Re: [asa] Level of certainty in science

From: Don Winterstein <>
Date: Tue Feb 06 2007 - 18:58:32 EST

Pim Van Meurs wrote: "What is so scary about science being correct once again? If one
believes this to be scary, imagine how scary it could be when
Christians, as Augustine pointed out, are observed spouting scientific

Who's got more to lose here? The climate data may be fairly firm, but the models, from which scientists make predictions, are likely much less trustworthy. They're Earth science models, after all. If Earth starts cooling within a few decades--as it's been known to do, climate scientists and other scientists by association will lose credibility for a long time to come.

I'm skeptical of the models, but my skepticism has almost nothing to do with my religious beliefs. It has to do with my experience with Earth science models. No, they weren't climate models; but they were models that involved Earth science data. Earth science data are not like lab data, taken under well-controlled conditions, but data recorded instead under often extremely challenging conditions. They are data that almost always need some conditioning in order to properly mesh with other data, and in those conditioning steps the scientist has some leeway to impose his druthers on the outcome.

Soon after I joined an oil company and switched from physics to geophysics, two of the company's most highly regarded geophysicists came by to tell geophysics jokes (they knew I was a fresh PhD in physics):

Joke 1: What's the difference between a geophysicist and a physicist? Answer: A physicist actually believes his models.

Joke 2: The boss goes to three of his reports, a mathematician, an engineer and a geophysicist, and asks, "What's two plus two?" The mathematician responds immediately, "Four, obviously." The engineer thinks it over, and acknowledging that there are always uncertainties and errors in the real world, says, "I'd say it's likely to be 4.02." The geophysicist paces and sweats and agonizes, paces and sweats and agonizes, until finally he looks at the boss somewhat sheepishly and says, "What would you want it to be?"

I didn't fully understand the point of either joke until I'd had more exposure to geophysics, but eventually the truths behind them became obvious and compelling: There's always an outcome that's more desirable than other outcomes, and in processing and manipulating geophysical data there are always opportunities--because of the nature of the data--for shifting results towards the preferred outcome without being dishonest.

Oil company geophysicists know better than to believe their models. Sometimes their models are quite good,
sometimes they're way off. But you never know until you drill the well, when you get to check out your predictions. The models are never perfect, and there are often surprises. It's not that the scientists are incompetent. Oil companies often hire the best that money can buy, from top universities. The problems lie in the nature of the data--its collection under uncontrollable conditions--and the nature of Earth.

Why do I think climate models are subject to similar uncertainties? I hear things that sound closely similar to arguments geophysicists use all the time: "I had this problem getting a good match here and here, but then I looked into the details of data acquisition and special recording conditions and this and that, and I realized a shift was necessary here and over here to compensate, and now we have a much better fit." And later, "Well, that result wasn't really as good as it could have been; I've thought of this new correction, and now the fit is even better."

Do climate scientists actually believe their models?

The final model may be pretty good, predictions from it may be pretty good, but on the other hand, maybe the model isn't that good after all, and predictions will fall flat.

Scientists who buy unhesitatingly into climate models are playing a high-risk game. They could lose big. How will Christians look if they eagerly buy in and then the climate models prove to be wrong for whatever reason? The wiser course is to hedge bets.



To unsubscribe, send a message to with
"unsubscribe asa" (no quotes) as the body of the message.
Received on Tue Feb 6 18:57:14 2007

This archive was generated by hypermail 2.1.8 : Tue Feb 06 2007 - 18:57:14 EST