[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math

View post   
File: 22 KB, 400x286, johnny5..jpg [View same] [iqdb] [saucenao] [google]
4878555 No.4878555 [Reply] [Original]

Whenever I read news articles about new studies, I always see someone say "correlation is not causation." but I don't know what that means. can /sci/ explain?

sorry for the bad english

>> No.4878561

>>4878555
just because two things happen together does not mean one is the result of the other.

>> No.4878563

It means that just because two things happen together consistently, doesn't mean that one causes the other in any way.

>> No.4878567

>>4878563

thank you for the reply.

I still don't really get it but thanks.

>> No.4878570

>>4878555
100% of dying people drink water.
Therefore, water kills.
As you may see, we have a relation between people who drink water and people dying, but I can't say for sure that drinking water is the main reason why all these persons are dead. This is why you shouldn't trust statistics without any reasonable evidence/explanation.

>> No.4878574

>>4878567

Say if you have 20 guys. 4 of them have red hair. Those 4 with red hair are also left handed.

Just because the men with red hair are also left handed, it doesn't mean having red hair will make you left handed.

>> No.4878573

Basically what it means is that just because two things are co-related (for example: smoking cigars and getting lung cancer) doesn't mean that one causes the other. This is the excuse that people used for decades (and still use today) to continue smoking cigars despite thousands of studies. In fact, it's already been scientifically proven without a doubt that smoking cigars causes lung cancer, but smokers still don't buy it (not all of them, but a lot of them still do).

>> No.4878578

>>4878573
Smoking cigars is more likely to cause throat cancer than lung.

>> No.4878580

>>4878574

obviously but what does that have to do with my question? you guys are confusing me

>> No.4878584

>>4878580

The two attributes correlate, but one does not cause the other.

>> No.4878587

>>4878567
they mean that two things can correlate but it does not mean one affects the other. say that as hours spent watching TV increase, grades decrease. TV itself doesn't make you dumber, but chances are you are spending less time studying the more time you spend watching TV. also, two things could correlate just by coincidence. did you know people who play the accordion have higher divorce rates?

>> No.4878588

>>4878567
They say there is an inverse correlation between the number of pirates and global temperatures or something meaning they follow a similar trend but it's ridiculous to propose that they have anything to do with each other.

>> No.4878595

thank you guys very much for the replies. I still don't really get it. and some of the things you said are confusing and don't seem related. probably i'll be thinking of it later and it will finally click. but thanks anyways for trying to explain, maybe it's just not easy to explain. again thanks

>> No.4878603

>>4878595
Holy shit, it's explained perfectly here, if you don't understand this, it's because you're retarded not because this is hard to explain.

>> No.4878607

>>4878574
Are you joking or are you really not getting this?
I jumped, Japan got hit by an earthquake. My jumping did not cause the earthquake.
A then B, A=/=B

>> No.4878610

>>4878603

there's no need to be rude please, i am just asking for information.

>> No.4878611

>>4878603

If English isn't his first language, then it might be harder to grasp consider the vagueish examples that have been given.

Some phrases and words do not translate well into other languages..

>> No.4878620

>>4878607
>A then B, A=/=B
Don't confuse him with stuff like that, just say,

"I jumped and then there was an earthquake in Japan. My friend turned to me and said that I caused the earthquake by jumping. I told him that correlation does not imply causation, and then I called him a schizo and a faggot and to fuck off."

>> No.4878621

I think I get it... it's like if you see a news story where a new study found something, but they're not sure what caused it. right? this seems like not a very important concept

>> No.4878632
File: 31 KB, 500x478, Watching with Drink.jpg [View same] [iqdb] [saucenao] [google]
4878632

>>4878620
My bad, I thought he went through basic maths like the rest of the world and remembered what a variable was.

>> No.4878628

>>4878621
no, you are failing at this.

>> No.4878629

>>4878621
A lot of people don't understand this concept.

They will see news stories about lots of birds, bees, and fish dying, and they'll think they're all related and being caused by one thing, and that the world is ending.

>> No.4878633

>>4878629

thank you. is there maybe a real world example of this fallacy instead of just hypotheticals like "i jumped and it didn't cause an earthquake"?

>> No.4878636

>>4878621

Imagine if every day you wore blue jeans (for example), it rained.

You wearing blue jeans does not cause it to rain. It still correlates.

>> No.4878635

From Wikipedia:

Sleeping with one's shoes on is strongly correlated with waking up with a headache.
Therefore, sleeping with one's shoes on causes headache.

The above example commits the correlation-implies-causation fallacy, as it prematurely concludes that sleeping with one's shoes on causes headache. A more plausible explanation is that both are caused by a third factor, in this case going to bed drunk, which thereby gives rise to a correlation. So the conclusion is false.

>> No.4878642

https://simple.wikipedia.org/wiki/Correlation#Correlation_vs_causation

>> No.4878643

>>4878611

>>4878574
>>4878574
>>4878574
That's baby English, if you're over 18 and can't speak enough English to understand this, you're retarded.

>>4878621
If you see a news story where a new study found out that one thing and another thing are both happening at the same time, then that doesn't mean one thing causes another thing.

>inb4 more "imagine..." posts
OP is mentally retarded and should be ignored.

>> No.4878646

>>4878633
The post you replied to gave you one. Awhile back lots of birds, bees, and fish were dying and people said that this was proof that the world was ending.

>> No.4878647

>>4878621
No! Re-read this thread. And it's a very important concept, especially for you, who seems to be the type that's easily fooled.
Aw man, I hope you're not trolling.

>> No.4878650

I think I don't believe this "correlation is not causation" thing unless there's a good example where it has been proved scientifically. maybe i'm just sketpical like that. lol

thanks anyways

>> No.4878656

>>4878650

Now get the fuck out of /sci/.

>> No.4878662

>>4878650
that's the opposite of skeptical

>> No.4878664

>>4878656

I guess I was expecting scientific examples that's all, not trying to make anyone mad

>> No.4878668

>>4878664

What are you even talking about now, you're obviously trolling as I don't believe a single person on this planet can really be this stupid.

>> No.4878672

>>4878643

> if you're over 18 and can't speak enough English to understand this, you're retarded

English isn't my first language and I know how hard it is for someone to learn, especially when it's written down. Using the word "retarded" repeatedly in one post is usually a good way of telling if someone is actually retarded.

"Correlation does not equal causation" doesn't go well into my native language, for instance.

>> No.4878678

>>4878633
>real world example
people who smoke weed, on average, have lower grades than those who don't. this does not mean that smoking weeds inevitably causes lower grades. The correlation could be due to self-selection bias. people who care a lot about grades are more likely to be "good" kids who don't misbehave.

another. study programs correlate with better grades. however, the people who take those programs are more likely to try harder at school than the average population, and therefore, get better grades even if they did not take the program.

>> No.4878683

>>4878672
isn't there some sort of internet forum for non-retarded people who speak you're native language that you can ask instead of failing to understand explanations in the simplest language we can provide?

>> No.4878685

If two things are correlated it means that they happen together.

For example smoking marijuana but NOT nicotine correlates well with increased lifespan.

However, this does not mean that smoking marijuana CAUSES life extension. Marijuana has an negative impact on health, however people who smoke marijuana but avoid tabacco are more likely to exhibit other traits which increase lifespan, such as vegetarianism, regular exercise, etc.

Basically, unless you have a CAUSAL mechanism identified to explain two things which you observe to be correlated, you have to be skeptical of anybody who asserts that one causes the other.

>> No.4878690

>>4878678

ahh I get it now... hmm, but if that's true then how can you find any scientific data useful? all we do is experiment so if we don't think one thing causes another we can't really learn anything can we? I think some things are caused by other things and you can show that scientifically... if you just believe correlation isn't causation then how can we have medicine, for example.

>> No.4878687

>>4878683
>insults another's understanding of the English language
>who speak you're native language
> you're
fucking lel

>> No.4878693

>>4878683

Perhaps OP is intrigued by the high brow scientific discussion that goes on here in /sci/

I know I am. I relish it.

>> No.4878700

>>4878690

See you're on the ball, but the key word is "doesn't". It doesn't mean it never equals it. There's a correlation between jumping on grenades and dying, but there's a cause from one that leads to the other.

The phrase basically means "just because something matches up to something else, doesn't mean it makes it act that way"

>> No.4878701

>>4878690

Science experiments are not statistics. Studies and statistics aren't how people acquire scientific knowledge.

>> No.4878703

>>4878700

but how can something's cause only "sometimes" cause it? to use my medicine example... if i eat Tums my stomachs PH will go down. it will never go up. because Tums causes the PH to go down. but if you say "correlation is not causation" then that would imply sometimes when you take Tums your stomach PH doesn't go down. see what I mean?

>> No.4878705

>>4878703
Are you deliberately trying to misunderstand or are you just naturally gifted at it?

>> No.4878706

>>4878672

Neither does it into mine, quite frankly, I'd have quite some trouble translating this.
However, OP didn't ask for any translations, he asked for explanations.
The meaning of the phrase has been translated into as simple English as possible, even a baby could understand what "if a human jumps on earth and suddenly an earthquake happens, it doesn't mean the human caused the earthquake" means.

>>4878690
If we have insufficient data(not enough data), we might be able to assume that the things are correlated. If we look further into it[examine it closely(find more scientific data)], we might be able to prove that indeed one thing causes another.
The whole concept of the phrase is that, and I quote sherlock holmes here, "It is a capital mistake to theorize before you have all the evidence.".

>> No.4878711

the number of murders and the number of ice cream cones eaten

both rise in the summer

ice cream cones cause murder

>> No.4878713

>>4878690
you're coming off as a bit of a troll but i'm not sure so ill try anyways.. in a scientific experiment, you control for other variables that could affect your results. for examples, i will use the ones i previously gave you. say that you measured the average grades of a group of students before they started smoking weed. then you used the same group of students, made sure they study the same amount of time, and the material they are learning is not any harder than before but you made them all smoke weed. If the only thing that changed was the fact that they smoked weed and their grades lowered, then it is highly probable that smoking weed indeed causes lower grades. this is why experiments have control groups and test as few variables as possible at a time.

>> No.4878720

>>4878713

I still don't agree. I know for a fact that people who smoke weed tend to have lower grades... i know a lot of people like that myself. unless there's a scientific reason it does not then i'm not going to just beleive they aren't related.

>> No.4878723

>>4878720
of course they're related, but that's not the same as saying one CAUSES the other.

>> No.4878728

Holy fucksticks, surely nobody can misunderstand so consistently?

>> No.4878731

>>4878720

What if they have problems at home(their father is beating their mother and they are quite disturbed by this)?
This makes them want to smoke weed, but also causes their lack of motivation to study, which is the cause for their low grades.
Smoking weed doesn't CAUSE low grades, but is related to the issue.

>> No.4878732

>>4878728

maybe it just isn't easy to understand. that's ok.

thanks anyways for trying /sci/.

>> No.4878733

>>4878732

Maybe you're just retarded?

Thanks anyway for posting, leave /sci/.

>> No.4878737

I really think it is a language barrier. Try to google it in your natural language. I have the same problem understanding technical ideas in foreign languages I have learned.

Although, it really isn't such a difficult concept. I really am hoping it is a matter of English translation.

>> No.4878750

There is no way this is a matter of English translation, it is either trolling or severe lack of mental capacity.

>> No.4878762

>>4878732

What is your first language?

>> No.4878770

>>4878762

Portuguese. it doesn't matter I'm going to try asking someone else for help instead of getting berated like this.

>> No.4878777

For OP:

https://amota.wordpress.com/2007/05/08/lgica-e-falcia-correlao-causalidade_climatologia_002/

Shoop da whoop motherfucker.

>> No.4878778
File: 56 KB, 500x325, Lemongraph.jpg [View same] [iqdb] [saucenao] [google]
4878778

This thread calls for a graph!

>> No.4878788

Imagine A causes B and A causes C, but we don't know this yet.
One could see from experiment that If B occurs, C has a higher chance of occuring than if B doesn't occur. One could now make the fault to think that B causes C, but that would be a wrong assumption. There is a relation between the two, but not one that says one is the result of the other.

I hope I made it a bit more clear to you.

>> No.4878789

>>4878788

He's already confused by ALL the previous examples.

>>4878777 gave him an article in his first language. I think it's enough.

>> No.4878790

>>4878777
>>4878778
>>4878788

thanks for the responses. it still just doesn't seem very scientific.

>> No.4878793

Correlation is when two variables show a trend. Suppose you have two variables, called x and y. They could be anything. For example, suppose x is the rate of obesity in different populations, and y is the amount of food consumed in different populations. Suppose you went around to different cities and collected data on how fat each person was and how much they ate. If you plotted each person on a graph of x vs y, you might find a positive correlation, meaning data points with a high value of x tend to have a high value for y, and vice versa. So x and y are correlated.

Causation is when one variable has a direct influence on another variable. In the example above, you might wonder if you could somehow independently force a change in x, would that result in a change in y. With infinite funds and an evil army to back you up, you force liposuction on everybody in the population, thereby lowering the obesity rate (x). After performing forced liposuction on everybody who gets over a certain weight, you try to see if people eat the same amount of food (y). Suppose you find everybody still eats exactly the same amount of food. Therefore you conclude (x) has no direct cause on (y).

But suppose instead you force everybody in the population to eat small rations, forcing a lower value of (y). After several months you find everybody becomes skinny, meaning a lower value for (x). In this case, you directly changed y, which resulted in a change of x. Therefore you conclude (y) has a direct cause on (x).

>> No.4878794

>>4878770

but you're on an anonymous board...

>> No.4878795

>>4878770
Brazilian reporting in.
It's not a language barrier, it translates just fine.

>> No.4878797

>>4878790

It's not about being scientific, it's about being logical (and I would be the first to argue that being scientific is among other being logical).

If you don't think that it is logical, then you do not understand the concept correctly.

>> No.4878798

>>4878795

maybe it's easier to understand in Brazilian

>> No.4878801

>>4878798

https://amota.wordpress.com/2007/05/08/lgica-e-falcia-correlao-causalidade_climatologia_002/
https://amota.wordpress.com/2007/05/08/lgica-e-falcia-correlao-causalidade_climatologia_002/
https://amota.wordpress.com/2007/05/08/lgica-e-falcia-correlao-causalidade_climatologia_002/

>> No.4878800

>>4878798
...What? Brazilians speak Portuguese. There are local minor diferences, but the formal language is exactly the same.
Or do you believe Americans speak "American"?

>> No.4878802
File: 30 KB, 500x350, piratesarecool4.jpg [View same] [iqdb] [saucenao] [google]
4878802

Global warming caused by pirates!

>> No.4878805

I think I see where this is headed.

Half of /sci/ is pulling their hair out, trying to understand how OP could not understand such a simple concept.
The other half of /sci/ believes he is trolling, but some of them can't help but try to bring him around anyway.
Some of the first group are also trying earnestly to help because they think that he's just t h i s close to grasping it.
Meanwhile, only OP knows for sure whether he is trolling or not. Or maybe he doesn't.
And we all watch the slow-motion disaster.

>> No.4878809

>>4878800

You silly, we all know that "Americans" actually speak "United-State-ian"!

>> No.4878812

>>4878805

I honestly don't think he's trolling. It's an abstract concept, so it seems obvious but only in hindsight to those who learned it formally. Honestly, how many people here do you think took a statistics course and could formally define "correlation" and "causation?" Everyone is restating the words in their explanations, but not defining what they mean.

>> No.4878816

Even if OP is joking, I personally know a few people this thick.

>> No.4878829

>>4878809
Fun fact: American and British English do have more differences than Portuguese and Brazilian's Portuguese. Our presidents signed a treaty unifying our written language's orthography recently, leaving only spoken peculiarities now.
>>4878816
Do they often come to /sci/ to ask questions and not clear their doubts?

>> No.4878836
File: 193 KB, 361x269, corzine-obama.png [View same] [iqdb] [saucenao] [google]
4878836

What this should say is, correlation between A and B does not mean A causes B.

However when there is correlation there is a causal relationship in there somewhere. What is not shown by a correlation is the nature of the causal relationship.

Eg everyone who drinks water dies eventually. Therefore drinking water kills you. The causal relationship here is that being a mortal human being causes you to eventually die, and it also causes you to drink water.

In this case

A is correlated to B but the relationship is one of a common cause. Common Cause is a common issue - it is also called Confounding Factors.

X->A
X->B

Being in prison is correlated with having committed a crime. However being in prison does not cause you to have committed a crime in the past. The causation operates in the opposite direction. Committing a crime may cause you to be put in prison.

So: Correlation does imply causation (or a logical relationship). But it does not specify the type of causative relationship.

>> No.4878846

>>4878829

They ask questions, the moment they lose focus on what I am explaining, they immediately scratch their head and yell stop and change the subject. I ask them why, they say "It's too hard to remember all of that, nevermind about it."

Some people seem like they have a very quick 'Instant brain shut off valve.'

>> No.4878879

>>4878607

Nobody really uses mathematical symbols to explain a logical concept at an introductory level, it's just something that gets thrown in there for easier memorization and expression.

Also, try and remember that this guy knows one more language than you do.

And to the OP, you've been accusing people of saying unrelated things, so of course you're going to get berated. You accused them of something, putting blame. Why they couldn't be honest about their annoyance, beats the fuck out of me, but that's why it's happening.

>> No.4879473

>>4878703
Ok here's the difference.
"Correlation is not causation" is something that applies a lot more to statistics, rather than scientific experiments.

We can prove that Tums lowers stomach pH by taking a tablet and placing it in a dish of stomach acid and measuring the pH. We can control the conditions of this experiment to a degree and repeat it enough times that we can be sure the Tums is what caused the pH to change.

Now, if we take a statistical study that for example says "people who are up-to-date on the news are less likely to be involved in car accidents", it does not mean that being up-to-date on the news protects you from car accidents. Even if they found people that didn't read the news were twice as likely to be in a car accident, there are too many uncontrollable factors to say that failure to keep up with the news is the cause of these accidents. More than likely, reading the news and not being in car accidents are both results of a larger cause, such as the type of person that reads the news being more observant or less of a risk-taker on average.

>> No.4879515

Here's an easy to get example.

Some studies showed that the prison population has higher circulating levels of testosterone (in their blood). This is a correlation between (A) - having elevated testosterone levels, and (B) - being in prison.

However other studies also showed that those who commit crimes don't have significantly higher circulating levels of testosterone than the rest of the population (before going to prison). What does that mean? It means that A doesn't necessarily cause B, and it might be some other variable which makes inmates have elevated testosterone levels inside the prison, but not before entering it. A violent environment for example could cause them to be vigilent and ready for aggression most of the time, which could increase their T levels.

>> No.4879530

It means 2 things can co-occur without being casually related.

For example, if I pray every time before taking a test, I might notice that there is a correlation between praying and getting a good grade. Yet, the correlation would obviously only exist because I study and thus always get good grades.

Aka There is a correlation between prayer and getting good grades but not a causation.