Since the victory of the President, Mrs McAleese, some newspapers have carried reports that the Government is to consider a ban on the publication of opinion polls in the fortnight before elections. This followed a claim by Mr Desmond O'Malley that many who felt the result was known before polling day failed to turn out.

In debating the issue with him on RTE radio last week, I agreed there was a possibility that some could indeed have been so influenced since the track record of pre-election opinion polls in Ireland has been one of accuracy and reliability. However, the issue is much wider.

During the campaign, four opinion polls were conducted by MRBI for The Irish Times, and on polling day Lansdowne Market Research conducted an exit poll for RTE. IMS also conducted a series of campaign polls for the Independent Group, all of which provided figures which were very consistent.

Commenting on the penultimate MRBI survey on October 22nd - eight days before the election - I wrote that Mary McAleese would almost certainly head the poll and remained favourite to be elected.

In the final MRBI poll, conducted with three days remaining, I concluded that Mary McAleese would be comfortably elected on the second count, on the elimination of the three lowest candidates, and suggested reasons why her first-preference support could be marginally overstated.

In passing it needs to be said that, not for the first time, the opinion polls have been misinterpreted. It is a well-known and frequently quoted axiom that an opinion poll is a snapshot on the day, and this certainly applied to the successive campaign polls which monitored candidate support.

While it's fully acknowledged that the RTE/Lansdowne exit poll was very accurate, the senior RTE presenter who said this was the only poll which did not understate Dana's support was off target. The comment would have been valid only if all polls to which he referred had been conducted on the same day. It conflicts with the axiom quoted above and carries a nonsensical implication that all polls should have identical findings, irrespective of the timing of each.

The final Irish Times/MRBI poll three days before the election positioned Dana in double figures for the first time and on RTE Radio 1 forecast that Dana would finish in third place on first-preference votes.

However, to return to the main issue. It is possible that some likely supporters of the lower-placed candidates were influenced not to vote; if this happened, I suggest it was for a combination of reasons.

Looking back to 1990, when the turnout was 64 per cent - the highest for a presidential election - the campaign polls provided similarly reliable indicators of the outcome. At first, Brian Lenihan held a commanding lead which, for reasons that are well documented, evaporated as the campaign developed. In the final Irish Times/MRBI poll, Mary Robinson had reduced the lead very satisfactorily; and my conclusion that she would be elected on the second count, on Austin Currie's transfers, was accurate.

What were the variables between the two elections? The first is that this year there were five candidates. Four were women; and two were nominated by the alternative to the party system. Only one of the five claimed to be a member of a political party.

From the outset and throughout the campaign there was considerable opinion-poll evidence of widespread cross-party voting intentions: considerably fewer party supporters, than one would expect, intended to vote for their party's candidate; a majority were noticeably influenced by the candidate's personal attributes rather than by political affiliations.

This is supported by the findings of the RTE/Lansdowne exit poll which showed that almost half of those who voted had not made up their minds at the outset of the campaign. It should also be seen against the background of the general election, where the combined first-preference vote of the parties which nominated Mary McAleese, Mary Banotti and Adi Roche amounted to 85 per cent.

THE second variable was the generally unfavourable political background following the recent tribunals which identified the extent to which a few senior people in the political arena had received financial contributions for purposes other than political campaigning.

From the outset and throughout the campaign there was considerable opinion-poll evidence that the electorate generally saw this election as much less political than was the case in 1990. The satisfaction ratings of all five party leaders dropped significantly over the three-week period.

It was evident, therefore, that many were unimpressed by the accusations and counter-accusations of senior politicians at a time when all candidates maintained a noticeable degree of courtesy towards each other. In my opinion, prior to and during the campaign, the political establishment lacked both credibility and influence, and voters got on with the task of assessing the five candidates based on their personal characteristics and attributes.

In these circumstances, it is not surprising that many adopted an apathetic stance or were more susceptible than heretofore to the influence of fellow-electors. A further revealing factor is that 66,000 voters decided to send a reproachful message to the politicians by spoiling their referendum ballot papers.

It is ironic that the achievement of the opinion polls in maintaining their professional credibility during the campaign is now seen in a negative context, and some politicians - and journalists - are calling for a ban on polls for a specific period before elections.

Many of the comments are both confusing and inaccurate. Confusing because I am not aware of any country where opinion polls per se are banned; inaccurate because the commentators who have made public statements are not knowledgeable as to the actual situation.

One commentator wrote that `most of our European neighbours ban polls in the last week' and another `most countries, including France, ban polls for weeks before elections'. Both statements are inaccurate. To my knowledge, four European countries ban the publication of opinion polls for the final fortnight or longer; five, including France, ban publication for seven days; and 15 either have no ban on publication or do so for the final day only.

I see nothing undemocratic about electors being made aware of what their fellow-electors are saying or what opinions they hold, and if some have been influenced by the findings of opinion polls, it indicates that the market research profession has a higher level of credibility than that of politics.

When the political tribunals complete their hearings and present their reports; when the profession of politics has dusted itself down and restores its level of credibility, then the electorate at large may again lend its ears to the political establishment, and the influence of what fellow-electors think or say as relayed by the opinion polls, will be balanced by the views of politicians and their parties.

Jack Jones is chairman of MRBI Ltd