Numbers

Back in November, when a bunch of emails between climatologists were acquired by the anti-global warming crowd, climate scientists were accused of a lot of nasty things–not the least of which was writing computer programs that were rigged to show global warming was happening regardless of what data were plugged in to them.  Global warming opponents urged scientists to simply rely on the data rather than using “tricks” to skew their graphs, and then danced around the Internet gleefully slandering empiricism.

At the time I thought it was a bit contradictory for global warming’s opponents to claim that scientists could write computer programs capable of skewing any climate data to make it appear that global warming was occurring, and that they should just observe the data without employing their “tricks.”  It seemed to me that if the data were so malleable that scientists could make it say anything they wanted, there would be no benefit from studying it in its raw form.  (It would be like telling someone to study the fundamental shape of clay.)  Likewise, if the raw data were understandable, then scientists should be able to make graphs that bring out the data’s true meaning, or graphs that don’t make sense, but not graphs that say anything they want.  (It would be like taking Lincoln Logs and trying to build a sports car.  The structure of the logs won’t allow them to form a car’s shape.)

But I’m not a scientist, and lots of scientists were handling the issue nicely, so I kept all that to myself.

Then, on Monday, Bad Astronomy posted a graph devised by Princeton astronomers containing the entire visible universe…one graph.  It was neat because the y-axis showing distance was based on logarithms, so each hash mark going up the side not only denoted an increase in distance but an increase in the increase in distance.  (First 10, then 100, then 1,000, and so on.)

It reminded me of the global warming emails since the graph had quite obviously been skewed to make 1,000 Kiloparsecs take up the same amount of space as 1,000 Kilometers using this logarithm “trick” to steadily change the represented distances.  It also made me realize why it can be important to fiddle with how data are presented on a graph.  After all, even objects as comparatively close together as the planets of our solar system can’t be graphed easily or meaningfully if the distances are kept proportional.  (Case in point.)  If you were to create a graph of the entire universe with accurately proportional distances, it would be a giant file of empty space.  And there wouldn’t be much use in that.

None of these “tricks” suggest a conspiracy regarding the size of the universe, just like the global warming emails didn’t suggest any conspiracy regarding the climate.  Data sometimes just needs to be tamed and organized before people can make sense of it.

Alas, this has been old news for months now.  I just thought I’d share.