chula, thanks, I stand corrected!
Nowhere in Wand’s reply does she state cherry-picking is not permissible, only that it is abused. Yet you imply she doesn’t understand this.
I think Wanda is aware of that, she like myself questions just how ideal a tool it actually is. Here you say it is “thought” to be an ideal tool which doesn’t convey all that much certainty. Earlier you stated, “A discussion about its validity as a technique is reasonable…” Given your own words, I wouldn’t be so quick to say she doesn’t quite understand.
I appreciate the meta-analysis math lessons you provided, but you still haven’t answered my question concerning the ETS tables I provided. Maybe I’m not making myself clear. We disagree on the number of studies that show a positive risk, you say there are many. When I follow the instructions for interpreting the studies, I find there are few. Now either you find something wrong with the instructions for interpreting the tables or you are doing something in addition to the instructions to come to your conclusion. It has been made clear to me that you couldn’t have pooled any of the studies without complete information on the criteria for each study which those tables do not provide. What I want to know is what procedure did you follow as you went down the list of studies that led you to conclude that many of the studies show risk.
Thank you, I’ll check it out.
Okay. Here is where I think that she either doesn’t understand or chooses not to …
It is not Clintonesque to say “it depends on how you define ‘statistically significant.’” Each study was statistically significant at a 90% confidence interval, but not at a 95% confidence interval. Each individual study suggests an increased relative risk but do not prove it to a 95% confidence level. Meta-analysis is a generally accepted tool to allow pooling of this data. A simple and extreme example-
John has suspicion that Phil’s coin is rigged to only come up heads. He does an experiment and flips Phil’s coin 4 times and all times it comes up heads. Phil says, “Well that is not convincing. That could happen by chance one out of 16 times.” “Yeah,” says John, “one out of 16 isn’t enough for me to kill you. It is less than 95% confidence.” “Uh huh” says Phil, “no statistical significance.” Joe comes by and does his own experiment and flips Phil’s coin another 4 times. Again 4 heads. Again Phil says, “But that could happen by chance one out of 16 times.” Joe is also reasonable; one out of 16 isn’t enough to prove that Phil is a lying scumbag cheat. Alice saw the whole thing. Neither experiment was enough to be convincing but she combines the data and says that one out 256 is conclusive and slugs Phil. She came up with convincing data out of two smaller and thus unimpressive studies. Meta-analysis is how researchers do that. My problems with the technique are in the other thread.
DSied, No matter how large a study, pooled or otherwise it will have a range of error. A result that falls within that range is “statistically insignificant.” This is how Forces explains it and I accept it. If there is a an official term, let me know so I can share it.
I have no problem whatsoever with pooling the results of flipping a coin, there’s only one factor involve which is a human initiating the flip. There are no other factors to be considered like age, race, gender, genetics, height, environment, or maybe a factor not even considered. Now let’s change the criteria, let’s say I have a machine in a controlled environment that is accurate enough to flip heads every time. Can Alice pool that study with the human ones and get an accurate result? Of course not. She would not be justified in killing Phil.
As I see it, the greater the number of factors that could possibly influence the outcome, the less likely one can know for certain which factors and to what degree they influence the outcome. I highly doubt that any ETS study can take ever single factor into account, so each study tests using its own criteria. Therefore, only those studies with the same criteria can be pool and as Wanda pointed out, "The likelihood of epidemiological studies being similar enough in design to be able to lump their results together and then do some statistical manipulation to come up with a sort of “average” relative risk is VERY small. Without knowing the criteria used for each of those EPA studies, you’ll have to evaluate them on their own merit.
DSied, been thinking and the definition I sited for “statistically insignificant” was from memory and it may be wrong. It may be within a certain percentage of uncertainty… not error. I left a message over there asking for the exact definition. I don’t want to make them look bad on my account. I’ll get back to you on it.
Peter, you really do not need to have every factor the same in order to do a meta-analysis. That is not one of its limitations. Her argument really is very specious. The odds that epidemiologic studies are similar enough to use in a meta-analysis is very large. She’d be the one in the coin flip story who came after another guy flipped the coin five times. The guy says, “Ha! This proves it. Only one out of 32 chance of this being by chance, that’s greater than a 95% confidence interval!” Wanda would respond “Nah. Two studies showed no statistically significant findings and only one study said there was. Two out of three! We can only conclude that this is a fair coin.”
Don’t worry about making Forces look bad. I’ve seen their site. They do a fine job of looking bad all on their own.
As I see it, determining which studies are similar enough is just another human factor that can bias the result. No wonder the results for ETS studies vary so much.
Once again, I have no problem with flipping a coin. I do have a problem with multi-factorial studies and the determination of what is relevant or not. Your example means zilch to me.
I don’t recall your having a problem using the ETS studies they provide. You accuse them of being hyperbole while the anti-smoking propaganda campaign poisons the airwaves is not with their insulting rat and doll ads. What a laugh. Your are about as biased toward Forces as the epidemiologists who are paid to do a job. It is obvious experts in the field have lost credibility when both sides of the smoking issue or any other issue can accuse the other of foul play.
BTW, do a net search for “statistically insignificant,” you’ll find thousands of references to it as it is used by statisticians. So don’t try to hoodwink me into thinking there’s no such thing.
I think I will end this rather useless dicussion with you right now.
These debates are always so ugly. The evangelical fervor of the anti-smoking lobby is outdone only by the violent bitterness of committed smokers. In my opinion, smokers win the “most irrational” award on points.
I remember seeing an ad a few years ago sponsored by some tobacco company or other showing a group of people standing outside smoking in a torrential downpour with the caption, “What Kind of Policy Makes People Stand Out in the Rain?” Everyone I know who saw this ad had the same reaction. “What kind of product makes people stand out in the rain?”
The issue isn’t “Can second hand smoke be bad for you?” Of course it can. It is completely specious to argue otherwise. First, note that the vast majority of second hand smoke isn’t “filtered” through a smoker’s lungs. It comes directly off the end of a burning cigarette. It is now (finally) accepted by even the tobacco companies that smoking causes lung cancer, heart disease, etc. If second hand smoke is, somehow, completely harmless, all these adverse health consequences must be caused by the bloody filter.
The issue is “How bad is second hand smoke for you in what concentrations?” Catching a whiff of smoke from across the street isn’t going to lower your life expectancy. Working in a smoke filled club for twenty years probably will. Whether, and at what level, this should be regulated is ultimately a policy question, not a scientific one. Most of the arguments in this thread, for example, don’t really turn on what level of increased disease is an acceptable trade off for allowing smoking in bars. They turn on a philosophical position that suggests banning smoking infringes personal liberty regardless of the health consequences.
As to why smoking was generally acceptable thirty years ago in the U.S., let me offer this anecdote. I was in Turkey for a few days in the winter. Now in Turkey, many people smoke, well, like Turks. Smoking on a fumigatory scale is largely inescapable. At first this bothered me quite a lot but I did get somewhat used to it, though I still couldn’t really handle being in a closed room with a serious smoker.
On my return, I bunged a jacket I had been wearing into a closet. About a week later, I opened the closet and literally staggered back across the room. That jacket stank. I do not mean that it smelled bad, I mean that it made your eyes water from fifteen feet. I had serious doubts as to whether the dry cleaner would even accept it.
The point here is that while in Turkey I couldn’t smell tobacco smoke. The smell was so pervasive that it simply faded into the background. But once I got back to a relatively smoke-free environment, I could smell it, and how.
Thirty years ago in the U.S. when smoking was very common, even non-smokers couldn’t really smell it. Now, however, they can. I can sometimes smell someone smoking upwind from me in the U.S. (especially in California) from a block away. Regular smokers, who can’t really smell tobacco smoke, tend to forget this. As a result, they equate complaints from nasally-proficient non-smokers as nothing more than an effort by busybody non-smokers to stick their noses where they don’t belong. In reality, it’s quite the opposite.
Truthseeker, you seem to be making the same argument that other anti-smokers on this thread have been making. Namely, that it doesn’t matter whether there’s any evidence that second-hand smoke is bad for you because non-smokers should not have to be subjected to the smell. It is ridiculous to suggest that one group’s annoyance is a valid basis for outlawing an activity.
I’m going to try for a recap here …
Argument #1 - Is ETS dangerous?
You have to
-say that meta-analysis is junk science, despite the fact that it is a widely accepted technique, and/or
-accuse both the EPA and the WHO as being part of a huge Nazi-like oppressive Big Pharma-bought enslave-the-smokers conspiricy
to not accept that there is a significant level of health risk to nonsmokers from ETS. You also have to avoid common-sense. (How much risk from how much ETS is harder to quantify.)
Argument #2 - Do one set of individuals have a right to engage in behavior that puts another group at risk? Is there a level of risk that is acceptable to impose upon others? Is it acceptable to impose nuisance behaviors upon others? Is there any limit to how much annoyance is allowed?
These are related questions and it is silly (IMHO) to argue in absolutes about them. There is some level of health risk that is insignificant even if real. And there is some level of annoyance that is unacceptable to allow (Going around naked, for example). The question is defining where those lines should be drawn and where ETS falls in relation to those lines. Not surprisingly smokers don’t find ETS all that annoying and don’t see the risk as all that huge. (chula, I doubt that you have a problem with outlawing your neighbor from blasting his stereo with open windows at 3 am … we do outlaw annoyances that are annoying enough). Not surprisingly a nonsmoker is more likely to believe that ETS’s risk are significant than is a smoker.
Argument #3 - Do workers have rights to be protected from dangers at work or do they accept them by taking the job? When should the government be involved in enforcing such protections, if ever?
Hmmn… As for Argument #1, there seems to be a sort of “all-or-nothing” approach in your recap. I wouldn’t say that meta-analysis is junk science, but rather that it is a flawed technique and that those flaws vary in importance according to what they purport to study. Also, I wouldn’t demonize the EPA and the WHO as per your recap, but I do think they have biased agendas.
As for Argument #2, I feel that non-smokers use flawed ETS studies to back up their self-righteous annoyance factor. I’ve observed non-smokers happily clustered around charcoal grills and woodfires; and noted that none of them seem outraged by exhaust fumes nor have stopped driving because of them.
Argument #3 I file under “informed consent and accepted risk.” I worked for years in travel sales and wore a headset. I was warned during training and periodically thereafter that the headset might affect my hearing in the long run. I accepted that risk. My father burned structural steel with a cutting torch and accepted the risk of getting burned by flying sparks. Similarly, I think waitstaff who work in a bar or the smoking section of a restaurant accept the risk (if there is any) of ETS.
I think the employer has the responsibility to inform employees of potential risks and provide for safety devices and safe practices when possible (in the case of bars and restaurants, a good air filtration/ventilation system). I think the employee has the responsibility to either accept the risks and do the job, or find a different job.
Not at all. I’m merely pointing out one of the reasons for the disjunct between smokers and non-smokers.
Non-Smoker: Could you put that out? It’s really bothering me even from across the room.
Smoker: You’re just a prissy health nazi! Cigarettes don’t really bother people that much. Thirty years ago, no one minded at all.
**
You can make this argument, if you like, but it doesn’t reflect the common practice in most western countries. Try raising a large flock of chickens in New York City if you don’t believe me.
All kinds of activities are outlawed or regulated because they are “annoying.” The fact is that as a smaller and smaller percentage of the population smokes, it will be more and more reasonable to regulate smoking in public places. To take an extreme example, if one person out of a hundred chain smokes cigars, is it unreasonable to prohibit chain smoking cigars in movie theaters?
Yes, Truth Seeker, we all realize that it’s common to outlaw or regulate things that the majority finds annoying. So what? It’s still ridiculous. If I can convince 51% of the voting public that wearing pink is annoying, does that mean that I should get wearing pink regulated? Certainly I could, but should I?
In fact, as the percentage of smokers decreases, smoking probably will be increasingly regulated; that this is so does not in my opinion make it reasonable. Would it, in your opinion, be reasonable to decreasingly regulate smoking as the percentage of smokers increases?
Not exactly a flock, but…
BTW–“chain-smoking cigars”? Indeed. Learn something about the evils you deplore before you inveigh against them, please.
Yes, of course. The “cost” to smokers of second hand smoke is almost nil. Even putting aside potential health risks, the “cost” to non-smokers of second hand smoke is relatively much greater. Therefore, if everyone smokes, the social “cost” of second hand smoke is low, especially when balanced against the “value” smokers get from being able to smoke in restaurants, etc.
If, however, only a few people smoke and if everyone else is bothered by second hand smoke, the cost to society is high, especially when considered relative to the value those few smokers get from smoking in places like restaurants. It is, therefore, more reasonable to regulate second hand smoke the fewer people there are that smoke.
Anyway, don’t you think “ridiculous” is a bit strong? You have a particular philosophical perspective. However, that doesn’t render all competing perspectives inherently absurd. In any case, if you think about it, I’m sure that you, yourself, would find it reasonable to regulate some types of annoying activities. There is nothing inherently illegal about raising pigs – it’s even, arguably, socially beneficial to do so. Nonetheless, wouldn’t it be reasonable to prevent someone from starting a hog farm in the middle of suburbia?
This all raises an issue someone referred to earlier: market failure. Once again, putting aside the health issues, I’m not necessarily convinced that outright banning smoking is the way to go. I’d like to see an experiment that outlaws smoking/non-smoking sections so that restaurants and bars have to choose to be either all smoking or all non-smoking.
Squish Sorry, I don’t take your point. I was looking for an outrageous example that even most smokers would find annoying.
Fair enough. I would actually say that even if the majority of people are smokers, it would be reasonable to regulate ETS to some extent, but I applaud the consistency of your stance.
As to the “ridiculous” comment, I’m not so sure it is too strong. Certainly it’s true that not all competing perspectives are inherently absurd, and most of them aren’t, but some of them, I feel, are. I just happen to think that a stance that it’s okay to regulate something once it becomes unpopular merely because it is unpopular is one of the latter, that’s all.
I would be interested in seeing the experiment you mentioned, and even more interested in a situation where restaurants and bars can choose to be all smoking, all non-smoking, or with separate sections where the separation between those two sections is real separation and not separation in name only. I, for one, would have no real problem with telling business owners that if they want to have separate smoking and non-smoking sections, the sections in question have to meet certain air quality criteria, and then leaving it up to the proprieter whether he wants to go all smoking, all non-smoking, or some mix.
That’s not a very convincing example, since until about a month ago I woke up every day in my Manhattan apartment to the sound of roosters crowing. Really! ") Should I have called the police? Should those people have been fined? Why the hell should I care if they have roosters?
I believe there are some actual health reasons why you’re not supposed to raise farm animals in a city. The other problem with your analogy is that none of us are arguing that there should be no regulation of smoking. We’re saying that smoking should be allowed (if the owner chooses) in one particular type of establishment. Just don’t visit a farm and then complain about the smell of the pigs. And don’t tell the farmers they shouldn’t be allowed to raise pigs because it might endanger their health.
I must be missing something here: is there truly anyplace that demands a restaurant or bar provide a smoking section?
TruthSeeker, you were looking for an extreme example–keeping chickens in the city wasn’t it. 
The smokers I know seem to be a lot more tolerant of ‘annoyances’ than the non-smokers, for some reason. I detest the smell of charcoal grills–do I want to ban them? No. I don’t like people wearing too much perfume in enclosed areas (the workplace being a good example)–do I want laws against that? No. Screaming baby laws? No.
As far as I know? No, there’s not. What Truth Seeker is interested in seeing, if I read him correctly, is a situation in which restaurants must be all one or all the other but can choose to be which. I suspect that the vast majority of them would choose to be all nonsmoking rather than all smoking, and I think it would be useful to offer them a third choice: have both but do so in such a way that the nonsmoking section really doesn’t get smoke.
I don’t really think this is the best way to resolve this debate, mind you, but it would provide some interesting information nevertheless.
And to be fair, of course, we shouldn’t be discussing this as if ETS was irritating and obnoxious but harmless. If it were, I’d say that it’s something people should just deal with, but it’s not.
Charcoal briquettes and perfumes are not harmless, to some people.