Tag Archives: polling

Media Fail

Over at the Spectator, Alex Massie analyses an Ipsos-Mori poll on some of the beliefs the British people hold about our country and the way the Government operates.

To pick just one statistic:

Then there’s this doozy: 29% of respondents think the government spends more on unemployment benefits – the Job Seekers’ Allowance – than on pensions. In fact, pensions cost 15 times as much (rightly so!).

Apparently the British public is similarly misinformed about the propoertion of Muslims in the country, the number of immigrants, the budget of the Department for International Development, the rate of teenaged pregancies, and the crime rate.

Massie shies away from calling the British people stupid, though the headline to his post ‘Abandon All Hope’ suggests this is what he thinks.  And clearly this poll shows that we are, indeed, a highly ignorant bunch.

But I think it would be wrong for we the public to take the blame.  To my mind, this poll shows that we are being failed by our media.  Their primary purpose is surely to keep us informed about what is going on in our democracy, and clearly they are not doing that job effectively.  That’s not something that Alex Massie cares to consider in his article.

Neck and Neck in the Polls?

This week, the BBC reports on the US Presidential this week have been consistently reporting the race “neck-and-neck”. This assertion is grounded on opinion polls: the latest BBC report trumpets an ABC News poll which places both candidates on 48% of the vote.

The problem is, one poll does not tell the whole story. Each polling outlet has a slightly different methodology which skews the results. For example, some poll only ‘likely voters’ and some ask everyone; some pollsters call cell-phones, while others use only landlines.

As a foreign media outlet, the BBC is not covering the race with the granularity of the domestic US media.

Rather than report the result of one poll, The corporation would do better to report on the polling averages between polling outfits, and the trend-lines of generated by each pollster over a given period. Both these macro views look better for President Barack Obama, whose polling in the last fortnight has been improving. More importantly, State Level polling shows the President ahead in battleground states like Ohio. Poll analysis site Five Thirty Eight (hosted by the New York Times) models the election on this basis, and is currently putting the chances of an Obama win at 85%.

This does not mean that Obama is coasting towards a second term. Governor Mitt Romney could still win. But given the totality of the polls, President Obama can be said to have the advantage. The situation does not really warrant the metaphor “neck-and-neck” which suggests either horse is equally likely to win. In the betting markets, Barack Obama is the clear favourite. He hasn’t won, but he is ahead.

So why does the BBC cherry-pick a single poll as its headline? Simples: “neck-and-neck” is a more sensational headline than “Obama ahead”. And the more sensational headline will deliver more viewers on Election Night.