In the aftermath of recent (and surprising) election results, it has become apparent that poll results do not tell the whole story of voter intentions. In a study published in EPJ Data Scienceresearchers from the University of Leeds they mapped voter sentiment across all UK constituencies based on e-petitions data, matching well with the results of the 2017 general election.
Wikimedia Commons
Guest posts by Stephen Clark, Nik Lomax and Michelle A. Morris
The EU referendum and the 2017 general election are two recent examples where polling companies have failed to accurately predict the outcome of voter sentiment. Most predicted that the UK would vote to remain in the European Union and that the Conservative Party would increase its parliamentary majority. When none of these findings surfaced, there was much criticism of the data sources and methods used to gauge voter sentiment and opinion.
More data on political sentiment than ever is becoming available and, in our view, provides an opportunity to better understand the views of voters in the UK. One source, used in our article published in EPJ Data Science is e-petitions data, collected by the UK Parliament. We use this data to rank Westminster parliamentary constituencies by distribution and concentration of signatories.
UK borough cluster map. Clark et al. 2017 EPJ Data Science
In our newspaper, we identify four classes of the Westminster parliamentary constituency: National Liberals; International Liberals; Nostalgic Britons and rural concerns. Both domestic and international Liberal constituencies oppose Brexit, supporting petitions calling on the government not to trigger Article 50 and to hold a second referendum. Domestic liberals are also concerned with domestic issues (such as the taxation of sugary drinks) while international liberals care less about domestic issues, favoring petitions focused on international causes.
Constituencies categorized as nostalgic Britons and rural concerns are far more conservative. Nostalgic Britons support petitions that reflect a bygone era (e.g. increased parental voice on school attendance) while Rural Concerns are anti-EU and support petitions relating to rural life (e.g. strong support for both for the prohibition and for the maintenance of capercaillie hunting).
We find our four classes fit well into the 2017 General Election voting patterns. The Conservative Party has the most support in the Rural Concerns class with little support in the International Liberals class. The Labor Party is strongly represented in both the international liberal and nostalgic British classes.
Our article demonstrates that there is significant information in responses to e-petitions and that this can have an impact that shapes perceptions of the political debate in every constituency. We argue that this data provides a picture of political sentiment across the country, eliminating the need to rely on outdated decennial census data, geographically scarce and incomplete opinion poll data, or sporadic household survey data.
Read the full study Here.
Stephen Clark has worked in the public, private and academic sectors, in fields as diverse as software engineering, computer aided design, transport planning and health. He currently works as a researcher at the Leeds Institute for Data Analytics and his research interests include spatial modeling of transport, public health planning and understanding election outcomes.
Nik Lomax is Academic Fellow in Spatial Data Analytics at the University of Leeds. His research interests include the estimation and projection of populations and their attributes, including ethnicity, housing, mobility and health.
Michelle A Morris is Academic Fellow in Health Data Analytics at the University of Leeds. Her research interests include how new forms of data and analytical methods can be used to inform health and policy.
