Romm's "Smoking And Cancer" Fallacy

Joe Romm is not the only one making the absurd analogy between the smoking-lung cancer link and the carbon emissions-global warming connection:

Everyone knows you can’t make a direct connection between carbon emissions and this January in Vancouver which is so damn warm it crushed the record set so long ago that toddlers can’t even remember it. It’s just a coincidence that we are now in the warmest winter globally in the satellite record.

It’s just like that chain-smoking guy who got lung cancer. The fact that he smoked two packs a day is a coincidence. You can’t prove it — so keep smoking, already. Sure the statistics show the warming footprint — Record high temperatures far outpace record lows across U.S. — but individual events are just coincidence. I’m telling you.

Romm’s sarcasm is wholly inappropriate. The  Relative Risk (RR) of developing lung cancer is around 23 for habitual (male) smokers. There is no reasonable “you can’t prove it” argument: indeed, here’s a checklist of what is needed to understand a phenomenon where most data are of a statistical nature:

  1. Epidemiology should find a strong association (i.e. a high value for RR, e.g. above 3)
  2. A very specific disease should be involved
  3. There should be a consistency between studies and with data from laboratory work, disease incidence trends and other sources
  4. The results should be preferably not involve a rewriting of biology and physics

(you can read more about the complexity of dealing with a statistical understanding of the world at this link)

In the case of smoking and lung cancer, every single point of the checklist is fulfilled. In the case of carbon emissions and global warming:

  • Point 1 is still ill-defined – notably, the fork between maximum and minimum expected warming has not decreased between the IPCC TAR and AR4. That’s a far, far cry from a RR of 23…
  • Point 2 is still ill-defined – we are given very generic statements “it’s going to get warmer”, “it will likely be a warm winter”, “the likelihood of heavier snowstorms and rainfalls will increase”, almost value-free if there’s no number attached to them

No need to talk about points 3 and 4. If there’s no well-defined data to work on, everything else is a moot point. All in all, it is sad to see just how misinformed somebody like Romm can be, when one is unwilling to find the time to understand the topic at hand. Hasn’t he got anybody helping investigating his own arguments???