Commons call for an investigation into the pollsters – now a big test

The group of leading Labour MPs that has put down a Commons motion calling for an investigation into opinion polling firms over fears that ‘questionable methodology’ is “skewing the political process and tarnishing the industry’s reputation” will be able to get some empirical information next week on the effectiveness of the internet-based pollster, YouGov. For they will be able to compare YouGov’s predictions on the Elections with how the conventional pollster, ICM, fared four years ago when one of the MPs, Frank Dobson (above), was Labour’s candidate.

ICM 2000 (Final week)

CON 27% (ICM) 33.2% (actual) +6.2%
LAB 39% (ICM) 31.6% (actual) – 7.4%
LIB 20% (ICM) 18.9% (actual) – 1.1%
Livingstone 51% (ICM) 39.0% (actual) – 12%
Norris 17% (ICM) 27.1% (actual) +10.1%
Dobson 16% (ICM) 13.1% (actual) – 2.9%
Kramer 12% (ICM) 11.9% (actual) – 0.1

YouGov 2004 (Two weeks to go)

CON 34% up 7% on 2000 poll
LAB 28% down 11% on 2000 poll
LIB 16% down 4% on 2000 poll
Livingstone 41% down 10% on 2000 poll
Norris 27% up 10% on 2000 poll
Hughes 17% up 5% on 2000 poll

Politicalbetting.com fully supports the MPs call that there should be more “transparency”. A number of pollsters, notably ICM, are good at publishing all the information from their surveys so that if there are issues then people can see the data that things are based upon. It is solely because of this that we have been able to discuss ICM’s approach to dealing with those who are not “absolutely certain to vote” .

We wished YouGov did the same. Thus three weeks ago the firm ran a London Mayoral Election poll showing that amongst those “certain to vote” Steve Norris was just 5% behind Ken Livingstone – vital information that saw Norris’s price tighten dramatically.

In last week’s YouGov poll, above, we were not given this data although YouGov must have collected it because it was able to produce an overall “intention to vote” figure. For that election, too, information on second preferences is critical and again this is not being published.

The result was a hardening of the Livingstone price. If we had had all the data then this might not have happened – or the price might have got even harder.

For political gamblers the polls are vital because they are our primary way of finding what’s going on and we often risk our money on what they say. The General Election and the current London Mayoral election person or party markets are driven by the polls.

If we were investing in our judgement on stocks and shares, rather than political outcomes, then there is a whole raft of regulation to ensure that the information we receive is fair and accurate. Just look at the way one of the oil giants is being pilloried over its oil reserve figures. Gambling is different from investing but with the opinion polls there is not even a requirement for tranparency.

One of the main objectives of Politicalbetting.com is to analyse and try to guide political gamblers on the polls and make statements where we think there are issues. Thus in recent weeks, as well as our comments on ICM, we’ve questioned the new Populus strategy of compensating for “Shy Labour supporters”.

Of all the pollsters we have become impressed with the results that YouGov produce political and they seemed to have dealt best with the issue of “shy Conservatives” who won’t tell interviewers what their views are. The proof of the pudding is in the eating – so next week will be really interesting. But YouGov do need to be more transparent and this, we believe, would help deflect some of the criticism such as that coming from the MPs who have signed the Commons motion.

US ELECTION UPDATE A new poll shows that Kerry would beat Bush by 53% to 39% if the Arizona Republican Senator, John McCain, was his running mate. Latest prices.