Sir, - The die is now cast and the electors north and south have handed down their sovereign judgment on the Northern Ireland Agreement. During the referendum campaigns, the public have been kept informed of the ebb and flow of dispositions towards the Agreement in both jurisdictions, through polls commissioned by The Irish Times, The Irish In- dependent, The Sunday Independent and by RTE. In the case of polls conducted by IMS and by Lansdowne Market Research in both jurisdictions and by MRBI in the south, the public have every reason to be confident in the findings published in the major commissioning media, based as they are on large and statistically robust representative samples, using the proven methodology of face-to-face interviewing of electors in their homes. In addition, however, The Irish Times commissioned and reported the findings of three polls carried out in Northern Ireland by the British company, Harris Research Centre.
The methodology adopted for the conduct of these polls gives rise to serious questions.
IMS have been responsible for conducting polls in Northern Ireland for over 20 years. Irish professionals are acutely aware of the stringent methodological precautions which need to be adopted in conducting political research in the uniquely sensitised environment of Northern Ireland. Our panel of interviewers in Northern Ireland is very carefully selected and trained. They conduct interviews within the communities from which they themselves are drawn. They are equipped with photo-ID cards and engender confidence in prospective respondents, through their physical presence and unthreatening demeanour. The net result is that refusal levels are very low. It is also important that sample sizes are sufficiently large to provide an adequate degree of statistical reliability. IMS considers a simple size of 800 interviews as being the minimum appropriate level to ensure an adequate depth of cross-community representation in Northern Ireland.
The three Harris polls commissioned by The Irish Times were carried out by telephone, from the company's call centre in London. Each one involved a sample of 500 people. Prospective interviewees were confronted with a request to expand upon their political opinions, by a faceless voice expressed in a London or British regional accent.
Against a background of decades of suspicion and mistrust, it requires little imagination to imagine the response of many inhabitants of the Ardyone, South Armagh, the Springfield Road or Portadown to being asked to openly express their political opinions to an unknown individual at the other end of a telephone line. Until all too recently, such indiscretion might literally have posed a threat to life and limb.
I hypothesis that polls conducted in this manner will over-represent the benign middle ground of opinion, and under-represent those whose political views veer in either direction towards the nationalist or loyalist poles of prejudice.
The final Irish Times/Harris was conducted on Monday and Tuesday of this week and suggested a very positive prognosis as to the likely "Yes" vote. Against the background I have described, I suggest that the poll will have overstated positive dispositions towards the Northern Ireland Agreement, as they prevailed at that time. Ironically, this does not preclude the possibility that the actual outcome of the vote, might tend towards approximating the pattern reported as at the beginning of the week. If that proves to be the case, however, I suggest the primary reason for such an outcome will have been the saturation coverage given to proponents of a "Yes" vote throughout the course of this past week (ranging from Bill Clinton and Tony Blair to Bono and Richard Branson). The "No" campaigners effectively shot their bolt almost a fortnight ago.
Professional Irish researchers, both north and south, will be saddened that The Irish Times should lend its reputation to unproven and questionable techniques brought to bear by a British company with negligible experience of or evident senstivity towards the very special requirements of conducting effective political polling in Northern Ireland. - Yours, etc., Charles M Coyle,
Director, Irish Marketing Surveys, Upper Pembroke Street, Dublin 2.