Tuesday, July 08, 2008

Shonky analysis, shonky polling from NZIER

During the debate over the ETS, the New Zealand Institute of Economic Research presented a report [PDF] claiming that the introduction of emissions trading would destroy the economy. While cited approvingly on the right, the report was spectacularly flawed, being based explicitly on an assumption that there would be no international action to combat climate change (it also assumed there would be no technological change, no benefits from action, and no costs of inaction - all of which are similarly dodgy). It was simply bad analysis.

Now the NZIER has released the results of a poll [PDF], purporting to show that the public are more interested in the cost of living than the ETS. It's already being cited approvingly by the usual denier mouthpieces at the Herald, and no doubt will show up soon in a National Party press release. But if they'd bothered to read the polls methodology, they'd have noticed a major sticking point: its compiled from a weighted online sample. Which means that, like the BCSD's ShapeNZ "polls", its not even social science. The first rule of surveying, drummed into any undergraduate's head from day one, is that in order to produce meaningful results, the sample must be random. Self-selecting samples - such as those you get online - simply produce junk. In other words, NZIER is palming off a tarted-up version of the daily Stuff poll. If I'd paid a polling company for that (and apparently they did) I'd be asking for my money back.

The NZIER claims to be "the premier centre of applied economic research in New Zealand". Perhaps they should start acting like it, rather than serving up such shonky "research" to the New Zealand public.