Did #Austerity Cause #Brexit? I have received some comments on the paper that have engaged with the data, but am more than happy to stand my ground. Here is why...
Recently, Noah Carl @NuffieldCollege studied national level opinion poll time series, arguing that the lack of evidence of an uptick in post 2010 support for Leave is evidence suggesting that austerity had nothing to do with #Brexit. See his comment here osf.io/zbgt3/
So what are the problems with this approach.
1) Opinion polls do not rely on the same sampling population, so a lot of time variation that is so crucial given the tight referendum results are well within the margins of error
1) Opinion polls do not rely on the same sampling population, so a lot of time variation that is so crucial given the tight referendum results are well within the margins of error
2) Inconsistent questions - the opinion polls do not always ask exactly the same question (the Eurobarometer being the exception), can easily distort things -- questions are full of priming etc.
3) Don't knows that are typically excluded make up a significant share of the respondent population ( ~easily within 10-20%), these are non negligible magnitudes and this picture suggests its those "dont knows" who drove support for Leave in 2016 
4) There is evidence suggesting that opinion polling data is distorted and the method of data collection matters a lot, there is a great paper on demand effects by my coauthor @jondequidt and colleague @cp_roth aeaweb.org/articles?id=10…
5) Polling data masks composition effects and a look into what type of individuals have changed their EU preferences is important as austerity and the social policies after 2010 had very much distributional effects affecting some groups in society more or less. In fact...
6) If you look at the actual Eurobarometer micro data available for the UK goo.gl/VpRwFg , you see that people with low education working in manual jobs (i.e. those likely to be on benefits and exposed to cuts) started to adopt anti-EU stances after 2010...
