Summary of the impact

Research conducted at Essex as part of the British Election Study (BES) successfully demonstrated the validity of internet-polling methods. This research has been integral to the development of YouGov and the growth of the online-survey sector in the UK. In addition, the BES has: informed public opinion about the likely consequences of introducing AV in UK elections; led to the creation of a spin-out partnership, BPIX, which conducts polls for major national newspapers; informed parliamentary debates about the causes of non-voting; and served as a template for national election studies in other European countries.

Underpinning research

Essex has a long tradition of electoral research: the British Election Study, which was based there throughout the 1970s, returned to Essex in 2001 under the leadership of Professors David Sanders and Paul Whiteley, in collaboration with two colleagues at the University of Texas at Dallas, Professors Harold Clarke and Marianne Stewart.

Funded by the ESRC with further support from the US National Science Foundation, the BES has conducted nationwide sample surveys during each of the General Elections 2001, 2005 and 2010, with monthly surveys, latterly panel in form, in between. The results of that research have been reported in major monographs co-authored by the four principal investigators, and in a raft of journal articles and book chapters. Many important substantive themes and findings have emerged from this research, concerning (among many other things):

• The role of policy performance, particularly in relation to the economy in driving electoral support;
• The decline of a sense of attachment to political parties in the minds of the electorate;
• A weakening of the sense of duty to participate, particularly among young people, which has produced a long-term decline in turnout;
• The likely consequences of introducing AV in UK parliamentary elections.

Many of those specific findings are picked up in public discourse and inform policy deliberations at the highest levels, in ways discussed in Section 4 below.

The BES is not only a national resource for exploring and explaining public opinion on the major issues before the electorate, but is also a means for pioneering methodological developments that have subsequently become accepted and widely used. For instance, in the early 2000s it was an open question whether online surveys could reliably substitute for traditional survey methods. The matter was the subject of a testy 2004 exchange in the pages of the International Journal of Market Research, with an editorial describing what was at stake:

"There is no issue more topical or more important in market research at the moment than internet surveys.... To put it somewhat crudely, if surveys via the internet can prove their worth, and given the increasing cost and difficulty of data collection both face-to-face and by telephone, the entire market research industry may be facing the biggest shake-up it has seen in decades" Warren, M. (2004) Editorial. International Journal of Market Research 46 (1): 1-2, at p1.

A methodological breakthrough of signal importance lay in establishing once and for all, through careful analysis of the 2001 and 2005 BES data, that internet-based surveys could be as reliable as in-person interviews (Sanders et al. 2004; 2007; Clarke et al. 2008). The key finding was reported succinctly as follows:

"We have conducted extensive statistical tests comparing the properties of the internet panel data and the in-person probability sample. Results of these tests (Sanders et al. 2007) demonstrate that the Internet and in-person data have very similar distributions on key variables and yield virtually identical parameter estimates for a wide range of comparable models of party choice and turnout" (Clarke et al. 2009, p. 21).

As demonstrated in Section 4, below, the BES was instrumental in first establishing, and then expanding, the use of the online survey as a tool for sampling public opinion.

References to the research

Research funding

Sanders, D., P. Whiteley and H. D. Clarke, British Election Study 2001/02. ESRC, 01.12.00 to 31.7.03, £588,950.

Sanders, D., P. Whiteley and H. Clarke, British Election Study 2001/02: supplementary award. ESRC, 07.03.01 to 31.07.03, £232,012.

Sanders, D. and P. Whiteley, British Election Study 2001. BBC, 01.10.01 to 30.09.02, £5,000.

Sanders, D. and P. Whiteley, British Election Study 2005-2006. ESRC, 01.04.04 to 31.03.07, £917,401.

Sanders, D. British Election Study 2005-06 — supplementary award. Electoral Commission via ESRC, 01.04.04 to 31.03.07, £95,000.

Sanders, D. British Election Study 2005-2006 — supplementary awards. ESRC, 01.04.04 to 31.03.07, £10,000 and £8000.

Whiteley, P. British Election Study 2009-10. ESRC, 01.09.08 to 31.08.12, £1,507,632.

Total: £3,363,995

Details of the impact

The BES, whilst at Essex, has had impact in five main contexts: pioneering online survey methods; informing public opinion on AV; enabling the formation of a commercial spin-out; informing UK parliamentary debate; and shaping other countries' election studies.

In demonstrating the validity of internet polling in the UK the BES helped alter the way survey organisations sample public opinion
The BES's pioneering of online surveys was not simply a breakthrough in survey methodology but it helped to change how public opinion is sampled. YouGov initially used online methods in teaming up with the BES in analysing the 2001 General Election. In March 2013 the President of YouGov confirmed that the success of this meant that YouGov was emboldened to expand its operations dramatically in an online direction, with increasing use of online surveys in subsequent elections. He also states that the BES's research convinced the survey sector as a whole of the validity of internet-based surveys and that "the effects of this research have been enormously valuable in the last five years, with the continued growth of online polling" [corroborating source 1]. Further to this, YouGov's Special Projects Director notes in a journal article that the "expansion of the online survey sector has changed the nature of public opinion research in ways that have important implications for scientific inquiry... The promise is a significant increase in high quality research, with data gathered in a very timely way" [corroborating source 2].

The BES informed public opinion on the likely consequences of introducing AV into UK parliamentary elections
The 2010 BES included a set of questions that sought to simulate how respondents would have voted had the 2010 election been based on AV. Extensive statistical analysis of these results enabled the team to provide a definitive picture, published in December 2010, of how AV would have affected the outcome of the 2010 general election. The result would have been only marginally different from the actual First-Past-the-Post result but, crucially, the Liberal Democrats would have been in a position to form a majority government with either Labour or the Conservatives.

This finding from the BES was picked up by The Guardian and, along with quotations from an interview with Sanders, was used to demonstrate the difference to the post-election landscape that AV would have made [3]. More significantly, this finding from the BES was used by both sides campaigning in the AV referendum. The `No' campaign used the BES order to support their claim that the introduction of AV would serve mainly to strengthen the `king-making' role of the Liberal Democrats in government formation. The No2AV campaign used this argument in its leaflets without crediting the BES, but the BES was cited by various Labour websites campaigning for a `No' vote [4]. However, the LabourYes campaign cited the BES in its claim that Conservative opposition to AV was based simply on self-interest as the party would have lost seats in recent general elections [5].

The BES led directly to the formation of a profitable, commercial spin-out, BPIX
The Essex BES team created a commercial spin-out, the British Polling Index (BPIX), in 2007 with the express aim of raising funds to support BES election surveys. BPIX offers a survey-commissioning service to major national newspapers and media outlets, involving the design of questionnaires and sampling frames. It has had a series of contracts with the Daily Mail and the Mail on Sunday. The polls that it conducts are typically included in measures of the UK `poll-of-polls', for example as reported by the UK Polling Report during the 2010 general election campaign. Since 2009, BPIX has raised over £60,000 in profits that have been ploughed back into purchasing more and better BES surveys [6].

The BES has informed UK parliamentary debates
The BES is not simply used by analysts and the media, but is also a tool used by politicians to better understand politics. The BES has been cited in parliamentary debates and been used as evidence in informing the decisions of politicians/political parties. For instance, Paul Whiteley worked as a Specialist Advisor to the Speaker's Conference on Parliamentary Representation, the report of which was published in 2010. The report cites BES findings on the strength of party identification in Britain (p. 32), but more significantly it cites findings from the BES's survey from May 2009 about the lack of public trust in politicians and parties (which had originally been cited in the Summary and Recommendations of the Conference's interim report) [7]. On the first sitting day of 2012, there was a long (over 140 minutes) parliamentary debate on diversity, growing out of and frequently referring back to that Speaker's Conference report. There was much discussion of the measures that various parties had taken to address those concerns since that report, and a resolution was passed encouraging them to do more [8].

The BES has also informed parliamentary debates about the funding of political parties. In its 2006 report on party funding, the House of Commons Constitutional Affairs Committee drew on findings from the BES. Those debates culminated in a set of recommendations from the Committee on Standards in Public Life, which were presented to Parliament in 2011. This 2011 report includes a note of dissent from Oliver Heald, the Conservative Party representative on the committee, which cites the BES's findings on the political affiliations of trade union members [9]. Heald ultimately refused to sign the report.

The BES shapes the way other countries sample the opinion of their citizens
Knowing what citizens think, and why, is central to democratic governance. National election studies are the premier scientific instrument through which democracies explore what lies behind the brute fact of the vote. National research councils generously fund such studies, as a matter of high national importance. The BES at Essex is widely regarded as a template for other national election studies, and the BES team is regularly asked to advise on the design of national election studies elsewhere in Europe. A Co-Principal Investigator on the Italian study has confirmed that "the design innovations introduced in the 2001, 2005 and 2010 British Election Studies had [a] very significant effect on the design of the Italian National Election Study from 2006 onwards" [10]. He goes on to say that "the studies have employed internet components in the overall design, particularly for rolling campaign panel surveys, which interlock with face-to-face surveys in the manner pioneered by the BES at Essex" [10].

Sources to corroborate the impact

All documents are available from HEI on request.

[1] President of YouGov.

[2] Twyman, J. (2008) Getting it right: YouGov and online survey research. Journal of Elections, Public Opinion, and Parties, 18 (4): 343-54. See p. 353. DOI: 10.1080/17457280802305169

[3] Patrick Wintour, 'AV would have given Liberal Democrats real choice of coalition partner: Essex University study says Nick Clegg's party would have gained 32 more seats under the Alternative Vote system', The Guardian, 15 July 2010.
http://www.guardian.co.uk/politics/2010/jul/15/alternative-vote-british-election-study

[4] Examples of BES being used to support `No' campaigns:
http://www.grassrootslabour.net/index.php?limitstart=42http://www.grassrootslabour.net/index.php?limitstart=42
http://labourlist.org/2011/05/vote-labour-vote-no/
http://conservativehome.blogs.com/files/notoav_ea2_final.pdf

[5] Examples of BES being used to support the LabourYes campaign:
http://labouryes.org.uk/why-is-cameron-so-desperate-for-a-no-vote/
http://labouryes.org.uk/wp-content/uploads/2011/02/Why-is-Cameron-so-desperate-for-a-no-vote-Labour-Yes-briefing.pdf

[6] BPIX accounts, 2009-2011.

[7] House of Commons (2010) Speaker's Conference (on parliamentary representation): Final report. London: The Stationery Office.

[8] Hansard 12 Jan 2012: Cols. 403-442.
http://www.publications.parliament.uk/pa/cm201212/cmhansrd/cm120112/debtext/120112-0003.htm#12011294000003

[9] Committee on Standards in Public Life (2011) Political party finance: Ending the big donor culture, Thirteenth Report of the Committee on Standards in Public Life, Cm 8208. See Appendix 8, pp. 109-110. http://www.official-documents.gov.uk/document/cm82/8208/8208.pdf

[10] Co-Principal Investigator, Italian National Election Study.