" - WeetatOr October 12, 1974
tae s merely serve to obscure these.
c As presently organised the polling companies , arinot come up with these important sectional breakdowns. If the media are interested in Predicting the election outcome they should SPrelY be demanding this additional informat.lon, and be prepared to pay for the work Involved. It would appear that the media are more concerned with the reporting of global and sensational percentages rather than accurately informing their public. If the polls are inaccurate and their presentation of limited value, can they be disregarded as tantalising palliatives or are there more serious questions concerning the effects of the Polls on the electorate? The head of one polling organisation recently, stated that the polls might have cost the Conservative Party the last election. Some research was carried out before the last election by one of the Present authors which showed that opinion. Polls'may influence the turnout at the election. Voters were asked how likely they were actually to vote in different situations as depicted by the opinion polls. We found that a Percentage of Conservative supporters were leS5 motivated to vote when their party had a uig lead over Labour, and similarly some Labour supporters were less motivated when their party was ahead. Although the research Pointed to small changes in motivation to vote, these could be important when one considers the small margins on which elections are won. There is also the question as to whether the Polls bias each other. It was notable in the last :two elections that the majority of the polls were 'flaccurate in the same direction. This could be due to accurate polling followed by last-minute changes in voting intention, or to some developed nqrm of convergence within the PPIling organisations. How could the latter arise? One possible mechanism known to social Psychologists is the self-fulfilling prophecy and the effects of knowledge of results on subse;Tient data collection. The interviewers will "crw the state of the parties and this prior
knowledge may in some way bias the opinion polling, In a survey in France, 35 per cent of the public admitted that their attitudes were influenced by the opinion polls. If in this country people are similarly influenced and some electors quite legitimately use the polls as a guide to their voting, then inaccurate polls can hardly be taken as a good basis for such decisions. As predictors of the election outcome opinion polls are at best half-truths and at worst nothing more than misinformation — one wonders why the media give them such prominence. Just as the media impose restrictions upon themselves in the reporting of unfairly worded, or misleading questionnaires, they should also produce a code of ethics for the reporting of opinion polls.
A recent survey quoted in the Times showed that 58 per cent were against further nationalisation. The polling organisation compared their finding on that question with some previous research and showed a trend away from support for nationalisation. They did not however quote or examine a significant finding of the earlier research. Only 39 per cent were consistent in their response to the issue over a one-year period. The authors of that research concluded that "it seems more plausible to interpret the fluidity of the public's views as an indication of the limited degree to which attitudes are formed towards even the bestknown policy issues" (Butler and Stokes, Political Change in Britain, p.225). When claims to majority support are often used in argument for and against a particular policy it is essential to have an accurate assessment of the public sentiment on the issue. The recent advertising campaign against Labour's plans for extending public ownership is a disquieting example of the use of an opinion poll posing as objective social science evidence but whose accuracy is open to question for propaganda purposes. Why it is that, when a range of more sophisticated and potentially more illuminating
techniques have been developed for examining attitudes, the opinion pollsters rely on the responses to a few simple questions which possibly reflect the attitudes and values of the questioner more than those of the public? Andre' Malraux's comment of 1945 is still relevant: "Opinion polls are like medicine, less scientific than they claim, more scientffic
than any alternative." But are they scientific enough to be of any value? Informative social research is generally a time-consuming and costly business. Are we prepared to pay this cost?
George Gask`ell is a lecturer in Social psychology and Rod Bond a research associate at the London School of Economics and. Politic& Science.
Previous page
