- Specialist Division:
- Client:Sunday Times
- Universe:Adults resident in GB
- Methodology:Online Sample
- Sample size:5481
- Fieldwork dates:15 June 2017 - 20 June 2017
Weighting: All the results were weighted by age, sex, region and SEG and voters were also weighted to match the 2016 EU referendum and 2017 GB election results. Age groups were weighted to match the BES voter profile (see below for details of the very limited impact of this weighting) and locations were weighted to match the 2015 turnout of voters based on stats from the Electoral Commission. Reported Voting Intention figures include only those who say they are likely to vote (8-10). NB if we had used the same method as in our final preelection poll (including only 10/10), the Labour lead now would be just over 3% rather than 5%. Conversely, if we had used purely nat rep quotas rather than turnout adjusted, the Labour lead would be just over 5.5%.
As these results are also weighted to the recent election itself, the impact of our age weighting is quite small.The method we have used for this poll would have given us more accurate figures than we achieved in the General Election – Con and Lib Dem votes to within 0.2% and Lab off by just over 2.5%. We’ll review this approach when more detailed information is available about 2017 turnout. As an interim measure between now and then we may revert to a straight nat rep approach – which would have given us a better reflection of the gap between Con & Lab in the election, but at the expense of understating the Con share (by almost 2%), a recurring problem which turnout weighting was intended to – and did (at least apparently) – help to address.