects on physiologic risk factors for disease, such as high blood pressure and blood lipoproteins, as well as beneficial effects on circulatory system functioning, blood-clotting mechanisms, insulin production and glucose handling, and caloric balance.

Experimental Evidence. Controlled clinical trials have not been conducted for the outcomes of mortality, CVD, cancer, obesity, or NIDDM. However, randomized clinical trials have determined that physical activity improves these diseases’ risk factors, such as blood pressure, lipoprotein profile, insulin sensitivity, and body fat.

The information reviewed in this chapter shows that the inverse association between physical activity and several diseases is moderate in magnitude, consistent across studies that differed substantially in methods and populations, and biologically plausible. A dose-response gradient has been observed in most studies that examined more than two levels of activity. For most of the diseases found to be inversely related to physical activity, the temporal sequence of exposure preceding disease has been demonstrated. Although controlled clinical trials have not been conducted (and are not likely to be conducted) for morbidity and mortality related to the diseases of interest, controlled trials have shown that activity can improve physiologic risk factors for these diseases.

From this large body of consistent information, it is reasonable to conclude that physical activity is causally related to the health outcomes reported here.


Layman, E. M. (1960). Contribution of exercise and sport to mental health and social adjustment. In Science and medicine of exercise and sports, ed. W.R. Johnson, New York

Cureton, T. K. (1963) Improvement of psychological state by means of exercise programs. Journal of the Association for Physical and mental rehabilitation, 17, 14-25.

Chernen, L., FRIEDMAN, S., GOLDBERG, N., FEIT, A., KWAT, T. and STEIN, R. (1995) Cardiac disease and ...

Morgan, W.P. and O'Connor, P. J. (1988) Exercise and mental health. In Exercise adherence: its impact on public health, ed. R.K. DISHMAN. Champaign.

(1) Harvey SB, Hotopf M et al. Physical activity and common mental disorders. The British Journal of Psychiatry (2010) 197: 357-364

  • 13
    486 words of pure, unfiltered sarcasm. Would you mind providing references to the part and lowering risk of all the following? – Maurycy Mar 27 '15 at 13:18
  • 3
    It it not shown that despite those factors, exercise makes you no healthier. I suggest rephrasing it as besides (...), there is no evidence that exercise makes you healthier at all. – gerrit Mar 27 '15 at 14:30
  • 3
    Ah - but to play complete devil's advocate, presumably exercise could also have negative results (higher injury rates, perhaps)? Would these automatically be smaller than the benefits? – xorsyst Mar 27 '15 at 16:57
  • 13
    The question explicitly argues that correlation is insufficient. The question is about causality. Generally, this requires randomised controlled trials (ideally double-blinded, placebo-controlled, although I'm not sure what that entails here). Not all of your references support that (I spot-checked. First and last references are examples.) – Oddthinking Mar 27 '15 at 22:54
  • 8
    Why does it need to be sarcastic? It's unpleasant to read and it sounds like you have an agenda, instead of providing unbiased factual information. – RomanSt Mar 27 '15 at 23:45

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .