Polls: Should we focus on of the few – not the many?

Polls: Should we focus on of the few – not the many?

Will Green-gate affect anybody’s vote?

Although there is still the odd story coming out the focus of the Westminster village has moved on from the Damien Green affair and it’s hard to see it bubbling up again.

But what of the impact on the voters? No doubt Brown Central was cheered by the above figures from this week’s Populus poll. These showed 45% agreeing that “this is one of those typical Westminster arguments that bear no relation to the lives of ordinary people ” compared with 39% who felt “There are important issues at stake in this case and it is right that politicians should be giving it so much attention”.

    But here as in other polling questions on issues the critical matter is whether these will affect people’s voting. Concentrating on the many and not the few can sometimes lead us to the wrong conclusion.

The fact that the proposition here was, as it were, defeated, does not in any way affect the minority – a factor that pundits quite often ignore. Drilling down to the detail can sometimes be illuminating.

Just look in the table above of the subset of those who voted for one of the three main parties in 2005. They were much less likely to say “don’t know” and registered bigger scores on each question.

But then look at the final column the views of those who voted Labour in 2005 but now say they are voting Tory – a segments that Gordon needs to eat into between now and polling day. There are not many of them but their split, which I’ve highlighted, is, I would suggest, telling. Yes the numbers are small and it would be great if we had a mass of data from a large number of LAB>CON switchers but we don’t. This is all we’ve got.

That could represent real voters which could be set not to move back.

I’m making a general point here not just a specific one on the Green affair. On responses to issues the key question is whether these are swinging votes at the margin? If so then that’s important – if not then we can probably ignore them.

UPDATE: For those who argue about the significance of such a small sub-sample please read the comment from Timothy (likes zebras) here in which he argues: “..that this is significant at the 90% level and the null hypothesis is rejected – ie this issue is more important for people who have switched their vote from Labour to Tory, than for the electorate as a whole.”

Mike Smithson

Comments are closed.