Untangling Chamber Statistics

Jersey's Chamber of Commerce has released the second set of results in their 2008 series of opinion polls. Designed to give business a say in election year, 154 members responded this time round, more than the previous poll. Treasury Minister Terry Le Sueur continues as the frontrunner for the Chief Minister's job, but the number of members who are still undecided has increased.

http://www5.channelonline.tv/news/templates/moneychannel.aspx?articleid=15636&zoneid=1

I heard this on BBC Radio Jersey, and it struck me as a much more balanced report.

This was mainly because they went into more details. They mentioned that Terry le Sueur was still in the lead of those named, but down from 25% in January to 21%, and that most members who responded did not want him as Chief Minister (up from 20% in January to 29% now). These were the "don't know answers" which the report said was a "significant rise". In this case, "don't know" meant none of the choices available - including Terry le Sueur.

Channel TV somehow forgot to mention that more didn't want him than did, just that "the number of members who are still undecided has increased." But those "don't knows" were in fact decided about one thing - they didn't want him. That was clear from the Chamber Report, and the BBC presentation.

A few caveats: This is only 154 responses from 550 members., a 28% reply rate of those who bothered enough to do the questionnaire, which means that all we know for certain is that about 6% of Chamber Members would like Terry to be Chief Minister. We can't assume the sample is representative, because it is self-selecting, not random. It could be anywhere between 6% and 78% in all. And we don't know how many businesses the Chamber represents as a percentage of the total either, and whether its representation is skewed in any way..

Incidentally, the one notable option was Guy de Faye, moving from 57% not wanting him to 78% and no "don't knows" there! I feel a bit sorry for him, even if he tends to be his own worst enemy.

P.S. For Nick Palmer (if he reads this), I don't see Channel TV's reporting as a conspiracy, simply that they were not good as the BBC as presenting the statistics unambiguously in this case.

Book of the Post:
Damned Lies and Statistics: Untangling Numbers from the Media, Politicians and Activists - Joel Best