Ordinary Gweilo

Some interesting analysis of the opinion polls from ESWN (via Simon). I wasn’t aware of this, but it seems that the DAB’s share of the vote is consistently under-estimated in opinion polls.

This is not a new phenomenon, and interestingly, there’s a piece in this week’s Economist about the differing results from opinion polls on the US election. They mention…

a problem that pollsters have long recognised: people who feel the winds of public opinion shifting in their direction are more confident about telling pollsters what they think than those who don’t.

Perhaps Americans like to back winners, but in the UK it was more a matter of apparent embarassment about supporting the party in power. This caused opinion pollsters major problems in the UK when the Conservatives were in power (notably in the 1992 election, which was widely expected to result in a Labour victory). When asked, people were most reluctant to admit that they were intending to vote (or even had actually voted) for the Conservatives. At first, opinion pollsters (reasonably enough) took the answers they were given at face value and found that their predictions were hopelessly wrong. Then they started adjusting the raw data to take account of this discrepancy (and I believe they also had some cunning scheme to allow people to express their preference ‘secretly’ when data was collected face-to-face rather than over the phone). Whatever they did seemed to fix the problem, and I guess the HK pollsters have no choice but to do something similar.

ESWN also notes that the results have been subject to vastly different interpretations in various media around the world. We aren’t electing a government and so there are no winners or losers in absolute terms, and in addition to that, comparing these results with 2000 is difficult because there are now more seats in the geographical constituencies. So, as ESWN notes, it largely comes down to how the parties did against expectations. That means that there’s ample scope for almost any interpretation you wish to make!

One area where I think EWSN is being a trifle unfair is in saying (even with the benefit of hindsight) that it was a mistake to field two separate lists of pro-democracy candidates on HK Island. The reason for having two lists is because some voters might support Audrey Eu (or Cyd Ho) but not be so keen on Yeung Sum or Martin Lee (or vice versa), so a single list might be off-putting for some people. As seems to have been the case in NT East, according to Christine Loh in today’s SCMP. She feels that it was a mistake to field a single list because the candidates were so diverse, and it certainly didn’t work as well as they had hoped. I don’t think there is any guarantee that it would have worked any better for the HK Island consitituency.

There is no simple answer to this conundrum, and unless we have a change in the voting system the democrats will always have this problem. The big error, obviously, was in encouraging voters to switch to Martin Lee’s ticket because opinion suggested he might lose. Which is where we came in, I think.