Watts Up With That (WUWT) is promoting (The Tol Poll – rating climate denizens) an online poll put together by Richard Tol that asks volunteers to rate a group of people on a scale of 1 to 5 where 1 is very nasty and 5 is very friendly. Now, I for one would be very pleased if people engaged in discussions about climate change/global warming were friendlier (or if the discussions were more pleasant) but I don’t think this poll is doing anything to achieve that goal.
Maybe Richard’s just having a bit of a laugh and I certainly enjoy a bit of giggle now and again. If you can’t have a bit of a laugh every so often, then you’re taking life too seriously. But Richard apparently intends to release the data. Hmmm, this doesn’t seem very funny anymore. Is this an attempt at a serious poll? Does he think it would be funnier to actually collect and release the data than to simply have a fake online poll to make some kind of point? Bear in mind that Richard Tol is a Professor of Economics at the University of Sussex and is also a co-ordinating lead author for the IPCC’s 5th Assessment Report (AR5), and has been involved in previous IPCC Assessment Reports. I was once advised by someone that I should not dismiss Richard Tol, as he is a serious academic. This isn’t really what I would expect from a serious academic but, I guess, there aren’t any rules so maybe one could argue that anything goes.
I’ll make a slightly more serious point though. Richard Tol has been a vocal critic of John Cook’s consensus paper, arguing that tests of the data highlight issues that indicate that the study is flawed. However, as I try to point out in this post it’s not clear that such tests are appropriate to the Cook et al. study. I also suggest that one could design a study that passed all the test but that would produce results that were meaningless.
I think that this may well be what Richard is illustrating with this new poll. I have no reason to suspect that the data will fail any of Richard’s statistical tests, but what will the results actually mean? Most of those taking the poll won’t actually know the people they’re rating so, at best, they’re expressing an opinion based on what they’ve read online. Given that the poll appears to be anonymous, you also have no idea what biases exist in those who’ve taken the poll (so really, Richard won’t be releasing all data).
So, apologies, if I’m not seeing what is meant to be a very subtle and clever joke. Doesn’t seem very funny to me, but then I’m happy to be convinced that it is. But, on the other hand, kudos to Richard Tol for illustrating that one can set up a survey that is likely to pass most (if not all) statistical test but that produces results that are essentially meaningless.