Today�s poll for The Herald offers a glimmer of hope for the Labour Party. But it also brings dilemmas. Does it prove that it works to be unrelentingly negative against the SNP?

James Mitchell

Today's poll for The Herald offers a glimmer of hope for the Labour Party. But it also brings dilemmas. Does it prove that it works to be unrelentingly negative against the SNP, or could this have something to do with a more positive mood following Gordon Brown's Budget?

Fieldwork was carried out between March 22 and 27, after the Budget had been well received in much of the media. The results in this poll agree with the pro-Labour bounce noted in another poll, conducted for the Independent newspaper.

This poll is more in keeping with Labour's support at the start of previous elections, and Labour can take comfort from the knowledge that it is generally the case, though not always, that the fortunes of governing parties tend to improve as the election approaches.

The dilemma will pitch those who think the attacks on the SNP are finally working, and should carry on as before, against those who think the positive approach after a well received Budget might pay better dividends.

Without doubt, negative campaigning works, but it usually helps to combine this with a more positive message.

Even in this poll, the SNP is sitting around 35% in both constituency and list votes, up more than 10 percentage points in both compared with 2003. As disappointing as it will be that the party has slipped behind Labour, the SNP looks reasonably assured of an increase in its representation in Holyrood. If this poll were replicated on May 3, the Nationalists would achieve their best ever performance. But second best will not be good enough.

The poll tells us much about the polarisation of Scottish politics. In 1999, the combined Labour and SNP vote was 67% for constituency seats and 61% for list candidates. In 2003, the respective figures were 59% and 50%: Election 2003 was the year of the "others".

However, in this poll, the combined Labour and SNP constituency and list votes are 72% each, indicating that the other groupings are being squeezed in the battle between the two main parties.

This election looks set to see fewer MSPs returned for other parties, but, ironically, those who are returned may be more powerful. If, as looks quite likely, no two parties will have an overall majority in Holyrood, the left, Greens, and independents who are returned will be the kingmakers. This seems most likely to be the Greens. We may be looking at a three-party coalition or minority coalition rule.

Care needs to be taken in interpreting poll support for smaller parties such as the Greens. The tightly fought nature of this campaign does not appear to be encouraging voter engagement. A high proportion of the electorate is undecided or refusing to express a preference. Those who are undecided are greater than those who have expressed a preference for Labour, SNP, and the Liberal Democrats put together.

The level of turnout may prove little higher than last time when just under half the eligible electorate voted. But it is early days yet, and there is plenty of time for voters to make up their minds.

The details of this poll suggest the SNP continues to poll better among men than women. It out-polls Labour among men on the constituency vote, but falls behind Labour by a larger margin among women. Alex Salmond's less combative approach in this election may be an attempt to rebrand himself and his party as more women-friendly, but so far it is not paying an electoral dividend.

But Mr Salmond has less to worry about when it comes to voters' preference for First Minister. He commands the support of just under half those intending to vote, with Jack McConnell lagging 10 percentage points behind. Mr McConnell would appear to be a drag on Labour's support. It is little wonder the SNP is contesting this election in a semi-presidential manner.

Consistent with previous polls, support for independence lags behind support for the union. This will be disappointing for the SNP, especially in a poll that asks how people would vote in a referendum on independence.

Polarising choice in this way still helps the SNP, in that the 46% who favour independence offer the Nationalists a substantial potential vote. Other polls show that support for independence is considerably reduced when a range of options are offered.

The pro-union finding of the poll will be welcome news for Mr McConnell. But this election is not a referendum on independence, and other parties are contesting the election. If it was fought exclusively on the constitution, the pro-union vote would be divided among a number of parties.

But, as we have seen across all elections, people will vote on a range of matters and it is far from certain that the constitution is uppermost in people's minds.

James Mitchell is professor of politics at Strathclyde University