A seal of clap-proval
At the Centre for the Analysis of Social Media, I study how open source intelligence can help understand society. New data from social media in particular can provide useful real-time insight into public attitudes toward events as they happen. But there is another source I’ve been overlooking all along: Question Time.
As everyone knows, this weekly programme asks a panel about the topical issues of the day; and applause is dispensed by means of approval. So can we gauge public opinion on current events based on the clap-proval each answer gets?
First, a quick statistics methods primer. Decent statistical polling works by taking a randomised, representative sample of the population. Question Time audiences are small - only around 150 people - but according to the BBC it is ‘a balanced cross-section of people’ from the city where the programme is filmed. From the application form this seems to be based on age, ethnic background and political preference. But there are also a number of confounding variables.
First, how far is clap volume an accurate and consistent gauge of attitudes? Panellists have a tendency to trail off, and Dimbleby frequently cuts-in. Audiences don’t just clap based on rational agreement; they respond to good delivery or acerbism, and some are better than others. There is also systemic bias: non-political commentators are always more popular than politicians, and thus likely to score a clap whatever they say. Any serving Minister is automatically penalised.
So to gauge public opinion I’ve factored in these various sampling and skewed data problems with some post-stratification weights, and converted each clap volume onto a ten-point Likert scale score, where 0 means ‘do not agree’ and 10 ‘strongly agree’. I assume a volume increase denotes more people clapping rather than the same number of people clapping more loudly, so a sound analogous to half the audience clapping would give a score 5. We cannot measure strength of disagreement, because we cannot measure volume of silence.
So with this in mind, what do we know about public attitudes this week?
Overall, Lancastrians are not too bothered about HS2, and are more concerned about lowering train fares overall (+6); or even better, nationalising them (+7).
Really what the North needs is more investment: (+6) and for southerners to stop saying the North is Manchester (+8).
Attitudes to Mali seem unclear, perhaps a mild displeasure toward UK involvement (+3), and a preference to trade over military intervention (+2).
Lancastrians are split about Charles becoming King, and about Nick Clegg possibly sending his children to private school: although politicians sending their children to private school while extolling the virtue of public education is not a vote-winner (+8).
And finally, the benefits of immigration – at least the economic benefits – are quite widely recognised and accepted (+6).