The unscrupulousness of the polling in Dobell approaches the wicked and may be criminal, writes Bob Ellis.
THE POLLING we did in Dobell, going up to people of all ages in shopping malls on a Saturday and Sunday, suggested a Labor/Green/Craig Thomson vote of 64 per cent. Newspoll on Friday suggested a Labor/Green/Thomson vote of 46.
How was this difference achieved?
Well, it looks like they rang, on landlines, 253 people in Dobell on late shopping night and got 102 for the Coalition, and then, ‘weighting’ the figures, got that up to 126. And then not asking people what their preferences were, but basing the figures on 2010 when there was no Thomson Independent, got them up to 136.
136 is 264 less than a sample should be — and therefore worthless.
But looking more closely, we find it doesn’t mean that in Dobell at all.
For the numbers are spread over two electorates, and could mean a Labor two- party-preferred vote of 136 that is 54 percent, and in Robertson of 117 — that is, 46 percent or 170. Or 160 and 93. And thus we are being told that Labor loses two seats, not one.
The unscrupulousness of this approaches the wicked.
Rudd’s campaign is being shown to be ‘running off the rails’ because of these figures, whose sample is eleven hundred less than it should be, and whose average age, unstated, is probably 70. Rudd is seen to be running scared from figures that have been, effectively, made up.
I ask again why the sample was 1,100 too few, why no mobiles were rung, no voters confronted in shopping malls, and what their average age was.
Refusal to answer is criminality.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Australia License