Sunday, June 27, 2010

Is It More Efficient to Replace Science with Public Opinion or Public Opinion with Science?

Recently, there was an article in the NYT about the declining belief in global warming, and/or its negative impacts in Europe.

I would like to suggest that all scientific problems be solved with polls. We simply ask a representative sample whether Yang-Mills theory exists and has a mass gap, whether the Riemann Hypothesis is true, or the best course of action regarding a quantized theory of gravity: is it a) Strings, b) Loops, c) Triangles, d) Entropy, e) I Don't Know, or f) I've Never Studied Gravity, But I'd Like to Say It's (a) Based on No Information Whatsoever. This is far easier than doing science.

[As an aside, society as a whole pays quite of bit of money to have scientists around in order to have opinions on scientific topics. I've never quite understood the propensity for people to then dilute their paid expert opinion with their own opinion. It's like going to a restaurant and insisting on cooking your meal yourself. The division of labor has given us so much!]

Of course, the problem with polling is always systematic bias, based on your wording, or in fact, in the case of the polls cited in the NYT article mentioned above, the outside temperature. Unfortunately, the belief that global warming is happening/bad/imminent is strongly correlated with the actual temperature.


In the article, first poll was taken in November 2009 after an above average temperature summer and the second poll was taken in March of 2010 after an unusually cold winter. That led to a drop in belief that global warming was happening and man-made from 41% to 26%.



And the German poll (42% feared global warming) was also taken in March and compared to a poll (62%) taken in one of the summer of one of the warmest years on record there.

It is entirely possible this NYT article may have been deliberately written to be deceptive. I found this.

Now climateprogress.org is a partisan site, so I'll have to do more digging. But it does explicitly call out Rosenthal (the author) and anyone who cites Viscount Monckton is pretty much a nutcase (no offense to nutcases).

Here's a new global warming poll showing increasing belief in the US as we go from January to June of this year.

The key to understanding the public's opinion of global warming would probably involve some kind of seasonal average, taking into account the average temperature in the location polled. Only then could you discover what the people really want. Better yet, bypass the polling of human beings and just measure the temperature and extrapolate public opinion.

Then we could move into other areas. Opinion of the president or the effectiveness of or trust in government is in fact strongly dependent on economic factors. We could account for this. Of course, if we did, we would probably find that every president's approval rating would likely be exactly the same.

And as the global temperature will likely continue to rise, eventually more people will believe it. The seasonal effect will also slowly disappear, because, well, one of the effects of CO2 heating the planet is that it retains heat in the winter months (as well as at night, and near the poles). So by the time global warming has started to wreak its havoc, we should have a well informed public regarding the havoc that is being wreaked by global warming.

2 comments:

  1. Can I suggest a companion word to "science" to use in the context of rampant pollsterism?

    Because absolute knowledge and public opinion are completely on par, this new word must have at least as distinguished an etymology as "science". Whereas science indicates knowledge (from the Latin), the corresponding word I propose is "accredence" (from the Latin verb accredo).

    ac·cre·dence
    –noun
    1. a branch of study dealing with a body of perceived facts or truths systematically arranged and showing the operation of public opinion.

    2. perceived systematic knowledge of general beliefs gained through leading questions and biased sampling.

    ReplyDelete
  2. I like it. I will create a post tag immediately.

    ReplyDelete