Fly Away Peter

Peter Mandelson's resignation last week was an embarrassment to the government, and naturally led to much speculation that the government is acquiring a sleazy image that will damage it at the election. Two polls conducted since the resignation have explored this. Unfortunately the reporting of the second of these, an NOP poll for the Channel 4 programme Powerhouse, has misinterpreted its findings and perhaps as a consequence its implications.

The NOP poll, published on Tuesday, was reported in the Evening Standard that day and in every quality daily the following morning. Peter Kellner's report of the poll on Powerhouse was, as always, accurate and incisive; but every one of the newspapers that reported the poll, with one exception, misinterpreted what it said.

NOP repeated a question that Gallup asked just before the last election:

Q Do you agree or disagree with the following statements:

"The Conservatives these days give the impression of being very sleazy and disreputable." "Labour these days give the impression of being very sleazy and disreputable."

Source: NOP/C4 Powerhouse

First, note the full wording of the statements tested. NOP asked about the impression that the parties give, not the reality. In the age of spin, this is hardly an academic distinction. (How ironic, indeed, that the press should blur the distinction when the story they are reporting is the sacking of the original and best-known Labour spin doctor!) But even ignoring this, almost all the reporting of the poll made a further - and very common - error, by confusing the prevalence of opinions with their intensity.

Labour's image on "sleaze", unquestionably, has deteriorated since 1997, while that of the Tories has improved somewhat. As many people think that Labour is "sleazy" (or, to be precise, think Labour "give the impression of being very sleazy and disreputable") as think it of the Tories. That is a powerful finding. You can even at a pinch justify, although the difference is within the margin of error, saying that more people think Labour is sleazy than think the Tories are. What you can't say - because the question didn't ask it - is that the public think that Labour is "more sleazy" than the Tories. Yet that is exactly what the newspapers told their readers that it meant. Several, indeed, headlined that version, repeating the misreading in the main text:

LABOUR MORE SLEAZY THAN TORIES, SAYS POLL "Labour is seen for the first time as a more sleazy party than the Tories…" (Patrick Hennessy, Evening Standard)

LABOUR SEEN AS SLEAZIER "Labour is now seen as more 'sleazy' than the Tories, according to an opinion poll …" (George Jones, Daily Telegraph)

LABOUR SEEN AS SLEAZIER THAN TORIES " … Labour is now perceived as more sleazy than the Tories" (Jenny Percival, The Scotsman)

The rest reported the poll as part of the wider story under a different headline, but made the same error in the text:

"an opinion poll suggested yesterday that … Labour is seen for the first time as a sleazier party than the Tories" (Brian Groom, Financial Times)

"a poll emerged suggesting that voters regarded Labour as more sleazy than the Conservatives" (Patrick Wintour, The Guardian)

"…a new poll found that voters viewed Labour as more 'sleazy' than the Tories …" (Paul Waugh, Independent)

Philip Webster in The Times at least avoided reporting a two-point difference in findings as "more than", but still explained it as a question of intensity:

"…a poll suggesting that the public now perceives Labour to be as sleazy as the Tories"

So congratulations to Michael Settle in The Herald, the only one of the broadsheet writers who covered the poll without suggesting that it measured a perceived degree of sleaziness.

The distinction is a significant one.

If a lot of the public think both parties are (equally) sleazy, then it provides no means to choose between them and perhaps encourages people not to bother voting at all. (Indeed, it is perfectly possible from the NOP data that while a lot of people think both parties are sleazy, most still think the Tories are more sleazy). On the other hand, for anybody who thinks that Labour is more sleazy than the Tories, that might be a rational motivation for voting Tory (though of course it would depend on how much sleazier the government is thought to be, and voters may have many other considerations they think are more important).

So what was the effect? Has the Mandelson affair, as many of the columnists seem to assume on the basis of NOP's figures, damaged the government's support in the country?

The immediate answer seems to be that it has not. MORI's regular monthly poll for The Times came out of the field last Monday. Peter Mandelson resigned on Wednesday. On Thursday and Friday we polled again, for the Mail on Sunday. Look at the voting figures - no change.

Q How would you vote if there were a General Election tomorrow? Q (If undecided or refused) Which party are you most inclined to support?

Source: MORI/The Times/Mail on Sunday

n.b. Voting intention figures given in the above table are, as is usual, the "repercentaged figures", i.e. figures given for each party are percentages of those naming a party, while the "don't knows" (would not vote, undecided or refused) are percentages of the entire sample.

That is not to say that the public does not disapprove of last week's revelations, as the details of our Mail on Sunday poll show. Three-quarters think the affair has damaged the government's reputation, and three in five that it has damaged the Prime Minister's. But on this matter, as on so much else, the government is insulated from real damage because even more of the public thinks badly of the Tories than do so of Labour. Our poll DID include a question measuring relative sleaziness of the parties, although we asked respondents to make the comparison between Mr Blair's government and the last government, rather than the present Conservative party. More than half thought that there was no difference between the two, and only one in eight that the Blair government is more sleazy than as the Major government.

Similarly, just as the voting intention ratings were not moved by the Mandelson resignation, nor were the government's satisfaction rating or Mr Blair's personal rating.

Q Are you satisfied or dissatisfied with the way the Government is running the country? Q Are you satisfied or dissatisfied with the way Mr Blair is doing his job as Prime Minister? Q Are you satisfied or dissatisfied with the way Mr Hague is doing his job as Leader of the Conservative Party? Q Are you satisfied or dissatisfied with the way Mr Kennedy is doing his job as Leader of the Liberal Democrats?

Source: MORI/The Times/Mail on Sunday

Interestingly, though, while the approval ratings of the government and of Mr Blair remained static over the period of the Mandelson resignation, those of both Mr Hague and Mr Kennedy rose significantly - indeed, both recorded the highest percentage satisfied that MORI has found in any poll since they became party leaders. (Mr Hague's previous best was 29%, Mr Kennedy's 42%. Possibly this reflects their respective performances in the resignation debate, although as they took such different approaches to it, it is surprising that both should benefit.

But what should be of much more concern to the government is not the risk of their supporters switching to other parties which are perceived as less sleazy, but of their not bothering to vote at all. Here the steady decline in so many aspects of the government's reputation could do real damage, and the situation seems tailor-made for an election dominated by highly effective negative campaigning. (The Tories' new "You paid the taxes …" campaign may not convince a single voter that they would do a better job of running the country, but two Labour supporters persuaded to stay at home on polling day are exactly as good to the Tories as one that switches and votes for Mr Hague.) Both MORI's poll for the Times last week and ICM's for the Guardian found fewer people at this stage certain that they will vote than even at the same stage before the 1997 election, which had the lowest turnout since the War. If the turnout pattern follows the pattern of certainty to vote in those polls, Labour will suffer.

Q How would you vote if there were a General Election tomorrow? Q (If undecided or refused) Which party are you most inclined to support? Q Can you tell me how likely you are to get along to vote at the next election?

Source: MORI/The Times 18-22 January 2001 Base for voting intention: 2,083 British 18+ Base for likelihood of voting: 1,104 British 18+

Every event like the Peter Mandelson resignation (whatever the truth of what really happened), whether the public sees it only as smearing the government or as an indictment of the entire political system, can only make it likelier that turnout will fall further still, and probable that Labour seats will fall with it. That is the real consequence of the last week's events.