Theresa May's commitment to Brexit may stand somewhat at odds with the source of her popular support, according to our first look at the Kieskompas--VOTEADVICE post-referendum study wave, dated 28 June–10 July.
The plot below shows her support is highest among those reporting regret about their Leave vote. Our figures are based on a subset of 519 respondents who reported having voted Conservative at the 2015 General Elections, out of a stratified-weighted sample of 2,251 respondents. Our model is predictive of May's support on a 10-point scale, as a function of respondents' referendum vote (discarding participants who did not vote), as well as whether they regret their vote choice.
Details on methodology
Sample strata reflect the UK's 12 regions ('NUTS1' including Scotland, Wales, and Northern Ireland), as well as the demographic categories of age (above 18) and gender. We calculated post-stratification weights to adjust the observed proportions to relevant statistics reported by the Office of National Statistics 2015 mid-year estimates. Online, opt-in samples such ours (participants signed up through our 'Election Compass' tools here and here) attract disproportionately many young and politically interested users. Thus we made further adjustments to account for other sources of selection bias: via Iterative Proportional Fitting, we adjusted for the distribution of political interest UK-wide as estimated by the British Election Study, for labour force status as reported by the ONS 2016 Labour Market Study, and for vote in the 2016 Brexit referendum, including turnout.
Theresa May's approval is estimated to be between 4.5–4.76 among the general voting age population, measured on a 10-point scale. The plot above is based on a linear model focusing on the subset of Conservative voters as of GE2015, and predicts approval (β0 = 4.09, t(489) = 26.28, p < .001.) as a function of the Brexit vote (β = 4.35, t(489) = 5.45, p < .001.), regret over the vote (β = 2.56, t(489) = 8.99, p < .001), and their interaction (β = -4.52, t(489) = -4.77, p < .001).
We have also looked at the sources of her support in the general population. We have found no substantive predictors of her approval other than partisanship, Conservatives being the most likely to support her while other party supporters are more prone to dislike her. We have found no support of a gender gap in her approval either.
Watch out for more updates as we are aiming to come back with further and more detailed reports on 'Brexit' political behaviour, as well as more information about our samples and data quality.