The politics of polling Contents


Since the first Gallup polls were run in the United Kingdom in the 1930s, political opinion polls have become an established aspect of British politics. For the media and political parties, political opinion polls are integral to understanding the political mood of the nation. Polling results have also come to dominate election coverage, driving the ups and downs, twist and turns of the ‘narrative’ of each election. The public and others are aware that there is always potential for some inaccuracies in polling. Nonetheless, polls that estimate voting intention continue to serve as prominent and highly influential sources of information during election campaigns.

In recent years, however, the polling industry has suffered a number of collective failures. The 2015 General Election was widely considered to be an embarrassment for the polling industry. In the weeks prior to election day much of the focus was on the possibility of a hung Parliament, only for a Conservative majority to emerge. 2015 showed just how influential polling had become, both in the extent to which we have come to rely on poll findings to understand political events, and the considerable impact that polling can have. This realisation prompted serious concerns about both the ability of the polling companies to make accurate and useful estimates of the outcomes of national political events, and around the prominence that polling has come to have. These concerns were only compounded when polling companies were seen to fail again, first in the EU Referendum vote in 2016 and then in the 2017 General Election.

Expert analysis has already been undertaken to understand the reasons behind these polling failures, most notably the Report of the Inquiry into the 2015 British general election opinion polls, chaired by Professor Patrick Sturgis. This inquiry and subsequent analysis of the 2016 EU Referendum and the 2017 General Election identified important recommendations on how the industry can continue to innovate and adapt to improve its approach to estimating voting intention.

Despite the efforts made to identify and address the methodological challenges associated with recent polling failures, important questions remain. Does the recent poor performance signal a broader trend that polling is getting less accurate? Is polling getting harder to do and, if so, why? Ultimately, does it matter if polling continues to produce inaccurate predictions? It was these questions that we sought to address, and the last of these we considered the most pertinent. Our central concern was that, if it is becoming less likely that polls can provide accurate estimates of the likely election outcomes, then there is a significant risk that future elections will be affected by misleading information, potentially distorting the democratic process.

The available data on longer-term trends in polling performance suggest that, currently, it would not be correct to say that we are witnessing a decline in the accuracy of polling. That said, although polling performance has not worsened in a statistically significant way, there is little doubt that confidence in polling has been shaken. We saw evidence of this scepticism in the last General Election. The question ‘can we trust the polls?’ featured prominently in the 2017 election coverage and a number of broadcasters told us that they had deliberately reduced the prominence given to voting intention polls in their election coverage.

A combination of factors is making it harder to estimate political opinion accurately. We heard that it is getting harder to persuade members of the public to take part in polls and surveys, and that shifts in demographic predictors of the vote, and an increasingly volatile electorate, have all made polling harder to do. This means that polling companies continually need to adapt and innovate. We believe that this can only be done successfully if polling performance is routinely and thoroughly reviewed after each election.

The polling industry is facing a number of challenges in polling the modern electorate but, in the overwhelming majority of cases, we recognise that polling companies make every effort to ensure accuracy and that they have no reason to deliberately distort poll findings. These efforts, however, can be undermined by the ways in which the media reports on polling. Media coverage of election campaigns has traditionally involved a considerable focus on polling information, but this coverage is not always an accurate reflection of polling data. Headlines that over-emphasise small, insignificant changes in party fortunes are misleading, but is a practice that remains widely prevalent. We believe that the British Polling Council, in collaboration with other expert groups and regulators, should use its considerable expertise to develop further guidance for the media on the reporting of polls, and work proactively to highlight particularly bad examples of media reporting on polling.

There are other types of polls which affect political discourse in the UK, such as those that measure public opinion on political and social issues. We found that some of the key problems we identified for polling, particularly the use of leading questions and misleading presentation of results, were more pronounced for policy issues polls. We feel that there is a clear need for more oversight of the conduct and reporting of such policy issues polls.

In the face of considerable challenges posed by the difficulties of polling the modern electorate and the misreporting of polls, and in light of the damage done to confidence in the accuracy of polling, the oversight of polling also needs to change. There is no overall framework for the regulation of polling. Instead, some oversight of polling is provided by professional bodies such as the British Polling Council and the Market Research Society, while the reporting of polling is regulated to a limited extent by broadcasting and press regulatory bodies. Oversight of polling is fragmented and disjointed, with different elements of regulation disconnected from each other.

The current system is not satisfactory and we therefore recommend a co-ordinated approach towards the oversight of polling, involving the British Polling Council, the media regulators and the Electoral Commission. The British Polling Council’s remit should be expanded to take on a greater standard-setting and oversight function. Media regulators should tackle quickly any instances of misreporting of polling. Finally, the Electoral Commission should take on an enhanced role in the monitoring of voting intention polling during election campaigning periods. We hope that these bodies in combination can provide more effective oversight of polling.

When considering the way in which political information, such as voting intention polls, is produced, disseminated and understood, it would be short-sighted to ignore the profound change in the way we access and share news, political developments and opinions. The enormous development of the internet and the rise of digital media has undoubtedly impacted on polling, opening up the industry and making the polling market easier to enter. It has and will continue to provide risks and opportunities for innovations in polling, such as the use of social media data to help predict or influence elections.

We were also alerted to very serious concerns about the impact of digital media on politics more broadly. These concerns included: the deliberate spread of political misinformation; the reinforcement of the ‘echo chamber’ effect exacerbated by social media; and the lack of accountability and transparency around political advertising and campaigning online. A wide range of issues were described but a clear consensus emerged on the need for further action to better understand the nature and scale of the impact of digital and social media on politics, and to identify the appropriate approach from Government, the regulators and the polling companies themselves. These issues stretched beyond our remit and we simply did not have the capacity to give due attention to them all. However, the issues raised are complex and rapidly evolving, representing a very significant threat to our democratic processes. We therefore recommend that the Government should address these challenges as part of its ongoing work on the Digital Charter.

Both political polling and digital media have the potential to influence the democratic process in the UK. With improvements to the system of oversight for polling, we believe that some of our concerns in this area could be addressed. However, we consider the issues relating to digital media to be far more serious, with grave implications for our democratic processes. We acknowledge that addressing some of these issues will be extremely difficult but we are clear that the challenges we have identified in this report should be tackled urgently and holistically, in order to ensure that the UK’s democratic process is protected and maintained for the future.

© Parliamentary copyright 2018