Opinion polls

Making sense of the swings and roundabouts

Conrad Jameson

Interpreting opinion polls is the most gratifying of intellectual pastimes: everyone wins and no one loses, since every interpretation must succeed for lack of evidence to refute it. Consider the turnabouts in the opinion poll in recent weeks and how the press responded. With the announcement of the three-day week, the Tories spurted decisively ahead for the first time during their entire administration. As Tory managers rubbed their hands in delight, the polls showed Labour moving into a three-point lead. Hardly had this shock been suffered, than the polls did another volte-face on the eve of the election and put the Tories back in front.

. When the Tories first moved forward, the interpretation was simple: the people were supporting the Government, Phase 3 and a tough line with the miners. When Labour regained dominance, comments then scattered: some claimed that the prices and incomes policy had lost its support; others that the public had lost confidence, not in Stage 3, but in the Conservative leadership; still others argued that public sympathy had swung behind the miners; and the Sunday Times, in the gamiest interpretation of them all, ventured that the turnabout saw the birth of a new volatility which, as discovered by a poll of its own, shows nearly a third of the electorate changing its party allegiance practically every two weeks. By the time of the third turnabout, commentators had the jitters. Whatever result could be predicted, they now argued, it was that the result is unpredictable.

Which interpretation is correct? If all one had to go on were current pollster findings it would be impossible to say. Nor would there be much chance of finding out in subsequent surveys: newspapers cannot afford public opinion studies in depth; foundations and governments do not want to get involved in what might be considered partisan research; and political parties must pretend so hard not to be interested in survey findings that they end up with a lick and a promise sort of research that is as likely to add confusion as enlightenment. All the student of public opinion can do is play one-eyed king in claiming to look behind the most recent poll but one. Pollster findings have a history. That history can be revealing. The gains and losses in the Government's Popularity in the present crisis are probably less significant than they seem. All that has happened, at a guess, is that for several weeks the Government played its hand badly in announcing an emergency and then frittering away its advantage by giving an impression of dithering or weakness. The reaction of public opinion seems to have followed a pattern first studied in the 'thirties. President Roosevelt, finding his legislative programme baulked by the Supreme Court, proposed to appoint new justices and retire those over seventy. The polls showed that the country rose fervently in his support. Confident that public opinion was behind him, the President left the battle raging in Congress and the country. Public opinion then did a U-turn and rose with equal fervour behind the President's opposition.

That this kind of turn-around represents more than a willingness to be persuaded became clear in a'study I myself carried out in 1969 for the Observer on reactions to Labour's trade union reforms, ',In ,Place of Strife.' The study was what I called an "anti-poll" as its purpose was not to record opinion as such but to probe beneath the yeas and nays to see what they really mean. The findings agreed with those of conventional polls in showing that nearly three-quarters of the country supported the Bill. When responses were crossexamined, however, two things became clear.

One was that public opinion was at such sixes and sevens as to be shaky on its pins. Asked if the government should require a cooling-off period before a strike, a majority agreed. But asked if the government should let management and labour resolve their difficulties without legislation, most people agreed with this proposal as well. The same sort of contradictory response came out in questions about who was responsible for industrial disputes. Most people, for example, thought that strikes were justified because workers weren't getting their fair share; yet of those people who agreed, 70 per cent also thought that strikes were the work of agitators and communists.

A second finding was that the public was so nervous about the state of the country it was in a mood to support almost any measure that promised decisive government. The particulars of the Bill were all but irrelevant. Hardly half the country could identify a single item in the Bill and only 15 per cent could

name two of its measures. Even more revealingly, half of those who wanted VI outlaw wildcat strikes — roughly a third 0; the country — wanted strikes to be outlawo altogether. Take this finding and add it to others — a majority who agreed that the country was in a state of emergency, for example, or a majority who thought prices were rising faster than wages (untrue at the time) — and one can make at least an educated guess as to what was on the country's mind: it was not the Bill the country was supporting but a general mandate NT, action. The country wanted to hear the smao" of firm government — and never mind WI° was smacked or why. With the public at once anxious and confused, was it surprising that supporters of the Bill turned out to be sail; mertime soldiers who, when the governmen` stopped promoting the Bill and trade union leaders rose against it, turned away and fle.‘1? A government, it would seem, easilY misreads pollster findings. Once it discovers that a crisis lifts it on a crest of approval. it then assumes that support can be taken f°1 granted. Instead of beating the drum louder and louder, it lowers the volume or stogs beating the drum altogether. This is 3 mistake. The public finds it difficult t°, tolerate a slackening of ardour in governmea` leadership during a crisis — not least, 011e, suspects, because a lessening of the noise !eye' allows it to hear its own doubts about whir,,i1 side is right or what the issues are about. r' government that arouses the public in a war dance can at least hope to maintain allegiancet if it then immediately plunges into battle. Oil it is daring the odds if it starts a war dao.ce, and then either loses its own fervour or stt' down with the enemy for a friendly chat. Wis„e the Government might be in attempting .° fairly quiet period of negotiation, but there no doubt that it lost political advantage In, delaying an election. It has now cried 'wolf: cl second time and again the country has rani% to the call. The key question remains: will t.11;, Government continue to profit from a cris'ci psychology? I would say yes. But I w°11.„le have said so much more emphatically if P'r Government had called an election just afte it spurted ahead.

The vagaries of public opinion during a cris,1 must not be confused with the much important phenomena of political volatility. ',"s assess political volatility would require pane2, of voters who are re-interviewed over tn; years. Little data is available. All ot,r!t knows is that the signals of volatility t"'t have been flashing over recent weeks do n°t go back several months or even to the lass election, but were evident at least as earlY the mid-'sixties: high 'don't know' ac°resi particularly amongst the young, a pencha st for cross-voting across class lines and, tn° a.larmingly, a roller-coaster pattern of sweepihg crests and dips of approval and disapproval as if the whole country were suffering from fits of manic depression. The month after the 1966 election the figures for those saying they would vote Labour was up to 53 per cent, and approval for the GovernMent 69 per cent; by the spring of 1968, voting intentions for Labour were down to 28 per sent and approval to 27 per cent. (The implication is that roughly half of Labour's supporters were fickle turncoats.) The last two years have seen additional evidence of volatility in the willingness of voters to supPort third party candidates. Political volatility has been with us for years.' We are just beginning to notice it. What is happening now is probably more active in the minds of political commentators than it is in the electorate: a major mental adjustment to a set of facts long refused as unacceptable to the conventional wisdom. A generation of political studies on both sides of the Atlantic had taught that voting behaviour Shows surprisingly little influence of issues and personalities, but is determined by deepset Party loyalties of class, family and ethnic group. There would be enough shocks to the ship of state to explain an occasional rolling Motion between left and right. But the larger movements in which the ship actually changes direction should best be explained, it was argued, by slow-moving currents within society itself. At the end of the 'fifties, Mark Abrams came forward with a thesis, highly popular ,amongst Gaitskellites, that a left-wing Labour r.arty was doomed to electoral failure: with ,rising prosperity, he claimed, the working crass was now locked in an inexorable trend towards "embourgeoisement." Did not the better offworker tend to describe himself as Middle-class? And was not this tendency most Marked amongst the working-class young? The writing was on the wall: unless the 1-:ahour Party moved rightwards with the tit-nes it would, in increasing measure, lose its Working-class support. At the end of the 'six

S, David Butler and Donald Stokes came

'Pr.ward with a counter-thesis: with all good thin-s

s and time, there would eventually be

what the authors called a "natural" Labour rh,ajority. Was it not true that party allegiances follow an almost genetic pattern rross generations that political activities do lttle to disturb, and was not the age-structure r Britain changing in a direction that 'av.oured the Labour voter? Here was the Writing on the other side of the wall: let either party do what it please, the success of the abour Party would still be assured. nut probably the most widely accepted analysis was carried out by two Cambridge seciologists, John Goldthorpe and David te'ckwood, who argued in favour of neither a rend to the left nor a trend to the right, but irliPly a trend down the middle: the ,10.rking-class voter would stay put — and 'us despite the pressures of high pay and Prosperity. These sociologists had studied the work and lt3Pl1tical habits of well-off car workers in Luacm; They found that affluence was breeding Iii alright, Jack" attitude towards work ulch, in its concentration upon pay, was Vnerating a decreased interest in collective wer,t12n. But only those who left their s;ricing-class milieu showed a tendency to _elect from the Labour Party. As long as these workers remained a homogeneous social fr,r°uP, so long would their traditional Labour riYalties remain. The implications of these iorldings were not lost on the Left: class Yalties and consciousness were not, as trintonly believed, eroded by prosperity and, 6°„1-igh interest in collective action might tli-je Paled, the Labour Party could still venit 'e left-wing policies without fear of losing s vvorking-class support. Shortly after their two monographs, The Affluent Worker, were published in 1968, the anti-collective workers in Luton staged a sitin. Then came the general election. And which was the key social group showing the most radical swing from Labour to Tory? The well-off workers of the same type who were supposed to be most steadfast in their allegiance to Labour.

The fact of political volatility would now seem to be accepted by political analysts — and accept it they must, as the evidence is compelling. Yet the same resistance that until recently showed in a begrudging and belated recognition of the growing instability of political allegiances now shows in sanguine speculations about what this turn of events might mean. The electorate is more volatile, it is said, only because it is disillusioned with political parties or the political process: let the Political performance improve and voters will return to the fold. This is a reasonable argument — reasonable, that is, were it not for the fact that if disillusionment were the major factor one would expect large-scale abstentions rather than the very slight decline in turnout since the war. It is also doubtful whether disillusionment would take the primary form of an erratic shuffling between one party and another. The electorate is volatile, it is said, only because it is now bet`,‘r educated and informed: volatility is only another name for a more critical political stance. This argument is also reasonable — reasonable, that is, were it not for the fact that stable voters tend to be found among both the most and the least informed: the changelings are those whose party ties are easily broken by small inputs of information. Until recently it was thought that voters selected political messages from the mass media according to their political prejudices. A newer crop of studies shows, however, that the mass media do not, as originally conjectured, convince only the converted; nor, for that matter, do the mass media persuade. What the mass media do is confuse: the viewer or reader swings back and forth with a mind so open, as Walter Lippmann once said, that there is nothing left but a draught. Interestingly, these studies also show that those least involved in politics are the most easily swayed.

Note that neither of the two main lines of speculation takes into account that volatility is observable outside the political sphere. Consumer behaviour has also grown more volatile during roughly the same period, a fact which explains why this period has seen an outbreak of frenetic product launches and re-launches as manufacturers respond to a more promiscuous pattern of shopping; one study of the food market, for example, showed that even when a brand share remained fairly stable about a third of its buyers entered or left nearly every six weeks. Whatever the social causes, political volatility cannot be simply put down to changes in politics itself. One must look to social changes of a more general kind.

What, then, is causing the loosening of political ties? No one knows. There are a number of theories knocking around and for every four analysts there are five opinions. Some argue that society has become atomised in much the way nineteenth-century sociologists predicted. At the other extreme are those who, while they accept a break-up of the extended family system, see new ties being reknitted: the newcomer to the housing estate may be temporily disorientated in his beliefs but he soon creates a new set of ties that are as abundant or nearly as abundant as the old ones. he has cast aside.

With so many theories to choose from let me only pick my favourite. It was developed by the brilliant philosopher-sociologist, W. G. Runciman, not as an explanation of political volatility as such, but as an explanation of why wage demands come forward or recede in ways that seem to have so little to do with actual deprivation. Runciman built on a famous wartime study of morale among military personnel. One might expect morale to be highest amongst those with the best opportunities for promotion. Exactly the opposite proved to be the case. The conjectured explanation: promotions created the jealous pangs of what was called "relative deprivation." Using this hypothesis as his guide, Runciman traced through the history of political and industrial agitation since World War I. He found that unrest coincided with just those conditions which detonate feelings of relative deprivation: a social group meets a new set of Joneses and starts making a whole new set of comparisons about what it should and should not expect: or, alternatively, the Joneses already known suddenly change their status. Both conditions have prevailed during the period in which volatility has become most marked, with social mobility and inflation mixing in what could prove to be a deadly combination. If relative deprivation can fire off angry social demands, could it not, in acting upon political ties already loosened by social change, also shake the unconvinced voter into political volatility?

The implications could be grim. For one, the theory suggests that the Lib-Lab argument that all would be agreed about a prices and incomes policy if the distribution of income were more fair is more than probably mistaken. But the other and even more disturbing implication of Runciman's findings is that, as social mobility and inflation are here to stay, so is political volatility. The danger is not the obvious one: that voters so easily persuaded to change their allegiance might eventually be persuaded to accept an allegiance that takes away their right to change it once again. The danger is more indirect though hardly les.s forbidding: that those so easily persuaded fail to provide sufficient consent for lin" popular but important politicial decisions. It !s the breakdown of rule by consent that !s threatened by volatility — and threatened in a more frightening way than we might care to acknowledge.

Conrad Jameson is the director of a social and market research organisation