How well do restaurant patios and smoking coexist?

I wouldn't base too much on that post. The first example didn't reflect any rigor at all and was more about pharmaceutical self-justification than epidemiology. The one about Republicans wasn't epidemiological at all, but sociological. The other two may be, in fact, instances of poor epidemiological studies, but that's a pretty thin sample to use to castigate the whole field.

A quick google search will demonstrate that there is a lot of self-critical analyses of epidemiology.

[Robin Garr]

For those interested, here's a actual statistician that I recently came across who is likewise skeptical about epidemiological studies.

Kind of cool. Thanks!

My analysis of this, however, is that it is largely a satirical effort to remind us that correlation does not equal causation, which expresses the same idea in a lot fewer words.

I don't think any sensible person denies this reasonable assumption. At the same time, though, correlation can certainly highlight possibilities. A lot of people who smoke get lung cancer. So do a lot of people who live around sources of radon. People who drink a lot of HCFS get fat. So do people who drink a lot of beer. These all point to ground that looks ripe for further research, at a minimum.

Simply yelling "correlation does not equal causation" doesn't prove a hypothesis wrong, it just reminds us that there's more study to be done before the paper gets published.

And "correlation does not imply causation" is actually a core principle of epidemiology.

Epidemiological studies can never prove causation; that is, it cannot prove that a specific risk factor actually causes the disease being studied. Epidemiological evidence can only show that this risk factor is associated (correlated) with a higher incidence of disease in the population exposed to that risk factor. The higher the correlation the more certain the association, but it cannot prove the causation.

http://pmep.cce.cornell.edu/profiles/ex ... ology.html

Epidemiological and sociological studies use many of the same statistical methods. So, they can be subject to the same flaws and blindspots.

And it's more than just "correlation is not causation". It's about not even knowing or measuring the independent factor before claiming there's a correlation with the independent variable, which is one of the main sins of the study discussed in the link that I posted. It's also about cherry picking data. And it's about meta-analyses begging the question by over weighting studies and data that agree and under weighting those that do not. It's about journals that accept such studies even though they the know the methodology is poor, because everyone else does it too. It's also about out and out fraud that can go on for years.

Faith in institutional science is touching. But like any human endeavor it falls short of perfection, and many of the feedback and self-correction mechanisms that are supposed to be apart of the scientific method are breaking down. And like much of our modern society, it's breaking down along lines of ideology and tribal type politics and affiliations. You don't have to understand or even listen to the science of the "wrong" people. But all the "right" people get to peer review each other papers. True scientific integrity is hard to find. Where is Diogenes when we need him?

Smoking sets up perfectly for this. Cigarette companies make perfect villains. Most smokers are rude and icky. So, it's *good* for "scientific" studies to show that second hand smoke is bad, and it's *bad* to show that there's little problem. So, most studies that want to show that second hand smoke is benign don't even get approved. And if a researcher should innocently reach a conclusion that it is benign, then his research grant is mysteriously not renewed.

Politicized science is the worse science. And and science paid for by the government to regulate our lives might be the most politicized of them all.

Another ideology free post from Steve.

And "correlation does not imply causation" is actually a core principle of epidemiology.

Epidemiological studies can never prove causation; that is, it cannot prove that a specific risk factor actually causes the disease being studied. Epidemiological evidence can only show that this risk factor is associated (correlated) with a higher incidence of disease in the population exposed to that risk factor. The higher the correlation the more certain the association, but it cannot prove the causation.

http://pmep.cce.cornell.edu/profiles/ex ... ology.html

So, when somebody suggests a bunch of epidemiological studies to support some new law or restriction, you will join me and say "This doesn't prove anything!" okay?

And if someone on a message board points out that epidemiological studies have flaws and limitations, then you won't bust out with teh sarcasm?

Another ideology free post from Steve.

Teh Sarcasm! It burns!

And meaning don't listen to Steve. He's the "wrong" kind of person. Right?

And "correlation does not imply causation" is actually a core principle of epidemiology.

So, when somebody suggests a bunch of epidemiological studies to support some new law or restriction, you will join me and say "This doesn't prove anything!" okay?

That question demonstrates that you do not understand the subject at all, in my opinion. Epidemiological studies cannot prove that I, a lifetime pack a day smoker, contracted lung cancer from smoking. Cannot do it. It can establish, with more or less certainty based on the data, that smoking a pack a day of cigarettes can make me more likely to contract lung cancer than if I lived the identical life without smoking. And when that is spread out over a population, people can make judgments on the best course of action. It is not certainty or proof, but it is more solid evidence for outcomes than we have for most of the decisions we make in life.

And if someone on a message board points out that epidemiological studies have flaws and limitations, then you won't bust out with teh sarcasm?

No one has argued that their are not limitations (see above). But to move from limitations to untrustworthy is a function of your ideology (I say based on inferences I have made from your posts, not from epidemiological data or laboratory demonstration).

That question demonstrates that you do not understand the subject at all, in my opinion. Epidemiological studies cannot prove that I, a lifetime pack a day smoker, contracted lung cancer from smoking. Cannot do it. It can establish, with more or less certainty based on the data, that smoking a pack a day of cigarettes can make me more likely to contract lung cancer than if I lived the identical life without smoking. And when that is spread out over a population, people can make judgments on the best course of action. It is not certainty or proof, but it is more solid evidence for outcomes than we have for most of the decisions we make in life.

I understand the subject well enough. Here's the thing though. You mention people making judgement. If these studies were used be people to make decisions about their own lives, then there would little for us to argue about.

The problem is that government bureaucrats pay for many of these studies, and then use the results not to allow folks to make there own decisions, but to enact laws and regulations that that the decisions away from the people effected.

Studies that we both know don't "prove" anything are used everyday by the government enforce their preferred policies onto people whether they like it or not. You may disagree, but I think there is a huge amount of confirmation bias in government funded studies. So, take studies that can't prove anything 100%, add the confirmation bias, and you've got government bureaucrats using the fig leaf of science to do whatever the heck they want to do. It's easy for you to trust them, because they usually agree with your point of view.

And if someone on a message board points out that epidemiological studies have flaws and limitations, then you won't bust out with teh sarcasm?

No one has argued that their are not limitations (see above). But to move from limitations to untrustworthy is a function of your ideology (I say based on inferences I have made from your posts, not from epidemiological data or laboratory demonstration).

Maybe. But why should I trust them if I don't trust the government bureaucracy that has them bought and paid for. Science is *supposed* to be about skepticism. When did that stop?

Look, I recognize when I am being snarked at and belittled. I might not be a genius, but I'm smart enough to see that. It's just part of the bias that you are enmeshed in. You see my posts as ideological. And you see your posts as being fair and unbiased. You are the defender of impartiality and justice! I'm am the racist violator of Godwin's law!

The thing is, you are are just as ideological , if not more so. You just can't see it. You just assume that you are the reasonable one and go from there.

I'm being snarky because you keep talking about liberal confirmation bias, then you write, "But why should I trust them if I don't trust the government bureaucracy that has them bought and paid for." Your posts demonstrate (or at least suggest) that your deep skepticism over science is because it tends to promote actions that go against your political positions, therefore it must be biased. And I can find two studies that prove that (of course, by the rules of the game, one couldn't prove it at all, but that is another matter).

Now I believe one should always follow the money. I don't trust any human endeavor absolutely.

But to quote the great philosopher Steven Colbert, "Reality has a well-known liberal bias."

Does all of this mean I can stop wearing the damn seat belt in the car? Talk about freedom killing legislation! Let's make a deal, you smokers can have at it wherever you want if I can stop wearing the seat belt. Unlike your deadly second hand smoke, my seat belt only inconveniences and potentially kills me. I can live with that. Do we have a deal? Yes, yes?

[Robin Garr]

t if I can stop wearing the seat belt.

Reminds me of the gent I saw riding out Frankfort Avenue the other day, darting back and forth in front of traffic. Wearing shorts; no shirt, no shoes, no helmet. He was riding no-handed, and talking on his cellphone.

Call for Doctor Darwin!!!

But here's the funny thing. If he had been hit by a car and squished on the payment, it wouldn't really prove that riding no-handed and distracted without a helmet on correlates with death in traffic. It's just anecdotal, after all.

Does all of this mean I can stop wearing the damn seat belt in the car? Talk about freedom killing legislation! Let's make a deal, you smokers can have at it wherever you want if I can stop wearing the seat belt. Unlike your deadly second hand smoke, my seat belt only inconveniences and potentially kills me. I can live with that. Do we have a deal? Yes, yes?

Long as you don't have any insurance that compensates your and/or your family in case of permanent disability or death paid for by the premiums of those who didn't make that choice.

Segregation was the law, passed by a majority of people, and we know what is on the menu when two wolves and a sheep vote on dinner...

Good stuff, Adriel!

I'll make an important distinction here, though, although I see that it could be argued on either side of this debate: Segregation was the law, but it was the law in spite of the majority. Whites were the minority in the Jim Crow South, as they were in apartheid South Africa. But they had property (ha, Ayn!) and they had power, and they abused it to keep the majority under control and keep them away from the ballot box.

Yeah I can see that. Same goes for current problems with corporate money, making minorities with power lord over us wee ones who could out vote them on paper, if only we weren't sucked into their political machinations that use media, and the cudgel of propaganda to herd us towards their bad ideas.

In my book the big no no is violating the non-aggression principle. It's the axiom that states you can't initiate or threaten violence against another or their property. Like the political golden rule. So anyone who does it, no matter how well intentioned is suspect or wrong. Governments are often big violators and so therefore you get people riled up when you say you don't like the moral justifications for use of violence behind big ideas like civil rights. NOT that you don't want everyone to be able to do all the things they wish in society, but that no matter how wrong someone is, telling them to not hate others while sticking a gun in their face is counter the spirit of what you are trying to engender in them. It's like using an electric chair to teach people not to kill.

Does all of this mean I can stop wearing the damn seat belt in the car?

Under my administration Ronnie you could smoke your car and drive wearing a cigarette... All things are possible at Anarchist Disney World.