William M. Briggs

Statistician to the Stars!

Zombies no joke: global warming can cause anything

A day ago my number two son and I sat over a bottle of wine and he suggested that if global warming caused temperatures to increase, then we would see an escalation in the number of zombie attacks because, obviously, there would be less cold weather, which everybody knows slows attacks from the undead. I wrote this up in the approved New York Times format, and, to my astonishment, some people thought I was joking.

I was not. The post was in earnest and was an attempt to put into perspective the hundreds, if not thousands, of “studies” that purport to show the ills that will befall us when global warming finally strikes. There are three problems with these studies.

The first is their ridiculous variety, nowhere better cataloged than at NumberWatch’sWarm List“. That page contains links, mostly to news reports to studies that ask us to believe that, for example, lizards will undergo sex changes, there will be “waves of rape“, a rash of camel deaths will occur, the Earth will spin faster (hold on!), and, worst and most frightening of all, there will be an increase in lawyers (to handle all the “who’s fault is it?” litigation, you see).

I listed only five of the hundreds on that page, which no doubt represents an undercount of the true number of worrying research reports. New ones appear daily. If you wanted to adopt a cynical attitude, you might think there is an unstated competition among researchers to see who can get the most of these or the most shocking of these things to press.

Can these studies all be true? Yes, it logically is possible, only it is absurd to think so. The probability of each and every one of these calamities coming to pass is as close to zero as you like. But that’s not the real problem. It is that each of these studies is usually taken as further proof, albeit indirect, that significant man-made temperature change (AGW) is true. “Why else, if AGW was not true, would these respectable scientists publish these studies?” people ask themselves. Only, it’s the wrong question.

Each of these studies claim to show a danger that might come to pass given that AGW is true. That is exactly backwards to answering the question of whether AGW is true, however. If instead these studies showed that these maladies already occurred then it might provide some evidence, however weak, to support AGW. But then again, it might also support the theory that the observed climate change is natural and expected. Few or none of these studies show what results to expect given that AGW is false (and climate change natural etc.). To be useful research both scenarios must be analyzed, else we are right to suspect the researcher has been sloppy and perhaps not a little biased toward a specific conclusion.

To be specific: because a study appears showing the harm that AGW might cause it is not, and cannot logically be, proof that AGW is true.

The second problem with these studies is their wearying specificity and confidence, which we alluded to by stating that our zombie researchers “calculated a 32.782412% increase in” attacks. Just joking? At the site Skeptical Science, hosted by an honest man, we find that, given AGW is true, there will be “Increased deaths to heatwaves (5.74% increase to heatwaves compared to 1.59% to cold snaps)”. Really? 5.74% and not 5.73%? Are they sure? Can they even be so confident to say 5.7% and not, for example, 5 or 6%? As a statistician, I can assure you, your model and sample have to be incredibly accurate for you to make verifiable statements to that many decimal places.

These studies almost never give any indication of their uncertainty about their results or assumptions. Instead, the “findings”, or results, are taken to be a given; they are just obviously true. Stating results of a study in this fashion has its intended affect: it increases worry. But it does so to an extent that is almost never warranted. To do risk analysis of a study’s results, accurate estimates of the uncertainty and range of possible effects, including positive ones, must be present, else the study is worthless. Reports without uncertainty and estimates of range of effects again bring up the question of the author’s possible biases.

To be specific again: a research report that does not include measures of uncertainty of its results, and an explicit list of the assumptions and the uncertainty in them, is of almost no use.

The last problem with this research is what you do not see. Again at the site Skeptical Science, the author compiles a list of bad things that will happen if AGW is true, and compares it to a list of good things that will occur. Good things? Yes, of course: because it is impossible that climate change can only be bad. Since about four billion years ago or so, the climate on earth has never been static, nor is it ever expected to be, so to claim, like many do, that any change in climate must be bad is just silly. Responsible scientists of course know this.

But many of them still suspect that any changes will be mostly bad. This is evident scanning the Skeptical Science list. Many items in the “good things” column are unfortunately facetious, for example “Record profits for pharmaceutical companies”, and a “thriving” trade in Mammoth fossils (truly, global warming can cause anything).

There is, naturally, an official psychological name for this sort of behavior, but we don’t need that label to see that if thousands of researchers rush to find the worst that can happen and none (except for corporate “shills”) try and find the best, that there will be an enormous bias in the published literature towards the worst. Add to that the search only for evidence that backs up theories of the worst, and you have the situation we are in now. Which is not quite panic, but extreme distress in some quarters; cries of “Something must be done!” are regularly heard, and any that dare question this necessity are raked over the coals.

It also causes some skeptics of the AGW hypothesis to perhaps lean too far the other way and make statements that they will probably regret later. But these skeptics feel forced to speak out to balance the overly-heated and anxious rhetoric that is generated because these studies get so much press.

To be specific once more: it is irresponsible and harmful not to seek the full range, both good and evil, of what will happen given the climate changes.

My prescription is for restraint among scientists who publish these studies: slow down and do a better job, include estimates of uncertainty, better emphasize limitations, and honestly describe what good might arise whether AGW is true or whether climate change is natural and expected. I predict, however, like the majority of medicinal regimes directed by doctors, that my patients will not be adherent.

7 Comments

  1. Long live the zombies … no wait… long dead the zombies … err … long undead the zombies?

    What do you mean I missed you main points?

    We don’t, or at least shouldn’t, put any weight behind a poll that does not publish its certainty, sample size, and sampling method let alone its assumptions like “given that AGW is true… .” Instead too many believe the one sided AGW stories that the media present without question because it is from ‘experts.’

    “…because it is impossible that climate change can only be bad”

    This reminds me of reports in Canada about the ever changing dollar. If it goes up, then the media report that manufacturers, tourism operators, and exporters will be hurt. When it drops with respect to the US currency, then consumer confidence is lessened and our economy will take a hit.

    John M Reynolds

  2. An amusing comment from J. Scott Armstrong at the Wharton School in Pennsylvania (he wrote the book on it, Long-Range Forecasting, and also edited the standard textbook in the field, Principles of Forecasting)

    “How good are climate scientists at predicting the future? On the whole, says Prof. Armstrong, they’re probably no worse than any other group of experts ? which is to say, not very good at all. Abundant research shows that experts ? whether they’re political scientists, stock market analysts, Super Bowl commentators or marine biologists ? are no better than non-experts at making accurate predictions. They’re just more certain that they’re right”

    http://www.theglobeandmail.com/servlet/story/RTGAM.20080201.cowent02/BNStory/Front/home

    .

  3. have a look at my findings on a PNG village being relocated due to Global Warming caused rising sea levels.

    But it is really because they are building a port where the vilage is!
    http://www.creditablecarbon.com/forum/viewtopic.php?t=30

  4. This form of research was pioneered by Blaise Pascal. Call it “proof by hellfire”.

    If you can’t prove that your thesis is true, then try to prove that failure to act as if it were true will result in x horrors lasting time t. (Begin with x = your skeptic’s death and t = time of death. Increment x and increment t when challenged. Repeat.) People have an internal function f(x,t)

  5. Interesting post, particularly the comment on the percentages to so many decimal places. To be honest, I find the knee jerk reaction to attribute anything bad to global warming a bit annoying and I’ve tried to only include the more substantial, creditable items in the negative side – preferably with links to peer reviewed studies.

    For items in the positive side, I’ve struggled to find many and believe me, I read many skeptical articles on global warming. Hence I’ve included a few of the more facetious ones mainly for humour (but also to pad out the positive side a little).

  6. In my field, ecology, many studies are done taking a 100 yr climate forecast and projecting ecosystem change as a result. Many many assumptions go into the analysis. The result is really a hypothesis, not a “result” and it will not be proven true or false until the author is long dead. Pretty good for getting tenure IMO, but not such good science.

  7. A complete list of things caused by global warming
    http://www.numberwatch.co.uk/warmlist.htm

Leave a Reply

Your email address will not be published.

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

© 2015 William M. Briggs

Theme by Anders NorenUp ↑