By-elections are exciting; they involve real decisions by real voters, so their significance is liable to be exaggerated. The 22.5 per cent swing from Labour to the Scottish National Party in Glasgow East is most unlikely to be repeated in the next general election. Usually, opinion polls have proved much better than by-elections at indicating long-term trends of public opinion.

Two national polls have been published since the Glasgow result was known; both were taken before the by-election. That may have had a shock impact over the past weekend. Labour's position may be even worse than in these opinion polls, but they are the best available guide so far.

On Saturday The Times and The Independent both published polls. Ipsos MORI in The Times gave figures of 47 per cent for the Conservatives, 27 per cent for Labour and 15 per cent for the Liberal Democrats. ComRes in The Independent gave 46 per cent for the Conservatives, 24 for Labour and 18 for the Liberal Democrats. These polls tend to confirm each other on the Conservative position, but differ on Labour and the Lib Dems.

When compared with the last general election, Ipsos MORI shows a swing of 11.5 per cent from Labour to the Conservatives, while ComRes shows a slightly bigger swing of 12.5 per cent. Ipsos MORI shows a swing of 11 per cent from Lib Dem to Conservative, while ComRes shows a smaller swing of 9 per cent.

They do not show any swing even nearly as large as that the SNP actually achieved in Glasgow East. Nevertheless, a series of by-elections, local elections and opinion polls all point in the same direction. These are the sort of results that predict a landslide. On present polls, the Conservatives can certainly expect a substantial overall majority at any early general election. Labour could lose as many as 200 seats.

For the Labour Party, this is a dangerously unstable situation. If it is thought that 200 Labour MPs could lose their seats, it follows that 300 or more must be afraid that they will lose their seats. The most sensible course might be for Labour to avoid panic and be prepared to go on until a general election in 2010.

Yet Labour MPs are already paranoid about their seats and those in the grip of paranoia do not make the best decisions. Unless there is a strong Labour recovery, of which there is as yet no evidence, Gordon Brown will be under continuous political pressure, just as John Major was before 1997. Such anxiety always reveals any splits that exist in a party.

In John Major's case, it revealed Conservative splits over Europe. In Gordon Brown's case, there is always the division in Labour between the old and the new, with the trade unions now reassessing their power. Old, trade union-dominated Labour would be even less electable than Labour is now.

There will inevitably be continued pressure for the leader himself to be changed. That pressure did not go away in the Conservative Party in John Major's period before the 1997 Labour landslide. Of course, economic conditions could change people's attitudes, but the Prime Minister has no reason to rely on an economic recovery in the next two years.

Early reports suggest that there are four front-runners for the Labour leadership. More might join later. Harriet Harman, the most convincing of the women candidates, has attracted similar support to that of Hillary Clinton in the US Democratic primaries. One can see why she is attractive to the same groups as Senator Clinton. Senator Clinton fought a tough and professional, but somewhat left-wing, campaign that impressed even those Republicans who would never dream of voting for her. Ms Harman fought an almost equally professional campaign for the Labour deputy leadership, with the added bonus that she won.

In America, hundreds of thousands of women, mostly of Senator Clinton's own age group, became passionately attached to her cause. When she lost, many felt personally betrayed. I am not sure that there is yet as strong a movement for Ms Harman, but it could easily develop. Women often feel a strong personal loyalty for a good and courageous woman candidate.

I do not think it matters that Ms Harman has a posh accent - Tony Benn has the poshest accent of modern politics and that has always been well liked by his supporters. I think that she would get strong support on the Left and in the trade unions as well as among old Labour voters and women. It is true that Ms Harman is disliked by some inside and outside the Labour Party. But strong candidates always have their enemies and their critics.

Labour has to change the climate of debate. It seems to me that a woman candidate would, from her gender, have a better chance of doing that than any of the men.

I have had considerable admiration for Jack Straw's work. I have always seen the necessity for good, professional politicians, who can get most of their judgments right over a career lasting for decades. The political hero of my youth, Rab Butler, was such a character. But these are the consolidators of political life. Mr Straw has been in high office every step of the way since 1997. If the voters want to clear out the old guard of Labour, why should they hesitate to get rid of him? As a senior minister, he is admirable, but he would not pretend to be a novelty.

David Miliband might be promoted as a British Obama where Harriet Harman is the British Hillary. Mr Miliband's trouble is that Senator Obama is beyond his level; the comparison could be damaging.

Finally, there is Alan Johnson. He seems a very nice guy; so did John Major. Nice guys cannot be expected to turn off the taps of Niagara.

Labour's choice should be made between sticking with Mr Brown until May 2010, which might be the wisest course, or switching to an early election under Harriet Harman - a high-risk strategy, but not unthinkable. It would make some sort of political sense.

Times Recommends

• Copenhagen will fail – and quite right too
• They’re still wriggling to avoid the flak on Iraq
• Rule the waves? Not any more we don’t