Polling in UK elections since 1945

Prepared by AMSR using the Archive

Voting behaviour and the media: Polling in UK elections since 1945

Introduction

This document is a list/summary of what the Archive includes on each UK General election on the influence of the media: for example data on how many people have seen Polls and know what they said.

Summary: Polling, Elections and the Media

A-Level Politics students must understand how the media influences voting behaviour, especially through its use of opinion polls. This includes the role of the media during General Elections, including media bias, persuasion, and the importance of opinion polls. The AMSR Archive provides a unique historical record of how polls have been used, reported, believed, doubted, and sometimes misunderstood — and how this has shaped public expectations and political behaviour from 1945 to today.

Across the decades, several clear patterns emerge. In the 1940s and 1950s, polling was new and often distrusted, but newspapers still published results and sometimes miscalled elections. By the 1960s, polls had become a normal part of election coverage, with television increasing public exposure and helping to create a sense of national momentum. In the 1970s, media reporting of polls exploded: more people saw polls, newspapers often misreported them, and the famous 1970 failure became a national story. The 1980s saw the rise of telephone polling and a dramatic increase in the number of polls published, with voters becoming more aware of polls and more willing to discuss their influence. By the 1990s and 2000s, polls were central to campaign narratives, and failures such as 1992 became major media events. In the 2010s, online polling and social media transformed how polls were shared, debated, and challenged, with 2015 and 2017 showing how volatile modern opinion can be.

Taken together, the Archive shows that the media’s influence is not simply about reporting numbers. It is about how polls shape campaign tone, public expectations, party morale, and voter perceptions of momentum. Polls can stir interest, create pressure, or even encourage strategic voting — and when they go wrong, the media amplifies the shock.

The Debate About Banning Opinion Polls Before Elections

Every few years in the UK, there is a debate about whether opinion polls should be banned in the final week before a General Election. The argument usually appears after a polling failure — for example in 1970, 1992, or 2015 — when critics say that inaccurate polls can mislead voters, distort media coverage, or create false expectations about who is likely to win. Some politicians claim that polls can influence voters by creating a “bandwagon effect” (supporting whoever seems ahead) or a “boomerang effect” (supporting the underdog). Supporters of a ban point out that several countries already restrict polls, including France, Italy, Canada, and India, where publication stops a few days before voting. However, UK pollsters and most media organisations argue strongly against bans. They say that polls are part of a free press, that voters should have access to information, and that banning polls would simply push them onto social media or foreign websites. The Archive shows that Prime Ministers from Harold Wilson onwards have rejected bans, calling them impractical and undesirable in a modern democracy. So although the debate keeps resurfacing, the UK has never restricted opinion polls — and continues to publish them right up to polling day.

Other Areas Students Could Research (Beyond the AMSR Archive)

Although the AMSR Archive focuses mainly on traditional media and polling, there are several wider issues that students might explore if they want to understand the full picture of how the media influences elections. These include the impact of social media, which now spreads poll results instantly and can amplify misleading interpretations; the debate about bandwagon and underdog effects, where polls might encourage voters to support the party that seems to be winning or, alternatively, the underdog; the role of polls in tactical voting, especially in multi-party contests; and concerns about horse-race journalism, where media coverage focuses more on who is ahead than on policy. These topics are not covered extensively in the Archive, but they are important parts of the modern debate and offer useful avenues for further research. Students who want to explore wider issues can look at modern sources like Ofcom, the Electoral Commission, or the Reuters Institute for the Study of Journalism.

Polling in UK Elections Since 1945

Polling in the 1940’s

1945 UK Election:

Political polling was in its infancy with British Gallup, originally known as the British Institute of Public Opinion (BIPO), and the first organisation of its kind, founded in 1937^1,2
Most questions were of ‘Yes’ or ‘No’ variety^1,2
First voting intention question was introduced in February 1939 when 54% answered they would vote for the Government ‘if there were a General Election tomorrow’, 30% for the opposition and 16% did not know²
In 1945, Gallup predicted a Labour victory under the leadership of Clement Attlee^2,3
When the General Election came, in 1945, Gallup showed that Attlee and the Labour party were going to win but no one believed this prediction, including News Chronicle who published the poll^2,3
However, as the very first results came in it was clear that Labour had won and Gallup were vindicated⁴
Nevertheless, polls were disliked at the time and market research itself not much loved⁴
This startling confirmation that sample surveys are capable of accurate work helped everyone⁴
Gallup’s reputation was secure and they have polled at every single General Election since then³

References:

Political Opinion Polling in Great Britain: Past, Present and Future. Robert M Worcester: Political polling in Great Britain: past, present and future – Opinion Poll Reports – The AMSR Online Archive (oclc.org)
Pollsters, the Press, and Political Polling in Britain. Robert M Worcester: https://amsr.contentdm.oclc.org/digital/collection/PapersOffprnt/id/1750/rec/37
Why a Sample of the Polls Suggests a Vote in Favour. Nick Moon: Research Plus issue 360 supp (50 years of market research) 1996 May – Research Magazine (1992-2012) – The AMSR Online Archive (oclc.org)
The Making of an Industry. The Market Research Society 1946-1986. A History of Growing Achievement. Ian Blythe: https://amsr.contentdm.oclc.org/digital/collection/Res-history/id/594/rec/3

Polling in the 1950’s

In the 1950’s, polls were all conducted using random samples¹
During the 1950’s, politicians began to take polls seriously²
Gallup achieved respect but the polling business as a whole did not¹
Morgan Phillips, General Secretary of the Labour party, warned against ‘a new technique of propaganda ….. the publication of polls’ (quoted in Butler, 1952)^2,3

1950 UK Election:

Newspaper polls at the General Election of May 1950 were close, even if 2 of the 3 called the wrong winner^2,3
The Daily Express entered the polling field in 1950 and, in 2 of the 3 General Elections won by Labour (1950, 1964 and 1966), they had the misfortune to put the losers, the Conservatives, ahead in the final figures^4,5
Three other polls appeared during the 1950’s: The Daily Express conducted its own; in 1951, Research Services Limited (RSL) published its first polls, and in 1957, RMP Shields set up National Opinion Polls (NOP)^2,3

1951 UK Election:

The 1951 UK General Election was called by the Labour government in the hope of increasing their parliamentary majority. In the 1951 General Election, all 4 polls wrongly forecast that the Conservatives would win most votes⁶
In fact Labour won most votes, but the Tories won more seats and so formed the next government⁶
The average error in measuring the party shares was only 2.2 percentage points⁶

1955 UK Election:

In 1955, 4% voted other than for the 2 main parties; in 1959, it was 7%⁷
Before 1955, there was no coverage of Elections on radio or television⁸

1959 UK Election:

The Daily Express had conducted perfectly reasonable polls of its own but announced in 1959 that ‘The Daily Express has no confidence in its own poll although it is conducted with complete integrity and all possible efficiency’ (Teer and Spence, 1973) which fuelled the doubts about polls^2,3
Election Day in 1959 was October 8, and from September 21 onwards polling results in sponsoring papers began to encourage the Labour party and shake up the Tories. A Tory lead of 6.8% slid down to around 2%, and 3 days before polling, Gallup reported a neck and neck race²
Polls in earlier elections had tended to underestimate Labour’s strength. Although the popular belief persisted that Conservatives would win, the City was more attuned to the polls’ reports resulting in a fall in stock prices on those mornings when opinion polls showed a decline in Tory lead²
In 1955 and 1959, the average error in measuring party shares was less than 1 percentage point⁶
According to Butler and Rose (1960), pre-1959, polls had an appreciable effect on the morale and expectations of politicians, but little impact on the electorate. However, in 1959, there were signs of a wider impact since ‘the effect of the polls, like that of television, seems to be to stir up interest, and hence participation in the election’^2,3

References:

Why a Sample of the Polls Suggests a Vote in Favour. Nick Moon: Research Plus issue 360 supp (50 years of market research) 1996 May – Research Magazine (1992-2012) – The AMSR Online Archive (oclc.org)
Pollsters, the Press, and Political Polling in Britain. Robert M Worcester: https://amsr.contentdm.oclc.org/digital/collection/PapersOffprnt/id/1750/rec/37
Political Opinion Polling in Great Britain: Past, Present and Future. Robert M Worcester: Political polling in Great Britain: past, present and future – Opinion Poll Reports – The AMSR Online Archive (oclc.org)
The Gallup Election Handbook: a compilation of statistics covering British General Elections and by-elections 1945-1970, obtained from Gallup Polls and other sources, and a description of the Gallup Poll, its organisation, its record and the methods employed: https://amsr.contentdm.oclc.org/digital/collection/p21050coll1/id/2746/rec/13
Journal of the Market Research Society vol 16 no. 4 October: https://amsr.contentdm.oclc.org/digital/collection/p21050coll2/id/3746/rec/14
Polling in Britain: A Historic Route to the Forecasting Future. Professor Robert M Worcester, MORI, traces the origins and track record of pollsters, politics and the media: https://amsr.contentdm.oclc.org/digital/collection/ResearchMag/id/4133/rec/1
The Volatility of Voting Behaviour. John Clemens: Journal of the Market Research Society vol 16 no. 4 October – The MRS Journal – The AMSR Online Archive (oclc.org)
Letter to our Readers. Robert M Worcester. British Public Opinion Issue 19.02 1996 March: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/652/rec/1

Polling in the 1960’s

During the 1960’s polls in Britain began to come of age, with several new organisations entering the field (eg Opinion Research Centre [ORC], Marplan), regular private polls were commissioned by the political parties, pressure groups used ‘sensible’ polling, and newspapers and television began to take polls more seriously^1,2
There were 2 General Elections in the 1960’s: 1964 and 1966 and during this period, continuous polling was conducted by Gallup and NOP, but in 1964, the Daily Express Poll and RSL also participated^1,2

1964 UK Election:

The 1964 election was close, with Labour winning by 1.9% of the vote; 4 out of 5 published polls correctly called Labour the winner in total votes cast^1,2
In 1964, 12.5% voted other than for the 2 main parties, and this Election marked the first big step in a Liberal revival³

1966 UK Election:

By the 1966 election, published polls appeared in the Daily Telegraph, Daily Mail, Daily Express, Observer, the Guardian, Financial Times, the Economist and the Sunday Times. The 2 most highly regarded polls agreed in showing Labour with a 10% or more lead throughout the campaign and this took the steam out of the contest^1,2
This led to calls to ban polls but Mr Wilson as Prime Minister flatly refused to consider any restriction on election polls (quoted in Butler and King, 1966)^1,2
In 1966, 10% voted other than for the 2 main parties³
Between 1966 and 1970, public opinion, previously extraordinarily stable, became more volatile. In the May 1966 General Election, Labour enjoyed a 20% lead but just 2 years later was 20% behind, with a long-term trend towards the Tories, despite dissatisfaction with Mr Heath as the Conservative leader^1,4

References:

Pollsters, the Press, and Political Polling in Britain. Robert M Worcester: https://amsr.contentdm.oclc.org/digital/collection/PapersOffprnt/id/1750/rec/37
Political Opinion Polling in Great Britain: Past, Present and Future. Robert M Worcester: Political polling in Great Britain: past, present and future – Opinion Poll Reports – The AMSR Online Archive (oclc.org)
The Volatility of Voting Behaviour. John Clemens: Journal of the Market Research Society vol 16 no. 4 October – The MRS Journal – The AMSR Online Archive (oclc.org)
NOP Bulletin 1967 August: https://amsr.contentdm.oclc.org/digital/collection/NOP_Reports/id/1572/rec/1

Polling in the 1970’s

By the early 1970’s, there was a change from using random samples to quota sampling and by the mid-1970’s, quota had almost entirely replaced random sampling as a method of conducting opinion polls^1,2
Although this represented a major change in technique, in most other respects there was little change between the methods used in 1974 General Election and that of 1992¹

1970 UK Election:

During the 1970 election campaign, there were 5 polls being published but they were read by an enormously enhanced audience as exclusive publication was abandoned. Prior to this, newspapers had jealously guarded the copyright they had on their polls^2,3
Polls dominated the 1970 election campaign reporting and, during the following September, The Harris Poll found 62% of the country remembered seeing polls (71% recalled seeing them on television)³
As a result, in 1970 when the polls got it so badly wrong, they were very much in the public eye; the Conservatives were the surprise winners of the ‘sunshine’ election of 1970^3,4
In May 1970, when the election was called, Labour was ahead of the Tories. However, Press reports during the interim period provided misleading data by rarely giving interviewing dates, hardly ever reported ‘Don’t Knows’ or ‘Refusals’, seldom offered breakdowns of top line figures, and showed little interest in issues. Sampling error was seldom mentioned, and headlines reported shifts of electoral support based on statistically insignificant changes³
At the beginning of the 1970 campaign, led by Humphrey Taylor of ORC, the pollsters banded together to publish a Code of Practice to curb this misreporting of data by the newspapers³
All the polls got the winning party wrong in straight poll figures, with all predictions being too Labour, but ORC adjusted its final raw figures by projected turnout to correctly predict a Tory victory^3,4
In 1970, 3 polls were conducted with random samples, 2 by quota.The mean error of the former on the Conservative party lead was 7.6%; by quota samples it was 5.4%³
Of the 5 final forecasts, only 1 (ORC for The Standard) correctly forecast the Conservative victory: the average share error was 2.2 points, the same as in 1951, but the average error on the Conservative lead over Labour was as high as 6.6 points²
Reasons why the polls got it wrong in 1970: The method of sampling did not appear to be the answer and there was only suggestive rather than conclusive evidence for a late shift. There were many hypotheses (eg differential voting by non-contacts, organized deception by Tory supporters, an anti-poll bandwagon, collusion on the part of pollsters, selective turnout etc), but these were mostly inconclusive, post-election analyses tend to support the late swing hypothesis (Abrams, 1970; Butler and Pinto-Duschinsky, 1971; Hodder-Williams, 1970; Market Research Society, 1972; Rose, 1970; Teer and Spence, 1973)^3,4
In a post-election survey of the 1970 Election, the Tories made considerable gains from those who had voted Labour in 1966, Labour lost more of their 1966 strength through abstention than did the Tories and, although younger voters are more likely to support Labour theoretically, there was substantial abstention in this group⁵
The Tories won support on their promise of beating the Cost of Living Situation and a pre-election NOP survey shows the problem retained its importance throughout the campaign. The majority who thought the Conservatives were the best party to deal with this problem doubled during the 3 weeks leading up to June 18⁵
The Tories seemed to have fought a substantially better campaign than Labour, particularly on television⁵

1974 UK Election (February):

Despite the 1970 fiasco, polls were as much quoted in the Press in the February 1974 General Election as they had been in 1970, perhaps even more so on television³
Both major parties mounted substantial private polling programs and 6 polling organizations conducted major national polls for the Press^3,4
The polls presented a confused picture: 5 polls showing a Tory lead ranging from 1.5-11% and every poll showed a trend to the Liberals³
On Election Day, the published polls were close but every prediction in February 1974 was too Conservative. The Conservatives won the popular vote by 0.8%, but Labour, with more MPs elected won the February 1974 General Election^3,4
In the first 1974 election, the difficulty of translating votes measured by the polls, into seats in the House of Commons was crucial. The Conservatives won most votes but it was Labour who won the most seats³
Every single prediction in June 1970 had been too Labour, every single prediction in February 1974 had been too Conservative^2,3
Harris and NOP were the 2 polling firms employing the largest sample sizes, random sampling and re-interviewing but were the furthest adrift. All the polls overestimated the Liberals, who did not contest 106 seats, and the pollsters failed to take this into account. All the polls underestimated Labour. All but Marplan adjusted for differential turnout and this increased the Tory lead erroneously (Worcester, 1975)³
An NOP opinion poll in March 1974 showed that most people made up their mind which party to vote for before the election campaign began, however 24% of all respondents made up their mind during the campaign itself⁶
This figure is high, ie 1 in 4, due to deliberate oversampling of Liberal voters who tend to be late deciders, but if Liberal voters are excluded, this falls from 24% to 18% which is still high (1 in 5)⁶
Nearly 1 in 5 Liberal voters did not decide to vote Liberal until the last few days of the campaign⁶
The polls over-estimate of the Liberal vote and under-estimate of the Labour vote at the General Election may have been the result of late changes from a Liberal intention to a Labour vote⁶
This was an election marked by change in that in 1974, 3 million more voters voted, and 4 million more voted Liberal compared to 1970⁷
Another source of inaccuracy in the pre-election polls was the adjustment for differential turnout. Following the post-mortem on the 1970 election, it was decided that election prediction would benefit from adjustments for differential turnout. In 1970, there was a strong relationship between stated intention to vote and actual turnout. This was not the case at the 1974 General Election where many respondents who expressed a very low likelihood to vote before the election actually did turn out to cast their vote⁶
Furthermore, most experimentation on weighting for differential turnout indicated a net advantage of 1% to the Conservatives so, learning from the 1970 Election, the majority of opinion polls, including NOP, adjusted their final figures to take this into account. However, the established relationship between turnout and Conservative party advantage did not repeat itself in 1974 and these adjustments made the poll forecasts less accurate⁶
The 1974 General Election saw more published opinion polls than any previous British Election. However, there was no significant increase in the penetration of opinion polls compared with 1970. 58% claimed to have seen poll results during the 1974 campaign. During the 1970 campaign it was 56%, There was no indication that respondents who made up their mind about their vote after the half way stage in the campaign were any different in their exposure to opinion polls than those who made their mind up early⁶
Most people who had seen poll findings remembered that they showed the Conservative party ahead (58%). However, almost half the electorate (48%) had no idea what the polls were showing⁶
In 1970, 11% voted other than for the 2 main parties; in February 1974, it was 25%. This trend has increased since 1955 showing the difficulties in forecasting⁷

1974 UK Election (October):

In October 1974, Labour Prime Minister Harold Wilson was forced to call a second General Election in a bid to turn his minority in Parliament into a solid working majority. During the 3 week election, the published polls carried out some 40,000 interviews, with well over 10,000 more conducted privately²
Every poll, apart from one (NOP, 16 September), showed Labour comfortably ahead. The results were more clustered than previously, of the 25 polls published, 20 appeared on a Thursday or a Sunday, none on a Monday or Tuesday, and only one on a Friday or Saturday²
The final polls all picked the winning party, but this time results were too pro-Labour and pro-Government²
The October 1974 election was the 4^th in a row in which every single poll overestimated the support for the party that was in the lead²
The October 1974 election demonstrated that the later polls are conducted in the campaign, the more likely they are to be accurate indications of the actual electoral outcome⁸
The NOP poll conducted between 2-5 October 1974 indicated little change in the state of the parties during the campaign so far, the Labour lead was still about 8% but there was still a high proportion of undecided (16%) who, upon further questioning, seemed slightly more in favour of the Conservatives than Labour⁸
Electoral volatility seems to have increased during the 1970’s, with NOP finding that 28% of all respondents sometimes change their party allegiance between Elections⁸
However, the actual Election result showed the Conservative party did surprisingly well, taking voters from the Liberals in key marginal seats⁸

1979 UK Election:

The May 1979 election was long expected and forced on a reluctant Prime Minister by Parliament. It came at a time when there were other important factors: Britain’s first national referendum on the EEC, a devolution referendum in Scotland, the resignation of Harold Wilson and election of Mr Callaghan as leader of the Labour party and thereby Prime Minister, and the ‘accidental’ election of Mrs Thatcher to replace Edward Heath as Tory leader^2,3
There were more published polls than ever before because the campaign was very long. Some 50,000 electors (registered voters) were asked their voting intention as well as other questions²
The 1979 election saw the first use of telephone surveys in election polls in Britain by Marplan^2,3
The Conservatives led by a substantial margin from the outset. Over the first 3 weeks of the campaign, the Tories were at 48% +/-2%, and Labour at 40% +/-2%. At week 4, the public narrowed the gap after sensing a handsome Tory victory. However, on election day, the polls proved to be very close to the actual result, with 3 of the 4 leading firms achieving results within 1% of the actual Conservative lead on the day^2,3
Overall, at the 3 elections before 1979, the polls are frequently considered to have done badly. In October 1974, polls correctly predicted the winner but over-estimated the size of Labour lead. In February 1974, polls were actually more accurate that given credit for, 5 of 6 main companies who polled in week before polling day showed Conservative lead between 1.5-5%, but because Labour won the election as they won more seats, the polls were heavily criticised. In fact, the Conservatives won more of the popular vote: 38.2% to 37.2% and all but one of the polls were correct in predicting this. For the 6 major polls, the average error on size of Tory lead in the popular vote was very respectable at 2.3%⁹
In 1970, all of the polls except ORC predicted a Labour majority of the popular vote but in fact there was a Tory majority which can be explained by a late swing to the Tories and this showed the necessity of conducting polls as close to polling day as possible to detect any late switches. This practice is also helpful since the electorate has become more volatile so polls closer to Election Day are more likely to account for this⁹
In 1979, polling companies polled right up to the day before polling day and were remarkably accurate: average error in the lead was 1.72% and average error on the parties was 0.84%⁹

References:

Why a Sample of the Polls Suggests a Vote in Favour. Nick Moon: Research Plus issue 360 supp (50 years of market research) 1996 May – Research Magazine (1992-2012) – The AMSR Online Archive (oclc.org)
Polling in Britain: A Historic Route to the Forecasting Future. Professor Robert M Worcester, MORI, traces the origins and track record of pollsters, politics and the media: https://amsr.contentdm.oclc.org/digital/collection/ResearchMag/id/4133/rec/1
Pollsters, the Press, and Political Polling in Britain. Robert M Worcester: https://amsr.contentdm.oclc.org/digital/collection/PapersOffprnt/id/1750/rec/37
Political Opinion Polling in Great Britain: Past, Present and Future. Robert M Worcester: Political polling in Great Britain: past, present and future – Opinion Poll Reports – The AMSR Online Archive (oclc.org)
OPINION POLLS AND THE 1970 GENERAL ELECTION. Frsnk Teer: NOP Bulletin – Opinion Polls and the 1970 General Election and Post Election Survey 1970, July (Link missing – sp. Frank??)
NOP Political Bulletin 1974 March/April: NOP Political Bulletin 1974 March/April – NOP Reports – The AMSR Online Archive (oclc.org)
The Volatility of Voting Behaviour. John Clemens: Journal of the Market Research Society vol 16 no. 4 October – The MRS Journal – The AMSR Online Archive (oclc.org)
NOP Political Bulletin 1974 October: https://amsr.contentdm.oclc.org/digital/collection/NOP_Reports/id/3744/rec/1
Political, Social, Economic Review NOP. The Accuracy of the Polls: https://amsr.contentdm.oclc.org/digital/collection/NOP_Reports/id/666/rec/1

Polling in the 1980’s

1983 UK Election:

In the 1983 General Election, there were 43 national polls, compared with 26 in 1979. There were also 5 waves of panel studies, 10 surveys of marginal, 3 surveys of SDP seats and 15-20 ‘respectable’ polls done in individual constituencies, making some 75-80 public polls in all. There were also about 35-40 private polls¹
A total of some 100,000 people were interviewed by pollsters during campaign out of an electorate of some 42 million¹
In 1983, most of the polls over-estimated the Tories’ share, and most under-estimated Labour’s, all, save Audience Selection, were exactly correct for the Alliance share^1,2
Audience Selection, polling by telephone, was the first to conduct a national telephone poll as an eve of poll prediction and their wide margin of error threw a shadow on the use of such polls in the period that followed²
Audience Selection overestimated the Alliance share by 3% due to problems with the use of telephone sampling, although in 1987, they performed much more in line with the face to face polls^1,2,3
Rumours from polls leading to massive impact on financial markets did not happen in 1983 because of the wide lead held by Tories so a Labour victory seemed incredible²
The 1983 General Election provided the Conservatives with the biggest share of the Con/Lab vote (60.6%) since 1931, but the smallest share of the total popular vote ever achieved by any party winning a working majority in the House of Commons (43.5%, with 397 seats) compared to Labour (28.3% with 209 seats)⁴
The 1983 election was not just about how the Conservatives won, but how Labour lost⁴
MORI interviewed 15,930 people for its nationwide polls. A total of 13,926 gave a voting intention⁴
Three groups of voters showed a swing towards the Conservatives. Among skilled workers and voters aged 18-24, huge Labour leads were converted into clear Conservative leads, among trade union members non-Labour trade union voters outnumbered Labour trade union voters 3 to 2⁴
Responses to polls conducted for the Daily Express on May 19 and June 2 showed most people thought there was too much coverage given to the General Election on television (47%, 52%, respectively), and about right by newspapers (47%, 50%, respectively), and the majority of respondents in the June 2 polls stated that they did not agree with a ban on all Election coverage on TV and Radio (84%), a ban on all party political broadcasts on TV and Radio (82%), a ban on newspaper coverage (87%) or a ban on the publication of opinion polls (71%)⁴
On Election Day, support for both Labour and Conservatives was at or very close to their lowest points of the campaign, and Alliance gained ground while Labour and Conservative support fell. However, the evenness of the Alliance vote meant the shift in votes had little impact in terms of seats, so the Conservatives entered the campaign with a massive lead over Labour and held that lead throughout the campaign⁴
The Tories won the big issue arguments (inflation, unemployment, NHS, nuclear arms), had more popular policies, and had the most popular leader, with Mrs Thatcher being seen as the best candidate for Prime Minister by a large margin⁴
Despite record unemployment, all the main political indicators ie leadership, policies, rating on the main issues, either looked favourable to the Conservatives throughout the campaign or moved the party’s way during it⁴
On June 9, Mrs Thatcher became the first Prime Minister that century to win 2 successive General Elections with working majorities⁴
Despite a proliferation of political opinion polls carried out during the run up to the 1983 General Election, NOP’s survey between 16-21 June 1983 of 1,823 adults aged 18 and over, found over a third of people claimed to have been very or fairly interested in them⁵
Awareness of the other pollsters was high: Gallup 64%, MORI 52%, and Marplan 38%, with men and those in the ABC1 social group having highest awareness levels⁵
The 1983 campaign seemed characterised by incessant polling, bombarding the public daily in papers and on TV. Some politicians deplored this, seeing it as a burden when the focus should have been on issues and problems of the country¹
Some, mostly from left of Labour party, called for a ban on polls during General Elections. However, although there would be clear economic advantages to banning polls the week before an Election, the conclusion was that the public should have the freedom to learn, or choose to ignore, the state of public opinion⁶
In fact analysis of coverage of polls on front of daily newspapers found just under 5% of front page space had been allocated to opinion polls¹
Moreover, with all the talk of intrusion of polls during campaign on the eve of polling day, Gallup asked ‘Have you seen or heard the results of an opinion poll in the last few days? One third, 33%, of the public said they had not seen or heard any poll. MORI on 16-20 June, asked ‘Have you seen the results of any opinion polls recently showing how people said they would vote in the General Election?’ and found by a week after polling day that 37% of the British public couldn’t recall having seen the results of any poll!¹
The 18-44 year age group were inclined to believe that polls could be instrumental by influencing peoples’ voting behaviour, while Conservatives were least likely to believe in the influence of polls; SDP/Lib voters were the group most likely to believe in their influence⁵
Of those who did claim polls could have an effect, less than 1 in 10 said it had helped or influenced them personally. The groups most likely to admit this were 18-24’s, DE’s⁵
In terms of accuracy, most people (63%) thought the polls would be very or fairly accurate, this being especially true amongst the 35-44 age group, AB’s and Conservative voters⁵

1987 UK Election:

At the time of the General Election of 1987, telephone penetration in Britain was only around 80% so many were sceptical of its value, especially for voting intention measurement²
There was considerable debate on the use of the telephone as a viable method for political polling. There was concern surrounding the inability of telephone polling to counteract the effect of those people who are not able to be contacted by telephone, with an acceptance that a significant weighting procedure had to be used. However, it was acknowledged that substantial benefits could be afforded through telephone polling, there was no justification in excluding this as a respected polling technique and that the polling industry had a duty to continue to develop and improve techniques that increased accuracy, reliability and precision of information finding⁷
In 1987, there were a number of occasions when the City of London was swept by rumours that one or other major poll was coming out showing a sharp swing to Labour, occasionally on the basis of false or non-existent findings. There was even a student-conducted poll for BBC’s ‘Newsnight’ that did leak accurately early enough in the day to affect market trends²
During the month-long campaign, some 54 national opinion polls were reported, plus 7 in Scotland, 1 in Wales, 4 national telephone polls, 12 or so regional polls and several score individual constituency polls published, plus at least 40-50 private polls conducted for the political parties and in some cases leaked to the Press²
At the outset of the campaign, the Conservatives had a 42% lead over Labour’s 32% and Alliance’s 24%. On Election Day, Conservatives were up a point, Labour up 2, and Alliance down 3^2,8
During the campaign, the polls were very stable, with Conservatives at 43% +/-3%, Labour at 34% +/-3% and Alliance at 22% +/-3%^2,8
However, the MORI/Sunday Times panel survey found that between the time the election was called and weekend before polling day, some 7.5 million people had changed their minds either between parties or on whether or not to vote during the 23-day election campaign, out of an electorate of some 43 million electors^2,8
The actual result of the 1987 General Election on 11 June was Conservatives 43%, Labour 32%, Alliance 23%^2,8
Thus, while the 1987 General Election saw the most professional and well-presented Labour campaign in the party’s history, the final result saw the third successive Election of a Conservative Prime Minister⁹
Overall, in 1987, all the major opinion polling companies did a creditable job; 6 out of 6 had the share of each party to within the +/- 3% margin of error generally accepted, just as they had in 1983, with the exception of Audience Selection by telephone, and in 1979²
MORI reported its accuracy in the 1987 Election with the average error per party (share] being +/-1.5% and in 1987, as in 1983, all the major pollsters were within +/-3% of the outcome⁹
Turnout was rose from 1983 to 75.3%, but there was no evidence to suggest that turnout helped any single party’s performance⁸
In the 1987 Election, the majority of the public expected ‘too much’ coverage of the election on television – and got it – with the undecided voters most likely to say there is too much coverage already⁸
Most people say they rely on TV for their election news and most (70%) believe all parties are treated fairly⁸
There was little difference in the extent to which supporters of the different parties were touched by the election campaign via leaflets, party election broadcasts, canvassing etc⁸

References:

Mrs Thatcher Won Again. So did the Pollsters – Didn’t they? Did they? Robert M Worcester: https://amsr.contentdm.oclc.org/digital/collection/Survey-Mag/id/540/rec/1
Political Opinion Polling in Great Britain: Past, Present and Future. Robert M Worcester: Political polling in Great Britain: past, present and future – Opinion Poll Reports – The AMSR Online Archive (oclc.org)
Why a Sample of the Polls Suggests a Vote in Favour. Nick Moon: Research Plus issue 360 supp (50 years of market research) 1996 May – Research Magazine (1992-2012) – The AMSR Online Archive (oclc.org)
Review of the Month. Tories Win More Seats But Fewer Votes. British Public Opinion Issue 05.05/6 1983 May/June: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/1506/rec/1
Opinion Polls. NOP Political, Social, Economic Review Issue no.43 1983 September: https://amsr.contentdm.oclc.org/digital/collection/NOP_Reports/id/814/rec/1
Political Opinion Polling Robert M Worcester MORI: The Market Research Society Newsletter Issue 250 1987 January – MRS Newsletter (1966-1992) – The AMSR Online Archive (oclc.org)
Telephone opinion polls. The Market Research Society Newsletter Issue 256 1987 July. Eamonn Santry, Derek Farr and Peter Hyett: https://amsr.contentdm.oclc.org/digital/collection/MRSNewsletter/id/2780/rec/1
Review of The British General Election of 1987. Edited by David Butler and Dennis Kavanagh. British Public Opinion Issue 10.04 1988 May: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/1231/rec/1
Review of the Month. Labour Wins the Campaign but Makes Only Limited Gains. British Public Opinion Issue 09.05 1987 June: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/1173/rec/1

Polling in the 1990’s

1992 UK Election:

The 1992 Election was described in the MRS report as ‘the failure … was the most spectacular in the history of British election surveys’ and it triggered fundamental questioning of the methodology of opinion polls¹
The polling disaster of 1992 led to a radical revision of all aspects of opinion poll methodology²
Although this led to further innovation, moving further from the 1940’s and 1950’s, in other ways there were moves back to some of the techniques used back then ie random sampling²
The failure of the 1992 General Election polling was blamed on use of quota sampling, although this argument is too simplistic as it fails to acknowledge the highly successful run of polling using quota samples throughout the 1970’s and 1980’s²
In truth, pollsters tend to believe that both random and quota sampling both have their place in different circumstances²
In the 1990’s, pollsters became more receptive to telephone interviewing due to the ability to make a very large number of calls in a short period of time and thereby achieve response rates necessary if random samples are to have any reliability²
The opinion polls published on the day of the 1992 General Election produced estimates of party support significantly different from the actual share of votes cast. The total discrepancy over-estimated the Labour share of the vote by 4%, under-estimated the Conservatives by 4.5%, and over-estimated the Liberal Democrats by 1%³
In 1992, the 5 major pollsters averaged a 1.3% Labour lead, whereas the final result was 7.6% lead for the Conservatives due to a very late Tory swing. The polls backed the wrong winner by a wider margin than can be accounted for by sampling error and were widely criticised for not being able to perform to expectation⁴
The eve of poll polls showed the Tories at an average 38% and Labour at 40%, but in order to meet newspaper publication timelines, the voters were asked their voting intention some 36-48 hours before they actually voted, ie they had a day or 2 to change their mind after being polled⁵
Furthermore, exit polls could not be conducted until polling ended at 10pm on Election Day, but finished by 4pm in order to allow time to enable data analysis. This clearly made the assumption that those voting after 4pm would show the same pattern as those voting before 4pm. But in 1992 this did not happen; in 1992 the Tories came out in their droves on the evening of April 9, possibly spurred on by the thought of a Labour victory⁵
The NOP poll was conducted in 100 marginals and did detect a swing to the Conservatives as the evening progressed, but polling in the marginals has not provided the best basis for predicting overall seat changes in previous elections⁵
Key factors were a late change in voting intention which accounted for 3% of the difference, refusal to disclose voting intention (around 5% compared with the usual 3%, and therefore more significant). Conservative supporters were disproportionately less likely to reveal their voting loyalties than Labour supporters – a so called ‘spiral of silence’ (Noell-Neumann, 1984), and de-registration of voters and overseas voters, accounting for more than 1% of the difference^4,6
There was no single major technical fault with polling methodology which made it more difficult to address the problems for future Election polling^3,4,6
Furthermore, in 1992, the electorate were far less set in their voting patterns than elections since 1979 and MORI found 8% of voters said they made up their mind in the last 24 hours, while a Gallup poll conducted just after polling day found 14% of voters decided who they would vote for ‘during the last few days’⁵
In 1992, the turnout was relatively high at 77%⁵
Part of the problem was that this huge error was so unexpected as polls had been so good at predicting final outcome in the previous 3 General Elections, eg MORI polls were each within 1% of the votes cast for each of the 3 main parties, so the expectation generally was that they would always be a good predictor of final outcome^5,7
All the polls indicated that the election appeared to be so close – seemingly the closest for 20 years – meaning that the overall outcome was in doubt which added to the shock; the fact that the poll error misled most observers as to which party was likely to form the next government considerably magnified its impact⁷
The lesson learned from 1992 was that in measuring public opinion, we are measuring something which is inherently subject to change, particularly during a close-fought campaign⁵
Many newspapers lost confidence in the polls because of the 1992 debacle, some drastically cut back, switched between pollsters or even conducted their own¹
As such, there were efforts by AMSO, MRS and the major polling companies to improve the public perception of opinion polls by starting an initiative involving the setting up of a small advisory group responsible for briefing the media and politicians as to how to judge a ‘good’ and ‘bad’ poll, deal with complaints about the design or reporting of polls and provide impartial comment to interested parties⁶
The 1992 enquiry showed pollsters need to take the voting question less at face value and look at alternative questions, such as including questions regarding party identity, party leaders and best party for running the economy²
The 1992 debacle led to all companies experimenting with modifications to their methodologies and some introduction of innovations to get it right at the next election, however, there was no consensus as to how to resolve the problem. This led to more calls to ban polls but the view was that this would lead to more private polls being commissioned by parties, the City would expand commissioning, leaking and the manufacture of polls for their own benefit. Those who would be most disadvantaged would be voters and to ban polls would be an illiberal act⁴
A MORI poll conducted in January 1996 showed 69% of the British public said they would oppose a ban and only 22% would support it; but 18% said they would ban all coverage of the Election on radio and television, and 12% would ban such coverage in newspapers⁴
Prior to the 1997 General Election, 1996 saw ICM switch their polling over to telephone sampling using, if not random, then a quasi-random method²
In 1996, questionnaire design remained largely unchanged, the only development being introduction of a follow-up question asking those who were undecided or who refused, which party they are most likely to support²
As at 1996, the impact of new technology was minimal with interviewers still using paper questionnaires and recording answers with a pen²

1997 UK Election:

In 1997, ‘New Labour’ won a landslide victory at the general election after 18 years of Conservative government under the leadership of Tony Blair⁸
Despite the length of the campaign (6 weeks) and the eventual (predicted) low turnout, fewer people thought that there was too much media coverage than in either 1987 or 1992. In fact,1987 was thought by more people to have too much about the election in the newspapers and on TV than either of the successive elections, both of which were characterised as ‘the most boring election since the war’⁸
There was a view that Labour’s use of private opinion polls and especially focus groups was a new ‘Americanisation’ of British politics. In fact there was little new in Labour’s private polling that was not being done in Labour’s private polls in the 1970s, when MORI was Labour’s pollster; what was different was that for the first time the polling information was used effectively by the party leadership⁸
The Tories employed negative advertising, such as the ‘demon eyes’ poster, but the public dislikes negative campaigning, and although it can sometimes work, this was not the case with ‘New Labour, New Danger’, which seems to have been entirely counter-productive: nearly 4 times as many electors said that the ‘demon eyes’ campaign made them more likely to vote Labour as said it made them less likely to do so⁸
There was a view that the opinion polls got the election wrong. However, the ‘poll of polls’ predicted the Tory share of the vote to the nearest whole number, and was within the standard margin of error for the other parties. Furthermore, polling in Scotland and constituency polls was extremely accurate. The polling companies had solved the problems which beset their ‘predictions’ in 1992. Polling is not a crystal ball that can offer predictions of supernatural accuracy, being subject both to the limits of statistical sampling and the behaviour of the British voter that is sometimes unpredictable even to the voter himself or herself. However, even with these necessary reservations, it remains a better guide to the political climate than any of the less scientific alternatives⁸
Conversely, the pundits got the election wrong in most cases. Of the 20 pundits on the Reuters panel of so-called political experts, the average eve-of-poll forecast was a Labour majority of 92, just over half of the 179 majority on the day, and compares to the average 159 seat projection recorded by the opinion polls. Not one of the 20 pundits was as close to the Tory share of the vote as the opinion polls’ average⁸
There were 47 national polls conducted in the 1997 election by the 5 full members of APOPO (Association of Professional Opinion Polling Organisations). Their average share throughout the campaign was 31% for the Tories; the Conservatives got 31%⁹
Just over 6 people in 10 (63%) said they were aware of any national opinion polls in this election, compared to 89% at the 1992 Election, when there were more polls over a shorter period⁹
Many more men (73%) than women (54%) were aware of having heard poll findings in the previous week (18-25 April), including two-thirds of those aged 45 and over, and three-quarters of middle class respondents, but only just over 55% of working class voters⁹
Nearly all (94%) of those aware of having seen any polls recalled that the Labour party was in the lead⁹
Only 16% of the public believed polls should be banned during elections. Those opposing a ban were more likely to be men (78%), younger (82% of the under 35’s, 77% of 35 and over) and more likely to be middle class (82%) than working class (75%)⁹
Over half (55%) said they trusted pollsters to tell the truth, compared with only 12% who said they trusted Government Ministers⁹
A report in 1996 showed that 30 out of 78 countries had some kind of restrictions on opinion polls, including an outright ban on their publication before Elections¹⁰
George Vassiliou, Chairman of The Foundation for Information and a former President of Cyprus said, ‘Speaking as a politician, we adore polls as long as they say what we like’¹⁰
Following the 1997 General Election, Nick Sparrow, MD of ICM, one of the UK’s 5 leading political polling companies, called on his fellow pollsters to ‘properly investigate’ the different methodologies used by pollsters in 1997’s General Election. Despite favourable media coverage of the polls’ performance, ‘we should avoid the temptation to bask in the perceived glory of success and learn from the experience’. He points out that most polls carried out during the 6 week election campaign ‘continued to display the same extent of pro-Labour bias apparent in 1992’¹¹

References:

Verbatim (Gordon Heald). Research Magazine issue 349 1995 June: https://amsr.contentdm.oclc.org/digital/collection/ResearchMag/id/4253/rec/9
Why a Sample of the Polls Suggests a Vote in Favour. Nick Moon: Research Plus issue 360 supp (50 years of market research) 1996 May – Research Magazine (1992-2012) – The AMSR Online Archive (oclc.org)
Report into the performance of Opinion Polls in the 1992 UK Election (initial report and press release: https://amsr.contentdm.oclc.org/digital/collection/p21050coll1/id/205/rec/1
Polling in Britain: A historic route to the forecasting future. Research Plus issue 358 supp (into the future) 1996 March: https://amsr.contentdm.oclc.org/digital/collection/ResearchMag/id/4133/rec/1
After the poll is over ….. Peter Hutton. The Market Research Society Newsletter Issue 314 1992 May: https://amsr.contentdm.oclc.org/digital/collection/MRSNewsletter/id/4707/rec/3
The Chairman’s view on the past year: AMSO Annual Report 1995-1996 – Company Reports & PR – The AMSR Online Archive (oclc.org)
The Opinion Polls and the 1992 General Election (Full Report)The Opinion Polls and the 1992 General Election (Full Report) – Opinion Poll Reports – The AMSR Online Archive (oclc.org)
Letter to Our Readers. British Public Opinion Issue 22.06 1999 August: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/1025/rec/1
Letter to Our Readers. British Public Opinion Issue 20.03/4 1997 April/June: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/1318/rec/1
Report Highlights Restrictions on Political Opinion Polls. Research Magazine issue 377 1997 October: Research Magazine issue 377 1997 October – Research Magazine (1992-2012) – The AMSR Online Archive (oclc.org)
Sparrow Call for Poll Probe. Noelle McElhatton: https://amsr.contentdm.oclc.org/digital/collection/ResearchMag/id/5122/rec/4

Polling in the 2000’s

2001 UK Election:

Turnout for the 2001 General Election was expected to be even lower than the post-war record set in 1997¹
In 2001, only 48% said that they were ‘certain to vote’, compared with 57% in the corresponding 1997 poll, when 64% were ‘certain’ or ‘very likely’ to do so, compared with 72% in 1997, when the eventual turnout was 71%¹
Labour seemed most likely to lose out from a low turnout as 63% of Conservatives and 62% of Liberal Democrats but only 52% of Labour supporters were certain to vote: such a differential would have cut Labour’s lead from 19% to 12%, equivalent to a 3.5% swing¹
In the 2001 General Election, the media decided, wrongly, that the public were finding it the most boring election ever, and there seemed to be the erroneous assumption that the campaign was pointless, that nobody had changed their minds and that nobody was ever likely to do so²
Although a higher proportion of the electorate had made up their minds before the start of the campaign than was the case in 1992 (74%), still 7% made their final decision within the last 24 hours (after the final opinion polls stopped interviewing), and another 7% in the previous week. Among those who voted Liberal Democrat, 10% said that they decided to do so on the final day and another 13% in the previous week²
The 2001 election saw turnout drop from 71% in 1997 to just 59%. The reason was not apathy but the political leadership of the parties and the media failed to connect with many potential voters³
Pundits and politicians often blame apathy for low turnouts but among the public as a whole, interest in politics has remained remarkably stable over the past 3 decades, with no significant decline on the part of the electorate³
In fact only about 1 in 6 people are ‘very’ interested in politics, and about 1 in 8 always wish it had never been invented³
In 2001, interest in news about the election was higher compared to some earlier elections³
The 2001 General Election was a lacklustre campaign and failed to engender the stimulus needed for people to go out and vote³
A MORI poll for The Times in the 4^th week of the campaign (29 May), 4 questions were asked that were also posed at similar stages of the 1992 and 1997 campaigns, measuring how many of the public said they were ‘very interested’ or ‘fairly interested’ in news about the election, what the polls were saying, the party election broadcasts and politicians’ speeches³
Surprisingly, in view of the general impression of apathy, more of the public said that they were interested in in each of these 4 than was the case in 1997. In fact, more of the public said they were interested in politicians’ speeches than in 1992, one of the highest turnout elections of modern times!³
A noticeable feature of this election was how early in the proceedings the media came to the unanimous conclusion that they were bored and so concluded this was must be the case for their readers or viewers. Could it be that that part of the public which takes an interest in politics and will vote – still a majority, apathy notwithstanding – far from feeling sated by the coverage, simply found it whetting their appetites? And as for politicians’ speeches, they rarely had the chance to hear them – only sound-bites, and then mostly from the leaders³
The numbers feeling coverage of the campaign on the television and in the newspapers to be too high were significantly lower in 2001 than in recent previous elections; and this was much truer of the press coverage than of television³
This indicates that if the public did not necessarily want more information in total, they did want different types of information, that is to say more candidate-focused and more policy-focused information, less on the daily campaigning circus in front of the cameras and notebooks³
Furthermore, a post-election recall of voters previously interviewed in the first week of the campaign found that, although the switching of voters between the 3 main parties was lower than in any other recent election, the total ‘churn’, which includes ‘don’t knows’ making up their minds and those who declared a voting intention deciding not to vote, was greater than it had ever been, 33.7%, more than a third of the electorate. Such figures make a nonsense of any analysis of the polls which assumes that nothing moved, or averages poll findings across the whole campaign²
There was movement, there was late swing (in both cases mainly to the LibDems) and there was a massive decision by ‘Old Labour’ supporters in safe seats not to turn out, thereby hurting Labour’s share but without damaging its majority²
The outcome of the 2001 General Election was a foregone conclusion, as it was in the 1997 and 1983 General Elections, so that cannot be the cause of the decline in voter participation. The levels of interest in politics and news about the election should give political and media decision makers pause for thought, and inform them that they have to consider what the public are telling them about the way elections are managed in this country³
Another myth that has been propagated, perhaps based on exaggerated estimates before the campaign even began, was the 20% of the electorate might vote by post. MORI found a fairly consistent 6-7%, but never more than 8%, intending to apply, and it appeared about 4% of electors actually did so, amounting to approximately 6% of the turnout, a far cry from 20%!²
There was criticism of the accuracy of the polls conducted for the 2001 General Election, particularly regarding the final result. The polls were criticised for over-estimating the strength of support for Labour, the final polls from Gallup and NOP over-estimating the Labour lead by 8 points, MORI by 6 points, with the closest being ICM giving Labour a 2 point lead over the Tories⁴
Commenting on the pollsters’ overall performance Nick Moon, Director of NOP, said: ‘They’ve done okay, not brilliantly, but there is no room for complacency’, ‘We got the big picture right but there is still room for manoeuvre’. The results suggest the polling community has still to resolve the issue of shy Tory voters, identified as one of the reasons for failing to predict a Conservative victory in 1992⁴
Meanwhile, Nick Sparrow, MD of ICM Research, suggested in The Guardian (10/09/01) that pollsters should shoulder some of the blame for Labour’s overstated advantage in the polls preceding the 2001 election. He said that a tendency for Conservatives to be under-represented in poll samples, as well as unprompted voting questions failing to remind people of the Liberal Democrat alternative, may have inflated the Labour lead⁵
There was less polling commissioned for the 2001 General Election than in previous years which was attributed to a downturn in advertising and rise of the cost of newsprint. ‘It was also because it was clear who was going to win’⁴
However, these criticisms did not take into account changes in mind that took place over the course of the campaign, and the shockingly low turnout meant that peoples’ views represented in polls chose not to turn out and vote on Election Day⁶
It is important to remember polls do not predict; they measure the present. Their aim is not to allow newspapers to print tomorrow’s news today, but to provide insights into motivations, trends and how the public reacts to particular politicians, policies or events⁶

References:

Election turnout may fall further. British Public Opinion Issue 24.01 2001 January/February: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/588/rec/1
2001: The Polls and the Election Myths. British Public Opinion Issue 24.03 2001 June: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/703/rec/1
Letter to Our Readers. Robert M Worcester. British Public Opinion Issue 24.04 2001 October: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/414/rec/1
Pollster Win Not Brilliant. Vivienne Wilson. Research Magazine issue 422 2001 July: https://amsr.contentdm.oclc.org/digital/collection/ResearchMag/id/2236/rec/1
Pollsters to Blame? Yvette Mackenzie. Research Magazine issue 425 2001 October: https://amsr.contentdm.oclc.org/digital/collection/ResearchMag/id/2414/rec/1
Letter to Our Readers. Roger Mortimore. British Public Opinion Issue 24.09/10 2001/2 December/January: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/307/rec/1

2005 UK Election:

In 2005, attitudes to politics were negative, with deterioration in in political engagement identified which had continued since 2003¹
A survey in the General Election year of 2005 found 13% of respondents were ‘not at all interested’ in politics¹
There was a feeling of powerlessness during 2005 with 44% of respondents surveyed feeling that they could not change the way the country is run by getting involved in politics. This was the highest level of dissatisfaction found across 5 surveys conducted by Ipsos MORI. It may be significant that this was the survey conducted at the end of an election year¹

References:

Audit of Political Engagement 5 for Hansard Society (December 2007). https://amsr.contentdm.oclc.org/digital/collection/Mori-reports/id/78/rec/1

2010 UK Election:

No relevant information found

2015 UK Election:

The British Polling Council (BPC) and MRS launched an enquiry into the performance of the opinion polls in the UK preceding the May 2015 General Election, particularly focussing on methodology, with a need for transparency from pollsters and access to unweighted data¹
Unlike the 1992 enquiry set up by MRS, the 2015 was totally independent from the polling sector, comprising mainly academics, whereas in 1992 leading pollsters predominated. It was hoped the 2105 report would be available promptly ie by March 2016¹
Representatives of each of the main polling companies (ICM, Opinium, ComRes, Survation, Ipsos-Mori, YouGov, Populus) agreed that being within ‘sampling error’ was not a good enough excuse in predicting the outcome¹
There was little evidence to support a late swing of any significance in 2015, based on the results of post-Election polls, but the results showed the situation of understanding the Conservative vote and overstating support for Labour, seen in 1992 and 1970, was still at the heart of the problem¹
Martin Boon (ICM) pointed out that there had been no Tory lead in the polls since 2012
Neither the panel nor the polling companies are advanced in their investigations and, whilst all the polls misrepresented the mood of the electorate, that does not mean there was a common view on the reasons¹
In today’s world the 2 main options for data collection by the pollsters are telephone or online. With the mobile revolution, landline based methods are no longer guarantee adequate coverage, so mobile top-up quotas are used, but the proportions varied in 2015 polling¹
Issues identified by pollsters were: sample bias (telephone and online methods), identifying the voting public and estimating likely turnout, weighting methods/modelling, differential turnout, tactical voting, factors possibly unique to the 2015 Election, such as a favoured Prime Minister, overstating Labour and understating Conservative support (a seemingly systematic problem)¹
Consideration was given to: seat projection modelling for the future, a review of the constituency level polls commissioned by Lord Ashcroft in Scotland which appeared to have produced a higher degree of accuracy, a question as to whether ‘modelling’ rather than sampling error lay at the heart of the problem, consideration of the importance and impact of the image of individual party leaders, and should the voting intentions question replicate the voting slip content for each constituency in which opinion polls take place?¹
Interpretation and media coverage must be considered and current evidence points to 3 key areas: sampling, question design, weighting/adjustment/modelling methods¹
The enquiry comprised a review of the overall research design used in opinion polling, especially the challenge of representing the voting public, rather than the general public¹
There was concern that solving the 2015 problem may not provide a generic, long term solution¹
A big challenge was predicting turnout (ie ICM predicted a turnout of 86% in 2015 but this was well above the actual level of 66%), with apparently distinct differences in the turnout behaviour between Conservative and Labour voters, which needed to be investigated¹
We now live in a much more complex world. In 1992, telephone polling using fixed line phones was a relatively new methodology and face-to-face polling was still used by pollsters, but now we employ the use of internet, mobile phones and social media¹
There was a need for ‘fit for purpose’ data collection methods and current methods needed to be scrutinised accordingly¹
The MRS Delphi report, ‘Deconstructing Bias’, recommended a hypothesis approach to building algorithms, taking into account relevant legislation, such as GDPR and equal opportunities. The report concluded with a case study summarising the salutary lessons learned from the enquiry into the failure of the polls to predict the outcome of the 2015 UK General Election, citing unrepresentative samples and confirmation bias at the heart of the issues²

References:

IJMR Editor blog 7 2105. Researching voting intentions: the BPC (British Polling Council)/MRS enquiry. Peter Mouncey: https://amsr.contentdm.oclc.org/digital/collection/PapersOffprnt/id/11590/rec/5
Applying artificial intelligence tools in research (2018). Peter Mouncey. https://amsr.contentdm.oclc.org/digital/collection/PapersOffprnt/id/10517/rec/1

2017 UK Election:

Just over three-quarters of the electorate (76%) voted in the1979 General Election. Since then, considerable concern has been expressed about falling levels of turnout, with just 59%of the electorate voting in 2001, the lowest rate in the post-war period. However, turnout has gradually risen again since, reaching 69% in 2017¹
Levels of political interest are higher now than they were in 1986, according to the British Social Attitudes survey. Then, 29% described themselves as having a ‘great deal’ or ‘quite a lot’ of interest in politics, compared with 42% in 2016. But we are less likely now to identify with a political party (although most of us, if pushed, still do); in 1986, 8% did not identify with any party, rising to 14% by 2016¹

Reference:

Four Decades of Social Change. Research: Matters 2018 December: https://amsr.contentdm.oclc.org/digital/collection/MiscNewslttrs/id/1211/rec/1

Banning the Polls

In Britain, Parliament is free to pass any legislation it wishes, and so has been described as a ‘democratic dictatorship’. Therefore in 1967 when the Speaker’s Conference recommended polls to be banned in the immediate period before an election day, it had to be taken seriously¹
In a free society, polls cannot effectively be banned. If they were, political parties would still do polling and leak the results, the City would do private polls and leak them, and foreign media would commission private polls in Britain and publish them. Certainly the impact on the City of polls, real or fictitious, can have a massive impact by wiping billions of pounds off share prices. In 1979, the Evening Standard (McGrandle, 1979) carried the headline ‘Should We Ban Opinion Polls?’ on the basis it would make life much easier for the fund managers and brokers and avoid panic in the City. The conclusion was that this was not a realistic option and the City would have to live with them¹
In terms of the public, a study by the Market Research Society in 1978 showed 84% of the sample said they had noticed the results of opinion polls in newspapers. 33% said they were pointless/take no notice of them. Only a third of the public gave polls credit for generally ‘getting it right’, with 46% saying the opinion polls are normally wrong¹
In 1986, the Market Research Society decided to assist in effort to obtaining proper understanding of uses and limitations of polls and to give guidance to journalists and others reporting polls and watchdog against rogue polls²
In a poll conducted just after the 1987 General Election, MORI found 24% of the British public in favour of banning political opinion polls during General Elections³
In 1987, the major polling organisations formalised a loose association that had been in existence for a decade as the Association of Professional Opinion Polling Organisations and reaffirmed their adherence to a Code of Practice originally agreed in 1971. All the Association’s members provided interested parties open access to their methodology and lodged their polls with Quantime, a data processing bureau, providing access to raw data, weighting factors and findings from members’ polls²
Over the years, and particularly in the run-up to and just after General Elections, MORI has measured the British public’s support for a ban on the publication of opinion polls during election campaigns. From time to time, voices are raised in support of such a ban. These calls come mainly from politicians who have taken the view that if opinion polls were banned, the public would not know how unpopular they really were³
Fortunately, the public has been opposed to a ban on the publication of opinion polls. Our most recent measure among a sample of 1,887 adults aged 18+, interviewed face-to-face, in home at 147 ward sampling points throughout Great Britain between 26-29 January 1996 found that 7 in 10 believe there should not be a ban on the publication of opinion polls during election campaigns. We must not be complacent, however, because 22% of the public would favour a ban on the publication of polls³
Except for a short-term peak immediately after the 1992 election, support for banning polls over the last 2 decades has remained very steady, with just over a fifth of adults being in favour³
After the 2015 General Election prediction failings, the possibility of a Bill in Parliament to limit polling in the run-up to future elections was mooted⁴
There are still many countries where it is not permitted to conduct and publish opinion polls. In some countries unprofessional polls are conducted which create problems, and in several others restrictions of one or another kind are in existence. Even among the Western democracies there are many which seek to prohibit the publication of opinion polls one, 2 or even 3 weeks prior to an election. We believe it is a fundamental democratic right to conduct and publish the results of public opinion polls⁵

References:

Pollsters, the Press, and Political Polling in Britain. Robert M Worcester: https://amsr.contentdm.oclc.org/digital/collection/PapersOffprnt/id/1750/rec/37
The Market Research Society 35^th Annual Conference 18-20 March 1992. MRS Conference Papers 35th Annual Conference 1992: https://amsr.contentdm.oclc.org/digital/collection/Conf-MRS/id/5178/rec/1
Letter to Our Readers. Robert M Worcester. British Public Opinion Issue 19.02 1996 March: https://amsr.contentdm.oclc.org/digital/collection/MORI-BPO/id/652/rec/1
IJMR Editor blog 7 2105. Researching voting intentions: the BPC (British Polling Council)/MRS enquiry. Peter Mouncey: https://amsr.contentdm.oclc.org/digital/collection/PapersOffprnt/id/11590/rec/5
The Freedom to Publish Opinion Poll Results: Report on a Worldwide Study; https://amsr.contentdm.oclc.org/digital/collection/Reports_/id/709/rec/64