Topic: Episode #257

[MountainManPan]

Anyway, it isn't my "claim".  Even the American Heart Association has cottoned onto restricting sugar to reduce the risk of heart attack do to coronary artery disease and the American Diabetes Associaton now says that carbohydrate restricted diets might be good for some.  It's pretty mainstream now.

isn't that a different claim?

Yeah, completely different.  I wanted citations for the claim that fat isn't actually bad for you, so when the AHA and the ADA say sugar is bad for you, that's not in any way relevant.  I know sugar is bad for you.  I never argued against that.

So yeah, unless and until you actually come up with some real references for your fat claims, I'm going to assume you don't actually know what you're talking about, even if you do happen to be right by accident.

it always seems that people will aruge that fat is horrible for you, or that sugar is horrible for you.  but in the end, it does just seem to come down to keeping your caloric intake at a level that maintains a healthy wait.   that is more important than what percentage of your calories come from fat or carb or protein

[kem]

One point we seem to be missing here: how much of a good idea is it to be over-fussy about heart health if the alternative to a quick heart attack is a long drawn out death by cancer, stroke or dementia?

But heart healthy (well at least a good way to have a good blood lipid profile) is so tasty and easy.  Meat, fish, cheese, vegies ... and lots of it!  And enough exercise to keep you active enough  to have a fatal accident rockclimbing, backcountry skiing or cycling at 90.  That's the plan, anyway.

[Trinoc]

One point we seem to be missing here: how much of a good idea is it to be over-fussy about heart health if the alternative to a quick heart attack is a long drawn out death by cancer, stroke or dementia?

But heart healthy (well at least a good way to have a good blood lipid profile) is so tasty and easy.  Meat, fish, cheese, vegies ... and lots of it!  And enough exercise to keep you active enough  to have a fatal accident rockclimbing, backcountry skiing or cycling at 90.  That's the plan, anyway.

Fine, if that was how I want to spend my life. But if I don't ... ?

[kem]

One point we seem to be missing here: how much of a good idea is it to be over-fussy about heart health if the alternative to a quick heart attack is a long drawn out death by cancer, stroke or dementia?

But heart healthy (well at least a good way to have a good blood lipid profile) is so tasty and easy.  Meat, fish, cheese, vegies ... and lots of it!  And enough exercise to keep you active enough  to have a fatal accident rockclimbing, backcountry skiing or cycling at 90.  That's the plan, anyway.

Fine, if that was how I want to spend my life. But if I don't ... ?

Fine.  I prefer to be heathy and active.  I'd like to think I can manage the Andy Weil "compressed morbidity" thing. 

[lippard]

Re: cursed cell phone numbers--local number portability began in the U.S. in 2003--don't know about Bulgaria.

Re: causal fallacies--I think most people who say that "correlation implies causation" is fallacious have in mind the claim that correlation *deductively* implies causation, which is indeed a fallacious inference.  It's not (or at least not necessarily) fallacious with respect to inductive implication, or implication construed more generally than deductive or material implication.

[PsyStat]

Can someone explain precisely and unambiguously what "correlation does not equal correlation" means?  After years of reading/hearing this, I still don't understand it.  I'm less concerned about lay usage of this phrase and more interested in what someone might mean by it in a serious discussion of research methodology, statistics, or epistemology.

To help explain my confusion, here's my view of why "correlation does not imply causation" aptly expresses caution against inferring direct/proximal causation from the existence of a correlation (as I've told my stats students): The fallacy in question commonly arises when someone observes that two events (or variables, etc.) are correlated (i.e., occur together over subjects, times, places, etc.) and concludes from this that one event causes the other.  In other words, he reasons that if the events are correlated then one causes the other; that is, he thinks the events' correlation implies their causal connection.

To correct this fallacy, it seems appropriate to assert that two events' correlation does not imply their causal connection -- that is, correlation does not imply causation.  Is there a similar explanation of what "correlation equals correlation" or "correlation does not equal correlation" means?

To describe this a bit more succinctly, let's define two events, A and B, as statements about some relationship between random variables X and Y:

A: X and Y are correlated
B: X directly causes or is directly caused by Y

(To be more rigorous, I should define "correlated" and "directly causes" more precisely.)  It's easy to pick a counterexample to show that the implication A ==> B is false; for instance, suppose a third variable, say Z, causes both X and Y and induces a correlation between them.  Hence, A doesn't imply B.  Here again, I don't know what statements such as "A equals B" or "A doesn't equal B" mean, because it seems unusual to make statements about the (in)equality of events A and B.  Can someone explain this to me?  Is "equals" just short for "is equivalent to"?

BTW, Philip Stark's online stats textbook (SticiGui) -- especially Ch. 2, 12, and 13 -- is a rare example of logic and logical fallacies integrated into intro stats material:

  www.stat.berkeley.edu/~stark/SticiGui/index.htm

[seaotter]

But that is usually the point of a correlational study. I hypothesis that x causes y. So I do a correlational study. It shows a relationship. That is evidence consistent with my hypothesis. It implies but does not establish causation. I think you have to have some idea of a mechanism to really imply causation. Many correlations are reported without such a hypothesis of a mechanism of causation.

[Trinoc]

Can someone explain precisely and unambiguously what "correlation does not equal correlation" means?  After years of reading/hearing this, I still don't understand it.  I'm less concerned about lay usage of this phrase and more interested in what someone might mean by it in a serious discussion of research methodology, statistics, or epistemology.

To help explain my confusion, here's my view of why "correlation does not imply causation" aptly expresses caution against inferring direct/proximal causation from the existence of a correlation

I think the problem is as simple as that: what really should be said is "correlation does not imply causation", but sometimes it is mistakenly stated as "correlation does not equal causation"* -- which of course it doesn't since they are different concepts.

[* I'm assuming this is what you intended to type.]

[gmalivuk]

Well for "equal" we could substitute "is logically equivalent", in which case correlation implies causation is Co -> Ca, whereas correlation equals causation is Co <-> Ca.  Neither of these are true, so it is equally correct to say "correlation doesn't imply causation" as "correlation doesn't equal (i.e. isn't logically equivalent to) causation".

[Caffiene]

im·ply   {im-plahy}
–verb (used with object),-plied, -ply·ing.
1.
to indicate or suggest without being explicitly stated

As Steve has discussed, correlation is one form of evidence that may suggest causation. It is a valid observation that can be part of a set of evidence for causation. I think its, generally, sensible to say that two correlated phenomena have a greater chance of being causally linked than two completely uncorrelated phenomena.

-> Correlation implies causation.

But of course not all cases of correlation are causative, so correlation is not logically equivalent to causation. Which is the fallacy: believing that correlation by itself is proof or overwhelming evidence for causation.

-> Correlation does not equal causation.

[GodSlayer]

Massimo would be a good guest for niggly little philosophy of science segments like that.

[Trinoc]

Well for "equal" we could substitute "is logically equivalent", in which case correlation implies causation is Co -> Ca, whereas correlation equals causation is Co <-> Ca.  Neither of these are true, so it is equally correct to say "correlation doesn't imply causation" as "correlation doesn't equal (i.e. isn't logically equivalent to) causation".

"Correlation does not equal causation" is a weaker assertion than "correlation does not imply causation", since the latter allows for the possibility that correlation does imply causation if only causation does not imply correlation.

Of course, if causation does imply correlation then the distinction becomes unimportant. I'm trying to think of a situation where there is causation but no correlation. Does anyone know of an example?

[quirk3k]

"Correlation does not equal causation" is a weaker assertion than "correlation does not imply causation", since the latter allows for the possibility that correlation does imply causation if only causation does not imply correlation.

Of course, if causation does imply correlation then the distinction becomes unimportant. I'm trying to think of a situation where there is causation but no correlation. Does anyone know of an example?

I was thinking the same thing but you beat me too it.

To answer your question, two examples come to mind. The first example is multiple causes. If A causes X, then it may also be the case that B and C cause X also. In this case A may not correlate to X. Another example is combined cause. If A, B and C are all required to cause X, then A by itself may not correlate to X.

[Dolomedes]

OK, it's not called "dark laser",  but laser cooling is fact!

http://en.wikipedia.org/wiki/Laser_cooling

Will the SGU come back to this one, again? I'm expecting nothing but a full retraction of this installment of SoF, and annullation of the results. 

[relic]

For anyone visiting Tokyo, there's a very cool show about the Hayabusa mission showing at a brand new state-of-the-art planetarium in Shibuya http://www.asahi.com/english/TKY201012080294.html

Went there with my family last weekend. Even if you don't speak Japanese, you'll probably get most of the story and the visuals are more than worth the 600 yen ($6) admission.