How close will that election be?

November 30, 2009

Labour is clawing back support and we may end up with a hung Parliament, according to a recent poll in the Observer [new window].

The IPSOS Mori survey puts the Tories on 37% and Labour on 31%, slashing the Conservative lead from 20 points in some polls last year. If the nation voted that way at the election, it would give David Cameron’s party the largest number of seats but still 30 or so short of a majority.

As with most news organisations, the Observer reports some details about the poll itself – 1,006 respondents were interviewed by telephone. But we aren’t told the margin of error. While this is common practice, it does disguise one very important fact – that the poll might be evidence that Labour support is actually ahead of that for the Conservatives.

How can that be? Well, luckily, IPSOS Mori do give the margin of error for this poll on their website,  where they caution that: “the ‘margin of error’ on these figures is c.+4% for each figure”.

What does that mean for the level of party support? On this margin of error, the Conservatives are in the range 33%-41% and Labour is in the range 27%-35%. Hence it is possible – though unlikely – that Labour support is 35% and the Tory support is 33%, putting Labour marginally ahead.

It is quite true that this is a slim possibility – both measures would have to be at the extreme endof the error bound. However, in order to understand the result fully, it is important to be aware of this possibility.

As IPSOS Mori say, “This is especially important to keep in mind when calculating party lead figures.” Of course, it is impossible to keep it in mind if you are not told about it. News organisations ought to include it as a matter of course, along with the sample size.

A further point of interest is that although IPSOS Mori interviewed 1,006 people, the result about party support is only based on half that number – this is because only the 513 people who said they were certain to vote in the next election were asked the question.

There is good reason for this – IPSOS Mori explain that, “We regard the voting intentions of those ‘absolutely certain to vote’ as the most useful trend indicator, since it includes only those voters whose frame of mind is nearest to
those who actually vote at elections.”

But when you look at the result of the full 1,006 respondents, the picture is that Labour and Conservatives are on absolutely equal footing, at 34% apiece.  (Note that this isn’t at all at odds with the headline figure, since Labour support is over-estimated when taking all potential voters into account).

So of the two key figures in any poll, one (sample size) is misleading and the other (margin of error) is missing.

I’m not suggesting there is anything wrong with the Observer article – it is almost certain that the Tories are ahead, and it’s standard practice to count only the views of cast-iron voters when it comes to party support. But the article would be fuller if it included that information or – on the website at least – if it linked to the full data set.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: