Scientists get frustrated when their work is dragged into the political arena, and for good reason. Scientists have an excellent set of procedures, which have stood them in good stead over the years - a process of vigorous debate and repeated efforts to look at data and theory in order to converge on useful truths.

If someone has an idea that overturns conventional wisdom, it can be tough going. But over time, if the idea works well enough - and by that I mean does a better job of explaining how the world works that the conventional wisdom it is replacing - it will win out. We have seen this over an over in science, from Einstein's relativity to Wegener's long-delayed realization that Earth's continents move.

But when science is dragged into the public square, a different set of rules apply. There, any old sciency fact will apparently do, as long as it sounds reasonable to the non-scientific audience to whom it is addressed, and supports the political agenda of the person making the argument.

Such is the case with round two of George Will v. the climate science community. In a column published today, Will responds to criticisms from experts who said he had misprepresented data about global ice levels, and what they might say about global warming. The scientists who actually deal with sea ice say Will is wrong (see Andy Revkin's discussion in today's New York Times for a discussion). But a comparison between Will and Revkin suggests why this is such a hard problem.

Revkin's long and thoughtful discussion explains satellite measurements, the difference between northern and southern hemisphere sea ice, the issues of thickening and thinning in the annual cycle. It's a great summary, and worth a read if you'd like to understand the science. But in the political arena, it's much easier to just grab a quick sound bite, as Will has done - Sea ice numbers show global warming isn't happening! - and run with it.

This is why I sometimes despair of our ability to use science to guide political and public policy decisions. If cherry-picking one bit of science that seems to sound good to suit your argument is acceptable, rather than looking at its totality, then science largely ceases to be useful.