Skip to content

What science has to say about weight loss

If you have the time and inclination, you can use Google Scholar to search academic literature on weight loss.  Here is my quick and dirty take on the literature out there.

For most people, it really is about eating less calories

Historically, a problem with diet research is that the researchers weren’t able to track what the participants are actually eating.  (Because many people lie or have unreliable memories.)  But by using something called doubly-labeled water, scientists are able to measure how many calories a person actually ate.  Those who are the most obese tend to eat significantly more than average.  They also tend to underestimate how much they actually eat (see “The validity of self-reported energy intake as determined using the doubly labelled water technique“).

A limited number of cases are due to genetics, medication, and other factors.

You can eat as much as you want and lose weight

Some studies are based on ad libitum (ad lib) diets, where the subjects are allowed to eat as much as they want as long as it follows the prescribed diet.  These diets may work because the subjects are trying to be healthier, they are trying to please their experimenters+nutritionist, or it could also be that they are eating less calories.  Some diets try to ‘trick’ people into eating less calories by making them eat less energy-dense foods (e.g. fruits and vegetables).  Or to put it differently… if you try to eat more fruits and vegetables, you will likely end up eating less calories every day.  Some studies show that ad libitum diets are slightly more effective than calorie-restricted diets, possibly because subjects find them easier to follow (and are cheating less?).

See “Provision of Foods Differing in Energy Density Affects Long-Term Weight Loss“.

Almost any diet will cause some weight loss

There are studies that look at:

  • Low-fat versus low-carb
  • High-protein versus low-protein
  • Low-fat vegetarian versus ADA diet for diabetics
  • “Paleo” diets
  • Low-energy density diets
  • Low glycemic index diets

Subjects show weight loss on all of these diets.

The bad news: subjects almost always (slowly) gain back the weight that they lost

Scientists haven’t quite figured out how to get people to keep weight off for the people who enroll in studies. This may be because subjects adhere less to their intervention.

Does that mean obese people are doomed?  I don’t think that it’s necessarily the case.  If you look at smoking, studies show that only roughly 10-33% are able to quit.  Yet surveys show that most people don’t have difficulty quitting smoking and that most do so cold turkey, without any special help.  Those who are more prone to success may be less likely to enroll in studies because they have already achieved their goals.

While researchers know that eating less calories will lead to weight loss, it may be difficult for human beings to actually follow this advice.  If we look at smoking, many smokers exhibit a desire to quit and may have tried several times to quit.  But they don’t.

Do we know for sure that overeating causes obesity?

Over the short term, there are many controlled studies that link calorie consumption to weight.  A potential flaw in the prevailing wisdom is that there are no long-term controlled studies (over several years) as far as I know.  Participants could be locked up in an institution while researchers carefully control and measure their food intake.  Such a study would require people willing to be somewhat isolated from the outside world and such a study may be very expensive to conduct (if done ethically).  As far as I know, no such study has been performed.  Scientists have performed such research to study dental health, but it was rather unethical as they tested mental health patients who were institutionalized.

What successful weight losers do

The National Weight Control Registry states that: “NWCR members [...] report continuing to maintain a low calorie, low fat diet and doing high levels of activity“.  However, this data may be affected by selection bias and does not scientifically prove that a low-calorie, low-fat diet with high levels of activity will work for everybody (studies show that it usually does not).

Low-fat versus low-carb

Surprisingly enough, low-carb diets seem to cause slightly more/faster weight loss than low fat.  However, side effects are higher in low-carb diets.  The issues of bias are tricky as most researchers have a bias against low-carb diets as it flies against the conventional wisdom.  It is in researchers’ interest to have their publications cited by others often, so there can be subtle incentives to conform.  On the other hand, when it comes to diet, some people have a religious attachment to their beliefs and may distort data to fit the desired conclusions.  So I would be careful about biases on both sides of the debate.

Studies show that those on a low-fat diet also slowly regain weight over time compared to conventional diets.  They are also largely ineffective in studies.

Weight loss pills

Amphetamines used to be prescribed for weight loss.  For very good reasons, this is no longer the case (amphetamines are considered a harmful drug).  Human ‘innovation’ will likely lead to more types of pills being proposed as a cure for weight loss.  However, I would personally be vary of trying such pills.  Historically, drugs have been prescribed and later on we discover that the cure is worse than the disease.  This has also historically been the case with artificial sweeteners and also for Olestra- almost all of these products have been pulled from the market.

There are a lot of affiliate marketing sites on the Internet that promote weight loss pills, mainly because [A] there is a lot of appeal to magical pills that solve our problems and [B] the margins on pills is incredibly high as they cost less than $1/bottle.  I would personally wait until other people try such ‘magical cures’ and wait for studies to show their effectiveness (or lack thereof).

I have not done research on Orlistat.


Studies do show that various types of surgery result in weight loss.  These studies typically track the level of complication rates and mortality rates following surgery.  The exact mortality rate is hard to pin down because surgery is typically performed on people who are less healthy than average.  Also, the obese have higher complication rates because it is difficult to operate on obese patients.

Those publishing the papers and statistics may have biases.  It is in surgeons’ interests to promote surgery (especially one that they have worked hard to pioneer) and play down its downsides.  Some people have biases against surgery (I am in this camp).

Some (most?) types of surgery do not have randomized, controlled long-term studies tracking whether or not patients are actually healthier following surgery.  Weight loss surgery tends to cause significant reductions in weight and most patients no longer have to take medication for diabetes.  On the other hand, there is a small chance of death, a chance of complications from the surgery, and the surgery may not necessarily be that effective in the long term as patients regain significant amounts of weight for some forms of weight loss surgery.

I would encourage you to do your own research.