A Climate Soap Opera in the Headlines
Bud Ward, the editor of the Yale Forum on Climate Change & the Media, weighs in on the Heartland billboard furor:
What stands out amidst the initial widespread revulsion is that the criticisms of Heartland’s effort came not only by the usual cadre of what climate skeptics dismissively call “warmists,” but also by those ideologically in synch with the group. In addition to their disgust with the message, of course, came their disappointment that the billboard had handed Heartland’s many adversaries a useful weapon of criticism.
Indeed, this disappointment is quite evident in the comment threads at various blog sites. I think I read one person ruefully lament that there was no way to unring this bell. And Heartland can’t stop the bell from ringing, either, because it has thus far refused to acknowledge its blunder, much less apologize. It seems it will get worse for them before it gets better (if at all).
On an unrelated note, I have a short piece up today, as well, at the Yale Forum. It’s about an enterprising journalism project that I’ll be watching with much interest.
Lastly, also new at the site is discussion of the “striking” results of a recent study that found six distinct climate change storylines that have played out on broadcast television news the last decade. It’s an important data point for climate media scholars and everyone else interested in how climate change gets translated on TV. As the piece explains:
The premise of the study is that we humans are, whatever else we are, story-telling animals: We make sense of, form beliefs about, and establish our stances on issues such as climate change less on the basis of reason or experience and more on the basis of the stories we subscribe to. Moreover, the news media are, whatever else they are, purveyors of story, always on the lookout for a narrative angle that will capture the attention of viewers, listeners, or readers.
That’s right. Journalists are always on the prowl for a good story. People who accuse us of left/right/center bias should never forget that. What we care about, above all, is the story.
That also explains why the current Heartland saga or Climategate, or the latest weather disaster are always natural story fodder.
I agree that it was over the top for HI to compare global warming believers to radical bombers.
Which is why I am surprised that the Chicago Green Festival featured a presentation yesterday by two unrepentant radical bombers- William Ayers and his partner from the Weather Underground movement, former FBI most wanted Bernadine Dohrn.
The pair were instructing environmentalists on how to unleash the radical. For this pair that meant killing people and blowing up buildings.
http://read.uberflip.com/i/61404/19
Interesting graphs in that article about media balance. It would appear that Fox covered ‘climate tragedy’ most up to 2006, then showed a big jump in the ‘hoax’ stories, but at roughly the same rate as the continuing ‘climate tragedy’ stories. Almost as if they were trying to be fair and balanced.
Fox News = fair and balanced. Discuss.
🙂
climate media scholars ?!
Seriously.
The internet blog world gives every niche an “Inside Baseball” forum to rant and rave about the end of the world. The truth is, most people don’t give a flip about 10:10, HI, or Hockey Sticks. Therefore, it is irresponsible for you Keith, as a serious journalist, to compare the latest kerfluffle with Soap Opera’s. Millions of people actually pay rapt attention to Soap Operas. I expect a full credited correction!
“What we care about, above all, is the story.”Not always ‘the’ story, but always ‘a’ story. Which is part of the problem.
I’m not disappointed. I never cared about these people and now I care even less.
HI has already recovered from the billboard incident. They have not done anything as egregious as supporting and furthering public policies that have retarded the nation’s economic growth, wasted billions of dollars, and contributed to thousands of people’s unemployed status.
No?
> […] not done anything as egregious as […]
A big But.
Not everyone agrees with Eli’s take on second-ghand smoke:
http://www.davehitt.com/facts/who.html
Another chicken test for kdk33:
Q:Why did the Heartland Institute came up with this smear campaign experiment
A:Because of battle fatigue. When you suffer from battle fatigue, sometimes you make mistake.
A related question:
How can you tell when an agent suffers from battle fatigue if that agent does not make any mistake?
That’s right. Journalists are always on the prowl for a good story. People who accuse us of left/right/center bias should never forget that. What we care about, above all, is the story.
And this is the self-serving story American journalists tell themselves. It is a favorite KK trope on this blog.
American journalists care about the story to a point, but not above all. Again and again, if there is a political angle, they care more about advancing the liberal narrative.
The most egregious recent example is of course the Trayvon Martin case, in which George Zimmerman was painted as a white vigilante killing a black child. Never mind that Zimmerman is some combination of hispanic, white and black heritage. Never mind that Martin is not an angelic small child as one frequently saw in photographs accompanying the story. Never mind that CNN incorrectly reported Zimmerman using a racially charged expletive (“f****** c***”). Never mind that NBC eventually had to fire a producer for selectively editing a Zimmerman quote in order to convey the false impression that Zimmerman’s interest in Martin was racially motivated.
There are nearly endless examples of journalists’ liberal bias far outweighing anything in the other direction. In fact, it’s an interesting “story” that we don’t find journalists investigating. Instead we get this spunky little narrative from journalists about how “they care about the story above all.”
Fail.
No again and no again and, of course, no but definitely no and some NO that bears on your link
Hey Eli, no disagreement that there are plenty of references in the lit to the dangers of second hand smoke. But, in your citations one report is from a company branded as “Secondhand Smoke Consultants” and another was from the “Ontario Campaign for Action Against Tobacco”. Do they have a bias? Maybe. If they have a bias does it make them wrong? Of course not. But, it would seem prudent to take their conclusions with a grain of salt.
Another of your cites is co-authored by Stanton Glantz PhD; in his article he talks about the Helena study on which he was a contributor……and not everyone is in agreement on the merits of that study:
http://www.davehitt.com/facts/helena.html
So, there are differing points of view out there, and who can claim to positively know the truth on the dangers of second hand smoke? But, I see nothing in your citations that looks as comprehensive as the WHO study, and that study didn’t show squat for a connection.
Matt. If you are so worried about Eli’s links than just go to Google Scholar, type in “second hand smoke” (pick a start year) and see for yourself. Be sure to report the results honestly.
Hey grypo are you worried about me worrying about Eli’s links? Plus a suggestion to use Google Scholar….is that you Hank Roberts?
No, only worried that you are spreading bad information and confusion and asked you to check the scientific record. But if you don’t want to, I’d ask onlookers to do so. The record is very clear on this subject, in several areas of health concern.
Hey grypo, I didn’t spread bad information. Neither did Eli; it’s just information. Intelligent people can look at all points of view & make up their own minds. As far as the record being very clear on the dangers of second hand smoke, there clearly are many studies done on it, there’s clearly not that many that show health risk with any statistical significance, and it’s clear that even in the studies that show the most danger the level of risk is in the “hit by lightning” category. It’s also clear that the second hand smoke boogeyman is used by nannystaters like Bloomberg in NYC to dictate lifestyle choices.
Sigh.
@ 20 Lazar – I hear you! Out of the 38 studies cited (most of which predate the more comprehensive WHO study) you have 10 where the confidence interval around the RR doesn’t include 1.0, so only 26% of these studies show statistical significance. Plus, only 4 (11%) show a RR higher than 3, and zero show a higher RR than 4. Here’s a quote to put that in perspective:
“My basic rule is if the relative risk isn’t at least 3 or 4, forget it.” – Robert Temple, director of drug evaluation at the Food and Drug Administration.
These studies certainly put the dangerous second-hand smoke issue into the “who really knows?” categoty……….
Ahhh, yes. The continuing confusions between what is written in magazines and what is science. Silly Rabbit..
Ah yes, the continuing confusion between wingnut blog commentary and the PRL.
Yes, BBD, name-caling is among your more accomplished skills. Sadly, there is little more. But carry on.
> The continuing confusions between what is written in magazines and what is science.
And what is science, kdk33?
The last No was there to point out that your friend Hitt did not tell the whole story, a great deal of your link was about a ruling by a Federal judge against the EPA which was reversed, and Eli was simply shocked that it turned out that before dressing in black, the judge had represented tobacco companies. Shocking.
The previous no was from the American Heart Association Journal Circulation, reviews cardio effects of second hand smoke. Plenty of references you can look up there
The next was a report of a presentation, which, with a little bit of googling skills refers to a paper Factors Associated with School Absence Among Children with Symptomatic Asthma, Akinbami Lara J.; Parker Jennifer D.; Merkle Sarah
Source: PEDIATRIC ALLERGY IMMUNOLOGY AND PULMONOLOGY Volume: 23 Issue: 3 Pages: 191-200 DOI: 10.1089/ped.2010.0013 2010
==================Overall, 59% of children with symptomatic asthma had >= 1 asthma-related absence in the past year. Among this group, presence of an adult household smoker was associated with an increased adjusted risk ratio for asthma-related school absence [adjusted risk ratio = 1.25 (110, 1.42)] as were race/ethnicity other than non-Hispanic white, low parental education, and reported health status less than “very good/excellent.” ==============
Shall we continue or are you going to stand there and proclaim more pap? Otherwise what grypo said
#26,
So that’s a risk ratio of 1.25, and not the 3 or 4 necessary to take a result seriously in an epidemiological survey?
http://xkcd.com/882/
Matt B,36/38 studies find a risk ratio of greater than 1. What happens to significance levels or confidence/credible intervals when you Bayesian combine those results?“”My basic rule is if the relative risk isn’t at least 3 or 4, forget
it.” ““ Robert Temple, director of drug evaluation at the Food and Drug
Administration.”It is probable that ‘Robert has a rule’ doesn’t classify as an argument. It is highly improbable that the mantra will win the internets. I see it repeated uncritically in the pro-smoking blogs with very little context and no discussion, unlike the following…“Robert Temple of the FDA is quoted by Taubes as saying:My basic rule is if the relative risk isn’t at least 3 or 4, forget it.However, Dr. John Bailar,from McGill,believes there is no magic dividing line.If it’s a 1.5 relative risk and it’s only one study and even a very good one, you scratch your chin and say maybe.It
is not size of the RR alone ( but we have to agree at some point low
is too low say 1.03 relative risk) but the results of other studies
addressing the same issue and concerns about biological plausibility
have to be factored in.”Thanks
#28,
The problem is that in epidemiological surveys, where you simply go out there and look at what’s happening, it’s virtually impossible to account for all the spurious confounding correlations. There’s a genuine correlation between the variables, but it’s not because of the effect under study.
It’s different in a randomised double-blind trial, where you have control over the presence/absence of the factor under study and can be confident it isn’t spuriously correlated with anything else. You can detect much lower risk ratios that way. That’s why they do them.
Epidemiology routinely throws up ‘significant’ low-level correlations with all sorts of random rubbish. I was taught the threshhold was 2-3. The FDA has more experience with these sorts of results, and is even more cautious than I am.
For comparison, the correlation between smoking and cancer has a risk ratio of around 20, if I remember correctly.
Stuff to worry about:The lifetime risk of dying in an auto crash in the US: 1 in 80.
Nullius,Experimental design can reduce the chance of confounding factors… yes surely.But why would there be less chance of confounding factors in a risk ratio of 20 versus a risk ratio of 2?Some effects will have a risk ratio of less than 2… you seem to be claiming that these can never be detected?… I don’t buy that.I also don’t buy the corollary; that increasing the number of studies cannot increase the weight of evidence.
#31,
There are confounding factors in any epidemiological study. The big ones you usually know about, but there are hundreds of small correlations that are generally unknown and far harder to get rid of. More studies can more precisely measure this spurious correlation, but they cannot get rid of it. For a reasonably well-designed study, the background noise is around 1-2 risk ratio, and results giving a risk ratio greater than 2-3 are worth taking seriously, at least to investigate further. But even the main smoking/cancer link took years to prove, with some really big names in statistics on the sceptical side.
Effects with a risk ratio less than 2 can only be reliably detected with more sensitive instruments. A randomised trial, where the putative cause is administered according to the toss of the dice, guarantees that nothing else influences whether the factor applies, and a double blind trial, where neither subject nor experimenter knows who is being treated and who is in the control group, guarantees that the factor under test does not influence anything else. With these guarantees of non-correlation, the test is far more sensitive.
The publish-or-perish ethos in academia does encourage many beginner researchers to churn out papers showing a variety of factors ‘causing’ a variety of conditions. Data dredges are an easy way to generate quick/easy results, but those experienced in the field know not to pay any attention to them. One factor in every twenty is statistically significant, and any factor will prove significant if examined closely enough. As researchers get established, they usually concentrate on only the stronger results, as they home in on the actual cause – unless they have an agenda.
The asthma study cited also listed race/ethnicity other than non-Hispanic white, and low parental education as significant factors. Do you really think that a person’s skin colour affects their tendency towards asthma? Or the school their parents went to? These are obvious proxies for poverty, which is commonly associated with poor health care, poor living conditions, and poor diet. There is no direct causal relationship between schools and asthma; it’s a spurious correlation. And of course, the poor tend to smoke more, too.
There is a common tendency amongst those who justify scientific credibility by its results that they judge scientific controversy by ends rather than means. Science succeeds because it demands that the method be rigorous. Your confidence that the result is true is founded on the care you put into testing and checking it.
But there are those who only care whether the conclusion turns out – with hindsight – to be correct. A correct result obtained by sloppy and inadequate means is a triumph. Those who challenged sloppy methods supporting a conclusion that later turns out to be correct are misguided enemies of science, and wrong to have done so.
Smoking was a classic case of this. The evidence initially put forward was sloppy and didn’t prove what it was claimed to. A number of very eminent statisticians objected. The methods were improved, and they still objected. The methods were improved more and more, until people ran out of objections, at which point the conclusion that smoking increased the risk of lung cancer became justified. But it was not wrong of all those people who held out and raised objections for all those years to have done so. It was only because they did so that we can now have confidence that smoking really does increase cancer. Only thoroughly sceptical attack allows science to progress.
Confirmation bias is universal. Science is the art of not fooling oneself; it is the catalogue of the incredibly elaborate techniques needed to overcome our own biases.
&tg; It was only because they did so that we can now have confidence that smoking really does increase cancer.
Confirmation bias is indeed universal.
#33,
But sadly, the scientific method is not.
Francis, Shea & Samet. 2006. Challenging the epidemiologic evidence on passive smoking: tactics of tobacco industry expert witnesses. Tob Control.
http://tobaccocontrol.bmj.com/content/15/suppl_4/iv68.abstract
Take-Home Exam: Compare and contrast with current thread’s dynamic.
#35,Like I said, belief in the scientific method is sadly not universal.
This thread might contain confounding factors between “the scientific method”, “eminent statisticians”, academic “publish-or-perish ethos”, “thoroughly skeptical” attacks, and scientific progress.
this is interesting. please elaborate.
Thanks!
Nullius,“For a reasonably well-designed study, the background noise is around 1-2”You appear to be stating that confounding factors, as a rule, cause risk ratios of between 1 and 2. Is that correct?Why would confounding factors be unable to cause risk ratios below 1?Thanks
@ 39 Lazar,You are right, you would expect to see to RR results coming in under 1.0 as well for the confounding factors. For example, in the WHO study childhood exposure to second hand smoke related to getting cancer came in at 0.78, with a CI of 0.64 – 0.96. So, the study shows there is a 22% less chance of getting cancer if you are exposed to second hand smoke as a child, and it is (just barely) statistically significant. Now, does anyone truly believe that? I certainly don’t, it just shows that these studies can swing either way if the link between cause/effect is tenuous.To get back to the start of this discussion, the original point made by Eli is that Heartland’s position is that “sound science” does not support the risks of second hand smoke, and he uses that Heartland position as slam-dunk evidence to shows Heartland as an anti-science death promoting machine. But, any rational look at the data shows that the case for second-hand smoke being a a clear & present danger as a health hazard isn’t there. May it be a danger? Yes! Should we limit second-hand smoke exposure as a precautionary step? Yes, why take the risk just so smokers can have an easier life? What I disagree with is the idea that significant danger from second hand smoke is obvious. It is not. And, I disagree with caricaturing people who hold the contratry position are anti-science dopes. They may well be but this is not the case that proves the point. The world is complex and simple worldviews like this do not help us advance.
Lazar,
While cleaning up my archives, I stumbled upon this episode that might interest you. The episode starts with this claim:
http://rabett.blogspot.ca/2006/04/effective-tobacco-
advertising.html
I’m not sure how we can reconcile this claim with this other one:
I’m quite confident that there is a way to reconcile these claims. But you have to admit that there are counfounding factors. I blame the lack of cafeine progress.
Caffeine has been found to reduce the risk of Alzheimers and Parkinsons. So, if you drink your coffee next to a smoker are you better off, or worse?
Amidst the confounding factors, I certainly feel less free.
So I will be less prosperous.
And I might even suffer from battle fatigue.
Perhaps it is too late for you.
Why did the chicken crossed the road?
Because it was too late.
#38,
The initial claims by Doll and Hill were made on the basis of 1500 patients in a number of different hospitals diagnosed as suffering from lung cancer, compared with a like number of hospital patients who were not. A survey was done of their smoking habits, and it was found that cigarette smokers were significantly more common among the cancer patients than the controls, and that heavy cigarette smokers more common than light smokers.
It was initially claimed that this suggested that smoking caused cancer (Hill didn’t claim a causal relation had been established at first) but they rowed back from this initial claim slightly when it was pointed out that pipe and cigar smokers did not show the same disparity, and that when you excluded the occasional cigarette smokers from them, the differences faded further.
It was also pointed out that the same study asked whether people inhaled, and contrary to what you might expect, inhaling was associated with less cancer – a statistically significant 10% reduction in risk. It was pointed out that by the same logic used to draw the smoke-cancer connection would also imply that inhaling reduced the risks. The sceptics did not think it did – it was an example to show the flaw in the logic and the selectivity of the researchers.
The research had been initiated because of a recent rise in the number of cases of lung cancer, particularly among men. However, it was pointed out that smoking among men had increased only a little, while smoking among women had risen rapidly. This of course is explained by cancer generally occurring much later in life. Although if it were simply that the damage is slow-acting, it seemed odd that people who had recently given up had much lower rates.
The main objection, though, is that it was a case of correlation being used to deduce causation. Correlation between A and B may be the result of A causing B, but it may also be the result of B causing A, or both A and B being caused by C. And C can include things not only out in the real world, but also in the experimental protocol, such as selection effects from biased sampling. The critics said not enough had been done to eliminate such possibilities.
And finally, it was said that human behaviour and biology is complicated, and does not fit neatly into tick-box categories. They overlap, change over time, or depending on the circumstances, are associated or dissasociated with other behaviours, environments, even genetics. Not enough had been done to isolate smoking as the only meaningful difference between the groups.
The complaint was that they had already launched the global campaign to try to stop people smoking before the ink had even dried on the first tentative evidence, which was a long way from being proven. The critics didn’t say that smoking didn’t cause cancer, or that there was no reason for concern, or that it wouldn’t be a sensible, rational choice for someone on hearing of a possible connection to give up smoking. But they said you needed to be very sure before claiming that science had proved something, and they weren’t.
The believers argued that if you told people the truth about the uncertainties, a lot of people wouldn’t be persuaded to give up, and that would cost more lives. They also came out with the usual ad hominems – it’s because you’re a smoker, or because you’re paid by tobacco companies. (Again confusing cause and correlation – maybe the tobacco companies paid them for advice because they were open to alternatives…) But eventually other researchers did follow up with many more studies to nail down and eliminate all the alternatives.
The tobacco controversy is so significant and influential because it subsequently became the standard template for the ‘health scare’ – a recurring meme in Western society. A data dredge is used to scrape up marginal evidence connecting some frightening disease with some disapproved-of behaviour, a paper is published, is hyped by the health control lobby, and pressure applied to bring in bans, regulation, or changes in behaviour. The same template has been followed for pesticides, food additives, plasticisers, high voltage power cables, radioactivity, mobile phones, vaccines, genetic engineering, saturated fats, cholesterol, salt, red meat, alcohol, and many more. Weakness of evidence is not mentioned as that would lend ammunition to those resisting action, and the precautionary principle requires we act early ‘just in case’.
They turned out to be right on smoking, and ever since then have built on that success, with sceptics dismissed as shills for whatever industry profits from the behaviour being attacked, and accused of using tobacco-lobby tactics. With considerable success, too.
But the point remains: the evidence against smoking is so strong because of the sceptics. Being right is not enough; you have to be right for the right reasons, and it’s no shame to be critical of the truth. Science relies on scepticism – about everything.
And the consensus among properly sceptical scientists is that ECS is ~ 3C. Heartland’s best efforts at misrepresentation notwithstanding.
🙂
#39, #40,
Yes, you do get risk ratios below 1, but usually if that happened the categories would be swapped around. If smoking gives a risk ratio of 20, then not smoking gives a risk ratio of 1/20. They go in pairs.
#41,
I had a quick look at the bunny’s scribblings. Apart from one appeal to authority near the start, he doesn’t actually talk about evidence. His case appears to be built on pointing out that the tobacco industry advertises its products. And – horrors! – it does so in a way that makes their product seem attractive.
There is a school of thought that says it’s none of the government’s business if people choose to risk their own lives doing things that are bad for them. It needs to be an informed choice, and I’m all for making people better informed. But if people know the risks and go ahead and do it anyway, the State needs to butt out at that point. This attitude that if people don’t do what you decide is good for them, they have to be made to is far more dangerous than a lot of the things they campaign against. Brer rabbett’s post drips with it.
But that’s besides the point. There’s no contradiction between my suggestion that the evidence was initially weak and strengthened as a result of criticism, and the bunny’s that tobacco firms have continued to market a product that everyone now knows is bad for your health. Or for that matter Pielke’s point that knowing the science didn’t have much effect on people’s behaviour.
Nullius,
Thanks. To reiterate my question so that I may understand what you’re claiming…
“You appear to be stating that confounding factors, as a rule, cause risk ratios of between 1 and 2. Is that correct?”
As in, it is highly unlikely that confounding factors would cause a risk ratio of 20?
Tucked away in NIV’s lengthy # 46:
They turned out to be right on smoking
Mustn’t forget this, must we?
🙂
Nullius,
“But the point remains: the evidence against smoking is so strong because of the sceptics.”
“because” elides the role of the scientists who did the work, and presumes that followup studies would not have occurred without criticism from alleged skeptics — a situation which I think is improbable.
#49,
Yes.
#50,
I wasn’t proposing that you should. 🙂
#51,
Are sceptics and scientists mutually exclusive classes?
So Heartland is helping to improve the current scientific understanding on AGW?
Heartland Institute:
The point is that believing in global warming is not “mainstream,” smart, or sophisticated. In fact, it is just the opposite of those things. Still believing in man-made global warming ““ after all the scientific discoveries and revelations that point against this theory ““ is more than a little nutty. In fact, some really crazy people use it to justify immoral and frightening behavior.
#53,
Heartland? No. I shouldn’t think so.
They’re down at the political end of things. They do for sceptics what Ted Turner does for mainstream climate science. Only on a much smaller budget.
Okay, let’s look at the robust scientific sceptical argument.
Oh.
So it’s back to the billboards, I suppose.
On which point, the HI is sorely in need of some fresh talent. Perhaps you could sharpen up their game for them 🙂 Unless you are already fully committed elsewhere, of course.
> There is a school of thought that says it’s none of the government’s business if people choose to risk their own lives doing things that are bad for them.
This might appeal to the authority of
http://en.wikipedia.org/wiki/Rational_choice_theory
This might not be the best of times to appeal to this school of thought, more so if we’re willing to mix it with Masonian public choice theory.
> There’s no contradiction between my suggestion that the evidence was initially weak and strengthened as a result of criticism, and […] Pielke’s point that knowing the science didn’t have much effect on people’s behaviour.
Let’s recall Pielke’s point:
And so we can either say smoking efforts to deny a link between smoking and health risks, or more simply “strenghtening criticism”.
Tomato, tom-a-to.
“But the point remains: the evidence against smoking is so strong because of the sceptics.”And, of course, because of the <a href=”http://rabett.blogspot.de/2006/04/tobacco-mortality.html”>millions who died in the mean time. Martyrs to science skepticism each and every one.