The opinion polls undertaken prior to the 2015 UK General Election under-estimated the Conservative lead over Labour by an average of 7 percentage points. This collective failure led politicians and commentators to question the validity and utility of political polling and raised concerns regarding a broader public loss of confidence in survey research. In this paper, we assess the likely causes of the 2015 polling errors. We begin by setting out a formal account of the statistical methodology and assumptions required for valid estimation of party vote shares using quota sampling. We then describe the current approach of polling organisations for estimating sampling variability and suggest a new method based on bootstrap re-sampling. Next, we use poll micro-data to assess the plausibility of different explanations of the polling errors. Our conclusion is that the primary cause of the polling errors in 2015 was unrepresentative sampling.
|Royal Statistical Society. Journal. Series A: Statistics in Society
|Early online date
|25 Oct 2017
|Published - 2017
Research Beacons, Institutes and Platforms
- Cathie Marsh Institute