16 October 2011

A graphical dalliance

Posted by Callan

I read an article in the current issue of Physics Today with interest. It deals with the nature of scientific controversies, as percieved by the public and by specialists in the field in question. The author, Steven Sherwood, compares the origin of the ideas of a heliocentric solar system, general relativity, and human influence on the Earth’s climate. Each of them follows a similar pattern, he argues, with the initial idea leading to increasing levels of expert consensus, organized public backlash, and then a growing public consensus. It’s worth a read. But this post is just to comment on one of the graphs included in the article, Figure 3:


Basically, the point here is that the more coal your country produces, the less likely your populace is to accept the notion that burning that coal causes changes in the Earth’s climate. So if you were to fit a trend to these data, that line would have a negative slope.

It reminded me of a widely-cited survey from 2005 on different countries’ acceptance of evolution. On a whim, I decided to compare the acceptance-of-evolution data with the acceptance-of-anthropogenic-climate-change data. But the two datasets were not identical (they included different countries), so I excised the countries in the climate-change dataset for which I didn’t have evolution data. Say goodbye to Mexico, Brazil, Canada, and Australia:

(modified from source)

The trend loses some “oomph” once those four countries’ data are excised, eh?

Then I eyeballed the coal production values, and added them to a spreadsheet I had already developed with the evolution values. And I plotted it, again, keeping coal production as the constant variable:

Not as clear a trend.

I wouldn’t expect coal production to necessarily correlate with rejection of evolution – my hypothesis was simply that anti-science viewpoints in different countries might correlate broadly. I was wondering if the evolution vs. coal graph would have the same pattern as the climate change vs. coal graph. It doesn’t, really, though there are some similarities: The U.S. and Turkey are near the bottom of each plot, and industrialized European countries + Asian countries are near the top. But I’m not sure I really learned anything new in this graphing exercise.

Oh well: it was fun to do. Maybe the more interesting plot would be rejection of climate change vs. rejection of evolution…

Yes, that gets closer to the relationship I was looking to explore – Italy, Spain, and South Korea show similar levels of climate change consensus as the expert climatologist community. The U.S. and Turkey are the most extreme outliers in this group of nations. However, you should note the difference in the horizontal percentage scale and the vertical one. Overall, the human influence on climate change is much more widely accepted than evolution. This becomes evident if you plot them both on the same scale (with 0% as the minimum axis value):

All the variation in climate change opinion takes place between 70% acceptance and 93% acceptance. Plotting the data on an axis that runs from 0% to 100% results in a lot of empty space at the bottom of the graph. Variation in evolution opinion, on the other hand, spans a wider range: 27% to 80%. It’s more diverse, but for all nations, there is still a lot of room at the top – the far right side of the graph. No nation shows more than 80% acceptance of evolution – an astonishing statistic in my mind. Call it the “glass ceiling of scriptural literalism” – reason cannot rise any higher given the current cultural mores.

Of course, the big story (if any) from this display of data is that both the polled public of the United States and Turkey are not as keen as peer nations on these two particular scientific insights. Both also have strong religious traditions: Christianity in the US, Islam in Turkey. Is that sufficient explanation? Perhaps, or perhaps there’s more to it than that.

But I’m done with my graphical dalliance,  so that’s all I’ll analyze this issue for the day. Any thoughts?