Correlation my ***

Post Reply New Topic

Did you know that chocolate causes depression? We know this is true because studies show that depressed people eat more chocolate, so the chocolate obviously causes the depression.

Really, I really hate how people jump the gun with statistics and assume that just because a study suggests a correlation may exist, that A must cause B. First things first, there is something called coincidences and random error. It is IMPOSSIBLE to PROVE that there is a correlation between two things through a study. You can only find rough probabilities that a correlation exists.

Even if there is a correlation between A and B it doesn't mean that A causes B, and chances are A doesn't B. There is three possible relation between A and B if they are correlated. Either A causes B, B causes A, or some third thing C causes both A and B. Hmm, does not having a girlfriend turn some one ugly? No it does not.

Second, there is multiple ways A can cause B. My first category is indirect vs. direct. If A causes be directly then it will cause A without going out side of this system. If the causation is indirect it means A causes D, which in turn causes B. Causation also depends on more factors then just the presence of A, there may need to be both A and E for A to cause B. As an example of an indirect cause is that a man is blind, so he doesn't see a cliff and walks off it, killing him when he hits the bottom. Him being blind caused him to die, but not directly. It caused him to not notice a cliff, which caused him to walk off it, which caused him to fall to his death. If there was no cliff he would still be blind, but he wouldn't fall to his death.

Another way of categorizing causes is one-time vs. continuous. If a cause is one time then it means once A happens then B will happen and it will not be removed once A is removed. If A continuously causes B then B can be removed by removing A, though the effect often isn't immediate and it may take some time for B to go away. Taking a dagger out of a stab wound won't heal the cut, but removing an obstruction from an airway will allow the person to breath again.

Lastly, just because A can cause B doesn't mean that A is the only cause of B, chances are it isn't. Yes, strangling someone will kill them, but someone can also die because they were poisoned. There is lots of ways B can come about. A can also can cause multiple things, not just B. Also, just because A can cause B does not mean that A must cause B, this is very common with indirect causes. Really, go think about things for a moment and be more open minded before you conclude that the vaccination caused your sons autism.

Also, chocolate makes people happy, so if you are sad you should go eat some chocolate.

It's human pattern recognition coupled with confirmation bias. People go looking for evidence that marijuana is dangerous (that's what your chocolate example made me think of), find that many cannabis users are somewhat depressed, and immediately jump to the conclusion "cannabis causes depression!"

This has long been propagated as a fact. Nobody ever considers the inverse case, namely that people who are depressed are more likely to start self-medicating with cannabis (and that they would be even more depressed if it wasn't for their pot consume). And of course nobody considers that there simply might not be a causal relationship either way.

People should be forced to repeat the mantra "there really are coincidences" twenty times before each meal. Because many things are just coincidence. Hitler, Mussolini and Franco were non-smokers. Coincidence. Lack of nicotine use does not turn people into a fascist dictator. And vaccines don't cause autism.

I've noticed that I've been a lot happier since I've stopped eating chocolate so much. If it causes depression, why do women eat it when they have PMS? Wouldn't that make the moodiness even worse?

That was sarcasm. Chocolate does not cause depression, its a misinterpretation of statistics.

Do they really censor "ass"?

correlation =/= causation...

so im sure there is a correlation, like there is a correlation between ice cream sales and hot weather, but obviously ice cream sales don't cause hot weather.

If marijuana were a gateway drug due to most heroin users having been marijuana users, then water is a gateway to alcoholism since most alcoholics drink water.

No, I censored it myself.

Good point

Personally, I think that chocolate is the #1 gateway drug. Not only did almost 100% of all drug users consume chocolate from a young age, it also really is a drug.

Chocolate contains a cannabinoid called anandamide, which works exactly like THC. It further contains a serotonin precursor, an amphetamine called phenylethylamine, and a muscle relaxant called theobromine (that's the ingredient that is toxic for cats and dogs). It is quite the psychoactive cocktail. By giving it to children, we condition them from a young age to improve their mood with drugs, especially when they are feeling down.

But perhaps that's just as crazy and unfounded as the hypothesis of cannabis as a gateway drug.

One could however argue that hot weather causes higher ice cream sales

People swimming causes ice-cream sales, as more ice cream is sold on days when more people are swimming.

Or that ice cream sales mean higher crime rate.

It is hot in the summer, so people buy more ice cream. All the highschool and college kids are not in school so crime goes up.

Some Iranian women show their ankles. There are earthquakes in Iran. Can this really be a coincidence?
http://www.guardian.co.uk/world/2010/ap ... ran-cleric

Summary of OP:

1. Just because statistics suggest something doesn't mean it's true.
2. Even if there is a correlation, chances are it is not causation.
3. Even if there is causation it probably isn't in the way you expect.

The #1 rule in statistics: You can't prove anything.

Oh, yeah, and it's biased to no end.

More people would be swimming on hot days one would think. People tend to swim more on hot days.