Tuesday, August 25, 2015

Correction to Opinion Poll post

Latest opinion poll result, from COMRES for the Daily Mail, should have read as follows.

Conservatives 42%
Labour 28%
UKIP 9%
Lib/Dems 8%

Conservative lead 14%


Earlier today there were versions of this poll result floating around giving Labour a lower share which I quoted in good faith. I then noted that other people were quoting different numbers.

If you look at the  COMRES's website at

http://comres.co.uk/polls/august-2015-daily-mail-political-poll/

it gives the correct figures as above.

Apologies for any inconvenience.

2 comments:

Jim said...

an interesting post, if only because it pretty much forces me to comment and ask, why do you now trust opinion polls?

Chris Whiteside said...

If by "trust" you mean "blindly assume everything they say is perfectly accurate" then in that sense I most certainly don't trust them.

I did actually say in the original post to which this is a correction that, quote "I know we've just been given good reason not to put too much faith in the polls at the moment."

I think that until the polls have demonstrated that they have corrected some of the errors they fell into during the run up to the 2015 general election, we should be particularly careful to remember that they might have a much bigger margin of error than pure statistics might suggest.

The error will not necessarily be infinity, which means that if a poll suggests a significant movement, and that indication both makes sense and is consistent with other evidence, it may be telling you something real.

It makes sense that a party which stages a huge public fiasco like the Labour party leadership election might lose support, and the finding that the Conservatives are doing reasonably well at the moment is consistent with the evidence that two of the three council by-elections on Thursday were Conservative gains.

So I think this poll might be right. That does not mean I would bet my shirt on it !