When is it cause, not correlation?

User avatar
glucose101
Posts: 423
Joined: Sun Mar 29, 2009 12:23 am

When is it cause, not correlation?

Postby glucose101 » Thu Jun 23, 2011 10:18 pm

"Correlation does not imply causation." I can't tell you how many times I've heard that. On the LSAT, it seems like causation is never the "correct" answer, and that correlation is always mistaken for causation. How would causation be "correct" on the LSAT?

User avatar
AlabamaIceman
Posts: 94
Joined: Mon Feb 14, 2011 3:32 pm

Re: When is it cause, not correlation?

Postby AlabamaIceman » Thu Jun 23, 2011 10:21 pm

I'm not sure what you're asking. Most of the time when that phrase appears, it's in reference to a "weaken" question, and you are pointing out that the faulty argument in the prompt shows correlation but not causation. If it shows causation, then the argument being put forth is most likely correct.

Conversely, if the question asks you to prove something or strengthen an argument, you would be looking for causation rather than correlation.

User avatar
NYC Law
Posts: 1569
Joined: Thu May 26, 2011 3:33 pm

Re: When is it cause, not correlation?

Postby NYC Law » Thu Jun 23, 2011 10:22 pm

There's almost never causation, researchers have a very difficult time finding actual direct causation. There are some research and statistical methods that can get you close to finding a causation, but it still usually can't prove it.

Just assume there's never a causation, there can always be an extraneous variable.

User avatar
glucose101
Posts: 423
Joined: Sun Mar 29, 2009 12:23 am

Re: When is it cause, not correlation?

Postby glucose101 » Thu Jun 23, 2011 10:25 pm

NYC Law wrote:There's almost never causation, researchers have a very difficult time finding actual direct causation. There are some research and statistical methods that can get you close to finding a causation, but it still usually can't prove it.


Will the LSAT ever have such scenarios? I realize the LSAT will probably only use correlation/causation on flaw/weaken questions, but I just want to make sure there aren't instances that the LSAT would defy this.

User avatar
suspicious android
Posts: 938
Joined: Tue Feb 09, 2010 4:54 pm

Re: When is it cause, not correlation?

Postby suspicious android » Thu Jun 23, 2011 10:40 pm

Are you asking when a causal conclusion could be validly drawn on the basis of a mere correlation? Never. A causal conclusion based on a mere correlation may be true, but never validly drawn. So I'm not sure what you mean by "correct" here, but there are definitely some instances where a causal argument based on a correlation has not hinged on that correlation issue, but on some other flaw instead.

TMC116
Posts: 287
Joined: Mon Jun 06, 2011 6:08 pm

Re: When is it cause, not correlation?

Postby TMC116 » Thu Jun 23, 2011 10:46 pm

suspicious android wrote:Are you asking when a causal conclusion could be validly drawn on the basis of a mere correlation? Never. A causal conclusion based on a mere correlation may be true, but never validly drawn. So I'm not sure what you mean by "correct" here, but there are definitely some instances where a causal argument based on a correlation has not hinged on that correlation issue, but on some other flaw instead.


Exactly.

If you see a causal argument, assume its wrong/flawed. If it is the basis for a conclusion, then that's where the flaw is (so act accordingly when asked to weaken, strengthen, or identify the flaw)

User avatar
EarlCat
Posts: 610
Joined: Mon Mar 12, 2007 4:04 pm

Re: When is it cause, not correlation?

Postby EarlCat » Thu Jun 23, 2011 11:10 pm

glucose101 wrote:"Correlation does not imply causation." I can't tell you how many times I've heard that. On the LSAT, it seems like causation is never the "correct" answer, and that correlation is always mistaken for causation. How would causation be "correct" on the LSAT?

You can only conclude causation if the premise tells you there's causation.

User avatar
glucose101
Posts: 423
Joined: Sun Mar 29, 2009 12:23 am

Re: When is it cause, not correlation?

Postby glucose101 » Fri Jun 24, 2011 12:20 am

EarlCat wrote:
glucose101 wrote:"Correlation does not imply causation." I can't tell you how many times I've heard that. On the LSAT, it seems like causation is never the "correct" answer, and that correlation is always mistaken for causation. How would causation be "correct" on the LSAT?

You can only conclude causation if the premise tells you there's causation.


Ok! Thanks! This is what I assumed, but just wanted to make sure.

MCockerill08
Posts: 6
Joined: Tue Jun 07, 2011 9:44 am

Re: When is it cause, not correlation?

Postby MCockerill08 » Fri Jun 24, 2011 9:32 am

Causation can rarely if ever be "proven" on an LSAT LR question (though in sufficient assumption questions, the correct answer could "assume" the casual link that would justify an argument's conclusion can be assumed), but arguments alleging causation can be strengthened by eliminating alternative causes or buttressing the credibility of the link (e.g.: if expanding and diversifying the sample size of our original study yields the same correlation, it is more likely that some casual link exists) .

Audio Technica Guy
Posts: 317
Joined: Tue Jun 29, 2010 5:21 pm

Re: When is it cause, not correlation?

Postby Audio Technica Guy » Fri Jun 24, 2011 12:05 pm

yeah, there has never been a single question on the LSAT were causation was successfully proven. LSAC would likely shy away from this, because there isn't a logically agreed upon definition of the word causation. It would be too ripe for counterargument.

So, on the LSAT, the only possible way you can have causation is if a premise straight up tells you it is, without a doubt, causation. Off the top of my head, I can't actually recall any question that did that either.

skip james
Posts: 264
Joined: Sat Sep 19, 2009 2:53 am

Re: When is it cause, not correlation?

Postby skip james » Fri Jun 24, 2011 9:45 pm

The above isn't necessarily true. Since we are told to accept premises at face value, if we are given causal premises, like 'smoking always causes cancer' and 'cancer always causes death', we could be asked what 'must be true' given those facts, and in the example above, it would be absolutely true that smoking always causes death.

And naturally, I'm smoking a cigarette as I write this..

skip james
Posts: 264
Joined: Sat Sep 19, 2009 2:53 am

Re: When is it cause, not correlation?

Postby skip james » Fri Jun 24, 2011 9:49 pm

Audio Technica Guy wrote:yeah, there has never been a single question on the LSAT were causation was successfully proven. LSAC would likely shy away from this, because there isn't a logically agreed upon definition of the word causation. It would be too ripe for counterargument.

So, on the LSAT, the only possible way you can have causation is if a premise straight up tells you it is, without a doubt, causation. Off the top of my head, I can't actually recall any question that did that either.
whoops.. Didn't read this last bit.

But from my recollection there have been a couple (though rare) valid causal arguments.

From what I remember the answers are usually phrased something like, 'doing blah increases risk of something blah'

User avatar
suspicious android
Posts: 938
Joined: Tue Feb 09, 2010 4:54 pm

Re: When is it cause, not correlation?

Postby suspicious android » Fri Jun 24, 2011 11:20 pm

skip james wrote:But from my recollection there have been a couple (though rare) valid causal arguments.

From what I remember the answers are usually phrased something like, 'doing blah increases risk of something blah'


So what are you saying? That a causal statement can be validly concluded? That's already the consensus of the thread, it seems. Just need a premise that indicates causality. Absent that, it's impossible.

skip james
Posts: 264
Joined: Sat Sep 19, 2009 2:53 am

Re: When is it cause, not correlation?

Postby skip james » Sat Jun 25, 2011 1:23 am

suspicious android wrote:
skip james wrote:But from my recollection there have been a couple (though rare) valid causal arguments.

From what I remember the answers are usually phrased something like, 'doing blah increases risk of something blah'


So what are you saying? That a causal statement can be validly concluded? That's already the consensus of the thread, it seems. Just need a premise that indicates causality. Absent that, it's impossible.

I take it you don't buy into the second analogy of Kant's Critique of Pure Reason then.

User avatar
suspicious android
Posts: 938
Joined: Tue Feb 09, 2010 4:54 pm

Re: When is it cause, not correlation?

Postby suspicious android » Sat Jun 25, 2011 1:38 am

skip james wrote:I take it you don't buy into the second analogy of Kant's Critique of Pure Reason then.


I'll admit I've forgotten almost everything about Kant I ever knew, but wasn't that position you're referring to the idea that we must assume causality as we intuitively understand it to be real? That's not to say that it's really real, just that it's how we're stuck interpreting the world. Regardless, I always found Hume to be on much firmer ground on this sort of thing. That might be because I'm not smart enough to really understand Kant, but well.. I'm okay with that.

User avatar
westinghouse60
Posts: 392
Joined: Mon Aug 02, 2010 2:27 am

Re: When is it cause, not correlation?

Postby westinghouse60 » Sat Jun 25, 2011 4:05 am

suspicious android wrote:
skip james wrote:I take it you don't buy into the second analogy of Kant's Critique of Pure Reason then.


I'll admit I've forgotten almost everything about Kant I ever knew, but wasn't that position you're referring to the idea that we must assume causality as we intuitively understand it to be real? That's not to say that it's really real, just that it's how we're stuck interpreting the world. Regardless, I always found Hume to be on much firmer ground on this sort of thing. That might be because I'm not smart enough to really understand Kant, but well.. I'm okay with that.


Didn't Hume say we can never know that one event causes another, only that one event seems to always follow the other? I'm probably oversimplifying that, or maybe I'm way off entirely.

User avatar
suspicious android
Posts: 938
Joined: Tue Feb 09, 2010 4:54 pm

Re: When is it cause, not correlation?

Postby suspicious android » Sat Jun 25, 2011 5:17 am

westinghouse60 wrote:Didn't Hume say we can never know that one event causes another, only that one event seems to always follow the other? I'm probably oversimplifying that, or maybe I'm way off entirely.


Yeah, that's basically right. It went like this:

Hume writes a bunch of stuff, says that induction and causation and just reason in general are all bullshit, we can only know what we can see.

Kant says "Man, you just blew my mind." He then resolves to prove Hume completely wrong because he is the most boring man in the history of the world and doesn't like Hume ruining all the nice ideas everyone has about God, the soul, rationality, etc.

Ten years later, he puts out a book that says it turns out all that stuff is real after all, and we can sorta prove it if we just think about it really, really hard. It doesn't make much sense, but he sounds so smart that everyone just kinda goes along with it for awhile, because they don't want to look dumb and it's what they wanted to believe anyway.

That was my take anyway, I didn't do too great in that class.

skip james
Posts: 264
Joined: Sat Sep 19, 2009 2:53 am

Re: When is it cause, not correlation?

Postby skip james » Sat Jun 25, 2011 6:28 am

westinghouse60 wrote:Didn't Hume say we can never know that one event causes another, only that one event seems to always follow the other? I'm probably oversimplifying that, or maybe I'm way off entirely.


Yup. Hume's the guy who made correlation => causation a flaw. Major lsat props to him.

suspicious android wrote:Yeah, that's basically right. It went like this:

Hume writes a bunch of stuff, says that induction and causation and just reason in general are all bullshit, we can only know what we can see.

Kant says "Man, you just blew my mind." He then resolves to prove Hume completely wrong because he is the most boring man in the history of the world and doesn't like Hume ruining all the nice ideas everyone has about God, the soul, rationality, etc.


I don't think this is fair. I'm not sure if Kant is right about the whole ordeal, but the dude did put a decade into thinking about it, and I think to be charitable, we should at least think about it for more than 10 minutes before making a dismissive forum post about it.

I struggled with Kant a lot myself, but only because I thought there might be something there that I wasn't seeing. It's easy to just shrug it off as philosophic nonsense, but it's quite difficult to encounter some complex, very well-thought out arguments and to really try and understand where it's all coming from.

I'm not sure if I buy into Kant myself, but I'm definitely not as dismissive as you about it, or at the very least, I try not to be..

Who knows.. I could be wasting my time.. I've definitely done that before.

jlee282
Posts: 17
Joined: Tue Mar 16, 2010 11:28 am

Re: When is it cause, not correlation?

Postby jlee282 » Sat Jun 25, 2011 7:43 am

If I remember correctly, there were few problems on the LR that asked to look for the necessary assumption of a causal relationship. There is a great chapter on PowerScore Bible about causal relationships. Perhaps you can look that up.

User avatar
suspicious android
Posts: 938
Joined: Tue Feb 09, 2010 4:54 pm

Re: When is it cause, not correlation?

Postby suspicious android » Sat Jun 25, 2011 2:07 pm

skip james wrote:I don't think this is fair. I'm not sure if Kant is right about the whole ordeal, but the dude did put a decade into thinking about it, and I think to be charitable, we should at least think about it for more than 10 minutes before making a dismissive forum post about it.


What makes you think I only thought about it for 10 minutes? I wasted three good months of my life trying to understand Kant before I dismissed him.:)

skip james
Posts: 264
Joined: Sat Sep 19, 2009 2:53 am

Re: When is it cause, not correlation?

Postby skip james » Sat Jun 25, 2011 8:08 pm

Sorta embarrassing.. but I can top that. By a lot.

And three months really isn't too long, especially for Kant. Three months is tantamount to saying 'I took a Kant class once'. That's barely enough time to a grasp of what's going on in the Aesthetics of the Critique.

But I'm biased. I like Kant, regardless of whether or not he ultimately succeeds. He's got heart, plus he's probably a nice guy, having invented the categorical imperative and all.

User avatar
westinghouse60
Posts: 392
Joined: Mon Aug 02, 2010 2:27 am

Re: When is it cause, not correlation?

Postby westinghouse60 » Sat Jun 25, 2011 9:56 pm

skip james wrote:Sorta embarrassing.. but I can top that. By a lot.

And three months really isn't too long, especially for Kant. Three months is tantamount to saying 'I took a Kant class once'. That's barely enough time to a grasp of what's going on in the Aesthetics of the Critique.

But I'm biased. I like Kant, regardless of whether or not he ultimately succeeds. He's got heart, plus he's probably a nice guy, having invented the categorical imperative and all.


Only Kant I've read is Groundwork of the Metaphysics of Morals. Although I'm about to start Critique of Pure Reason now that the LSAT is done and I have tons of free time lol.

skip james
Posts: 264
Joined: Sat Sep 19, 2009 2:53 am

Re: When is it cause, not correlation?

Postby skip james » Sat Jun 25, 2011 10:14 pm

you're gonna need a lot of coffee.

User avatar
suspicious android
Posts: 938
Joined: Tue Feb 09, 2010 4:54 pm

Re: When is it cause, not correlation?

Postby suspicious android » Sun Jun 26, 2011 12:29 am

skip james wrote:But I'm biased. I like Kant, regardless of whether or not he ultimately succeeds. He's got heart, plus he's probably a nice guy, having invented the categorical imperative and all.


I read that he was a pretty nice guy, and apparently a beloved professor, which is pretty atypical for big deal philosophers. I'm sure there's lots to be gained from a thorough, years long study of his works, but . . . I haven't even seen season 3 of Breaking Bad yet, I gotta prioritize.

skip james
Posts: 264
Joined: Sat Sep 19, 2009 2:53 am

Re: When is it cause, not correlation?

Postby skip james » Sun Jun 26, 2011 2:31 am

mad disrespect.

how can you not realize that season four starts in a matter of weeks?




Return to “LSAT Prep and Discussion Forum”

Who is online

Users browsing this forum: BobBoblaw, Yahoo [Bot] and 4 guests