A word about leadership polling
The Labour leadership bandwagon seems to be rolling again. So, do the polls tell us anything about whether Labour would be better off without Brown? The answer, I'm afraid, is often no.
Questions in polls about alternative leaders normally take two forms. The first is a straightfoward "who would you like to be leader of X", which is normally no more than a recognition contest. People like Jack Straw invariably head the list since people recognise their name, whereas few if any know who James Purnell or Jon Cruddas, for example, are. They should never be taken as a sign that, were people like Purnell or Cruddas as widely recognised as Straw, they wouldn't be as popular as him.
The same problem also affects the other main type of "pre-leadership election polling" - hypothetical questions asking how people would vote if X became leader of the party. Generally speaking people simply don't know enough about the alternate leaders and lots of people who might well consider voting Labour with a new leader accurately say "don't know", giving the impression that the party would do badly under less well known leaders when it isn't necessarily the case.
Even when the alternate leader under consideration is well known there are still problems, because people are not very good at imagining how they will react to future events. Take the example of Gordon Brown, prior to his election polls universally suggested that he would perform very badly indeed in comparison to Tony Blair. Yet most observers correctly predicted that he would in fact receive a healthy boost upon becoming leader. The public were answering the question simply by transferring Gordon Brown's head onto the existing Labour government, with everything else remaining the same. In reality a new Prime Minister also brings changes of policy, emphasis and party image - and in the case of Gordon Brown at least a huge swell of sympathetic media coverage.
Gordon Brown is an interesting example though because in the long term he has done very badly in the polls. Does this suggest the polls were right all along, and hence would be just as useful this time? The answer is partially yes, but no. I wrote at the time that those hypothetical polls probably were a very bad sign for Gordon Brown because of the reasons behind them: he wasn't polling badly because people thought he was weak, incompetent or had the wrong ideas, to the contrary, people back then still had extremely positive perceptions of Gordon Brown as a strong, experienced and capable Chancellor. The reason he polled so badly appeared to be that people just didn't like him.
I said back then that a politican could probably make themselves look strong or capable by performing well in office, but if they weren't likeable there was little or nothing they could do about it. Brown's figures boded ill for him because they appeared to be based on something that he couldn't correct. In contrast, if we see polls showing Alan Johnson, James Purnell, Jack Straw or A.N.Other wouldn't do as well as Gordon it doesn't follow that it is for the same reason. It may be because they aren't well known enough, or people don't think they'd perform well in office, or haven't got good ideas or other percieved failings that a potential leader could correct in the future.
I would urge great caution in reading too much into these sort of leadership polls if they do turn up, which given the recent news coverage of the Labour government, should be quite soon! I would not, however, ignore them completely - while I don't think they show much, they do seem to carry a lot of weight with MPs and the media, in the early part of 2007 there was a period where the media would build up an alternative to Gordon, eventually do some polling on them, find they wouldn't do any better than Brown, and move on to the next alt-Gordon. So, even if leadership polls like this are rubbish, they shape the debate - just wait till the day comes when a poll does finally show X performing significantly better as Labour leader than Brown would.
