Polling and forecasting the general election of 2010

In the 2010 British general election the pollsters did well in correctly forecasting the Labour and Conservative vote shares, but they did much less well in predicting the Liberal Democrat share. The fact that the Liberal Democrat forecast alone was seriously out is reassuring in one sense because i...

Full description

Saved in:
Bibliographic Details
Published inInternational journal of market research Vol. 52; no. 5; p. 687
Main Authors Whiteley, Paul, Sanders, David
Format Journal Article
LanguageEnglish
Published Henley-on-Thames Market Research Society 01.01.2010
Subjects
Accuracy
Conflict
Forecasting
Measurement techniques
Parliamentary elections-UK
Public opinion surveys
United Kingdom > UK
Online AccessGet full text

Cover

More Information
Summary:In the 2010 British general election the pollsters did well in correctly forecasting the Labour and Conservative vote shares, but they did much less well in predicting the Liberal Democrat share. The fact that the Liberal Democrat forecast alone was seriously out is reassuring in one sense because it means that methodological problems with the polls are unlikely to be the explanation for what happened. If there are methodological lessons to be learned from this for the wider polling industry they relate to attempts to measure the probability that an individual will turn out and vote before the election day. A number of agencies have reported data on voting intentions from respondents who say that they will definitely vote, and this can be used to adjust the data. There may very well be survey mode effects with a willingness by respondents to claim voting varying between face to face, telephone and Internet surveys. The challenge is to derive measures that overcome these problems.
ISSN:1470-7853
2515-2173