The old polls act
Sir: In your issue of 2 May, Peter Pater- son raises a number of questions about opi- nion polls with particular reference to the quality of interviewing. The implication of Mr Paterson's comments is that polling organisations do not care to disclose in- formation on this score. This is certainly not true of NOP. We have on a number of occa- sions publicly stated our policy and pro- cedures in respect of interviewer recruit- ment, training and control and we are per- fectly willing to restate these to anyone who cares to ask. Briefly, however, interviewers are selected by a personal interview with one of our full-time supervisors and are trained by that person, initial training normally last- ing about three or four days. No minimal educational standards are laid down and though a degree of literacy is important. the main qualities in which we are interested are those of personality and ability to com- municate. So far as interviewers' political beliefs are concerned, we do not engage any- one as an interviewer who is a member of a political party, and we continuously check work for political bias. The ratio between interviewers and supervisors on a political survey averages 10 interviewers to one supervisor. 10 per cent of an interviewer's work is subject to a personal recall by the supervisors and a further 10 per cent is checked on by post.
Your writer also questions whether polling organisations should be governed by a pro- fessional institution. However, most of those people involved in running public opinion polls are members of the Market Research Society whose code of standards is equally applicable to political polls as it is to market research.
I am puzzled by Mr Paterson's illusions concerning our post-Budget survey. The date for this survey was selected early in April when we certainly could not have predicted a Budget with relatively little public impact. Finally, he complains of the disparity be- tween the results recently published by all the polls. In fact, however, all the polls have reflected a trend in favour of the Labour party and the polls estimate of the gap be- tween the two main parties lies in a range of about 6-1- per cent. This is within the range that you would expect to arise through the sampling error to which polls are subject.
Frank Teer Joint Managing Director, NOP Market Re- search Limited, 76/86 Strand, London wc2
Previous page
