Correlation versus Causation

Options
johloz
johloz Posts: 176 Member
in Food and Nutrition
They are not the same!

As an example, say I conduct a scientific study in which I conclude that people who wear size small clothes are highly statistically likely to weigh less than those who wear size extra-large. I think we can all agree with that.

Now say I take a huge leap away from logic and say that the CORRELATION between people wearing size small and weighing less, is instead CAUSATION -- people weigh less because they wear size small. I create a diet called the "Skinny Jeans Diet" in which you squeeze yourself into the smallest possible size, and you lose weight with no other lifestyle changes. Amazing, right?

No, of course you know better and would not fall for that.

Then why do we fall for it when it comes to eating grapefruit, celery, or protein; drinking magic shakes or juices; or taking vitamins? I've even seen commercials that push eating more whole grains or drinking milk using the EXACT same "logic" I used in my example above.

Take everything you hear with a grain of salt (not literally, too much sodium causes weight gain, right?). Please, before you risk your time, money, and body, check to see if the claims of your new diet plan or supplement are backed up by science, or at least logic. Correlation is not causation!

Replies

  • FredDoyle
    FredDoyle Posts: 2,273 Member
    Options
    Yes. There seems to be a correlation between shark attacks and ice cream sales on the beach. It seems when the stands are closed in the off season, there are less swimmers attacked. Those ice cream vendors must be the cause. Better shut them down.
  • johloz
    johloz Posts: 176 Member
    Options
    No!!! More ice cream!
  • Chief_Rocka
    Chief_Rocka Posts: 4,710 Member
    Options
    sounds like broscience
  • UsedToBeHusky
    UsedToBeHusky Posts: 15,229 Member
    Options
    :drinker:

    I've ranted on this before myself!
  • neanderthin
    neanderthin Posts: 9,969 Member
    Options
    The states with the highest obesity rate listen to the most rap music. Rap music causes obesity, who would have thought.

    Blood is found to be most prevalent indicator of violent crime, therefore blood is the cause of violent crime. :bigsmile:
  • myofibril
    myofibril Posts: 4,500 Member
    Options
    I create a diet called the "Skinny Jeans Diet" in which you squeeze yourself into the smallest possible size, and you lose weight with no other lifestyle changes. Amazing, right?

    Really? That sounds brilliant!

    Can you tell me what to do or where I can buy the book?
  • diodelcibo
    diodelcibo Posts: 2,564 Member
    Options
    Do you even many-one relationship?
  • iAMsmiling
    iAMsmiling Posts: 2,394 Member
    Options
    The states with the highest obesity rate listen to the most rap music. Rap music causes obesity, who would have thought.

    Blood is found to be most prevalent indicator of violent crime, therefore blood is the cause of violent crime. :bigsmile:

    Actually, that one is true. Rap music ruptures brain cells and causes them to swell horribly.
  • ArroganceInStep
    ArroganceInStep Posts: 6,239 Member
    Options
    impossibru
  • neandermagnon
    neandermagnon Posts: 7,436 Member
    Options
    everyone who breathes dies. therefore breathing causes death. we must all stop breathing RIGHT NOW
  • sissiluv
    sissiluv Posts: 2,205 Member
    Options
    There's a concept I learned in my philosphy class called Magical Thinking. It's an evolutionary thing that turns correlation, in our minds, to causation. We were built to notice patterns and live by them.

    Or something like that anyways.
  • AlongCame_Molly
    AlongCame_Molly Posts: 2,835 Member
    Options
    QFT! Great thread!
  • RenfieldX
    RenfieldX Posts: 87 Member
    Options
    pchart1_thumb.jpg
  • PomegranatePriestess
    PomegranatePriestess Posts: 2,455 Member
    Options
    Now say I take a huge leap away from logic and say that the CORRELATION between people wearing size small and weighing less, is instead CAUSATION -- people weigh less because they wear size small. I create a diet called the "Skinny Jeans Diet" in which you squeeze yourself into the smallest possible size, and you lose weight with no other lifestyle changes. Amazing, right?

    photo_zpse3c1ba1c.jpg
  • Firefox7275
    Firefox7275 Posts: 2,040 Member
    Options
    Then why do we fall for it when it comes to eating grapefruit, celery, or protein; drinking magic shakes or juices; or taking vitamins? I've even seen commercials that push eating more whole grains or drinking milk using the EXACT same "logic" I used in my example above.

    "I'm not looking for a quick fix but ..."

    I can't tell you why US citizens trust advertising executives and journalists over their own government. In the last couple of weeks I've read claims on MFP that the US government is pushing overly high protein diet for the benefit of the meat industry, overly high grain diet for the benefit of the grain industry and dairy products for the benefit of the dairy industry. Whereas obviously the advertising, journalism and fast food industries are only concerned for your health.

    Nothing inherently wrong with the average US/ UK citizen eating more wholegrains IMO, if the product actually fitted into the 'cereals/ grains/ carbs' category and not the '10% of daily calories as processed/ sugar and fat laded ****e' category, and was substituting for refined crap not leading to higher grain consumption overall. And there is some interesting research *suggesting* causation not simply correlation with dairy, depends how cleverly the commercial is worded.
  • bcattoes
    bcattoes Posts: 17,299 Member
    Options
    While correlation does not prove causation, it does not disprove it either. If you have several studies showing correlation at some point it's logical to theorize about causation. Assuming the studies are peer reviewed and valid, of course.
  • hearthemelody
    hearthemelody Posts: 1,025 Member
    Options
    Yay statistics!
  • neanderthin
    neanderthin Posts: 9,969 Member
    Options
    While correlation does not prove causation, it does not disprove it either. If you have several studies showing correlation at some point it's logical to theorize about causation. Assuming the studies are peer reviewed and valid, of course.
    That generally creates logical fallacies and there can be truth to correlations but it's the confounders that make it almost impossible to show cause and effect. Critical thinking I believe, and an open mind is paramount for connecting the dots.

Categories