In the blog, they attack Singer for taking stuff out of context, and the skeptics for misrepresenting their results. But they also defend the methodology of the 2003 survey, which Tim Lambert attacked [update: TL maintains his opinion]. I’m still dubious about the possible selection bias, though the good news is that they will address this in the next version.
But they also need to redesign the questions (as they note). The most headline-y one was “Climate change is mostly the result of anthropogenic causes.” The answer to this depends a lot on the timescale you’re thinking of. If you mean 1000′s, or 100,000′s of years, then the answer is clearly no. If you mean last-50, then the answer is clearly yes, unless you’re a wacko. The possibility of misinterpretation of the question (I assume that B+von S meant recent change) makes the results difficult to interpret.