Popular Posts

Caveat Emptor

The opinions expressed on this page are mine alone. Any similarities to the views of my employer are completely coincidental.

Friday, 29 July 2016


It's Friday and time for more Chet the musician 'bittersweet' was coined for. Looking forward to seeing the biopic Born to be Blue  that was released in the UK earlier this week. Chet is another example of the rule that a great artist is allowed to be an arschloch. The same rule does not apply to academics. Tenderly.

Tuesday, 12 July 2016

The Social Structure of Brexit 3

Last time around on this theme. It seems to me that  a fair amount of silliness has emerged in the discussion of what drove the Brexit vote. Two moves seem popular, neither of which are terribly enlightening. One is to condition on the outcome and then ask what are the people who voted Out like? There is nothing inherently wrong with this, it answers a perfectly well defined question, just not the scientific question of interest. The problem is that p(c_1, c_2, c_3...| y_1) where c_1 etc is a characteristic - class, gender, vote preference -  and y_1 is Brexit vote is influenced by the marginal distribution of the characteristics. What you actually want is  p(y_1| c1, c2,c_3...).

The second move I've observed is to argue that Brexit voting isn't really a matter of social structure at all, but has to do with holding certain attitudes.  As I showed in my first post it is certainly true that holding certain attitudinal views - on immigration, political inclusion and welfare is strongly correlated with views about the EU. However it is quite another thing to argue that one causes the other. Certainly it is not wise to condition on an endogenous attitudinal variable and then claim that this shows that social structure doesn't matter. You need to consider both indirect and direct effects and it has something of the flavour  of saying - to steal the words of a very perceptive sociological colleague - "God did it" (hat tip to BTH). Sometimes, as I say to my students, you don't want an explanatory variable that is too close to the dependent variable.

So this brings me around to my last bit of empirics on the subject. Again I'm using the 2015 BES cross-section with the same measure of attitude to the EU the construction of which I described in my previous post, but simplified so that I only consider the contrast between those that are supportive of the EU and those that are anti EU (I drop the ambivalent). I estimate a linear model for the probability of being anti EU with fixed effects for the 300 odd constituencies in the sample. I control for age and gender. I'm interested in: 1) coefficient for socio-economic characteristics of the respondents ( occupation, household-income and whether or not they are a graduate) mainly so we can get some idea of the magnitude of the differences between groups; 2) getting some idea of the predictive impact of these characteristics. 

In case the binary regression police are watching I should say that if you estimate a conditional logit model with fixed effects you  will draw exactly the same substantive conclusions (but feel slightly better about your standard errors).

So, if you estimate this regression with just the fixed effect and a constant you "explain" about 20% of the variance. I don't recommend this as a way to estimate a variance component - but I need some sort of baseline against which to evaluate the contribution of the individual level predictors. Next throw in the demographics -age and gender. That "explains" 21%. Add in the socio-economic predictors - occupation, income and graduate status - and this leaps up to 35%.  However you look at it adding more socio-economic structure improves the prediction by a substantial amount.

Let's look at the estimated slope coefficients and associated t values. These are shown in the columns headed  (1) in the table below (the fixed effects and demographic controls are in the model but I don't show them here).
Controlling for the constituency fixed effects and the demographics, occupation, income and being a graduate all capture substantively important differences in the propensity to be anti EU.  Being self-employed or working class makes you more anti, as does having a low household income, and being a non-graduate.

 Consider the magnitude of the estimated  %  point difference in taking an anti-EU stance for two typical cases. The conditional average difference between a graduate higher manager with a household income in the £40,000-£74,999 bracket and a working class, non graduate with household income in the £15,600-£39,000 bracket: it's 39 percentage points. Assuming that these data have some bearing on people's behaviour in June 2016, then anyone who wants to argue that something like social class is a minor factor has some explaining to do.

But let's test the case for individual level socio-economic differences a bit more. Does it stand up even if we condition on an obviously endogenous variable like party voted for in the 2015 General Election? The columns labelled (2) hold the answer. And it is, broadly speaking, yes. Controlling for political preference (and in fact participation) doesn't really change anything very much.  We get a similar outcome if we condition on the, again obviously endogenous, attitudinal variables  - towards immigration, politics and welfare - that I discussed in my last two posts. The socio-economic coefficients shrink slightly towards zero, but they don't go away.

It may be the case that something like class dealignment has happened with regard to party choice, but that does not imply that "class" per se is dead. 

"As surely as Water will wet us, as surely as Fire will burn,
The Gods of the Copybook Headings with terror and slaughter return!"


Friday, 8 July 2016

The Social Structure of Brexit 2

 I'm surprised by how much interest (judged by page views) there has been in my last post. Especially since I didn't put much effort into explanation or indeed  interpretation of the graphs. And  I didn't devote any effort at all to answering the $64,000 question which is why  we went from a situation in Spring/Summer 2015 in which roughly 40% of the electorate was decidedly pro EU and only 30% decidedly anti to in June 2016 52% of voters opting for Leave?

 As I said before, the best evidence on this is going to come from the ESRC's EU Referendum Study which among other things will, I assume, be able to tell us something about  change and crystallization of opinion at the individual level in the months leading up to the vote. So you should look out for the stuff coming out from those guys.

What I thought it might be useful to do though is return to the 2015 BES cross-section and explain a little about what went into the measures of  support for the EU, concern about immigration, alienation from politics and scepticism about welfare that I used last time round. 

To get anywhere you have to simplify, so this is my simplification. BES asks the following questions:

1) Overall, do you approve or disapprove of Britain's membership in the European Union?
2) If there was referendum on Britain's membership of the European Union, how do you think you would vote? Would you vote to leave the EU or to stay in?
3) Which of these comes closet to your own views? Britain should: - respondent should choose a position from an 11 point scale anchored at one end by the statement ' Do all it can to unite fully with the European Union' and at the other by 'Do all it can to protect its independence from the European Union'.

4) Do you think that too many immigrants have been let into the country, or not? How strongly do you feel about this?
5) Do you think immigration is good or bad for Britain's economy?
6) As far as your concerned what is the single most important issue facing the country at the present time? An open ended question from which I pick out all those who mention immigration. The latter are 35% of those that mention an issue. Immigration roughly 3 times more likely to be picked out than any other single issue.

I'd like to read a few statements about public life. Using an answer from this card, please tell me how strongly you agree or disagree with each of them.
7) Politicians don't care what people like me think.
8) People like me have no say in what government does.
9) Politicians ignore the issues I really care about.

How much do you agree or disagree with the following statements?
10) If welfare benefits weren't so generous people would learn to stand on their own two feet.
11) The welfare state encourages people to stop helping each other.
12) Many people who get social security don't really deserve any help.

This is my raw material. In most cases I chunk up the answers into three groups, putting, where it makes sense, the don't knows in the same groups as the neither agree nor disagrees. Measure 3) (in the EU group) naturally only has two categories - you either mention or don't mention immigration. Measure 4) in the immigration group has four categories since I distinguish between those who strongly and less strongly believe that there has been too much immigration.

If you assume that the answers people give to these questions reveal their latent underlying attitudinal disposition then you can construct a statistical model for the responses that recovers the underlying attitudinal groups that people belong to. So I estimate a latent class model for each of the four sets of items. In each case I recover 3 groups:

1) Pro EU; ambivalent towards the EU; anti EU.
2) Positive about immigration; ambivalent about immigration; negative and strongly concerned about immigration.
3) Believing that politics and politicians reflect their views; ambivalent about politics and politicians; alienated from politics and politicians;
4) Supportive of welfare; ambivalent about welfare; negative about welfare and the people that receive it.

What happens when you look at the world in this way is illustrated below (click to enlarge).

A number of things strike me about these numbers. Firstly, the 10 percentage point difference between those who were positive & those who  were negative about the EU is nothing compared to the potential for swaying the almost one-third who were ambivalent.  There was lots of potential for the campaign to draw voters to one side or the other.

But now look at the scale of the persuasion job the Remain side had to do. Forty-five percent of the electorate strongly believed that immigration was a major problem and a further third were ambivalent which actually meant they thought it was a concern but expressed their views a little  less strongly than the diehards. More than half of the electorate felt alienated from politics and politicians, meaning, among other things, that they felt ignored and that their views  (which presumably included their views about immigration) were not of concern to the political elite. Finally just under a half had negative views about welfare and somehow felt that it wasn't going to the right people.

As my last post showed, concerns about immigration, alienation from politics and negative views about welfare are predictive of taking a negative view about the EU. In fact they are  all part of  the same ideological syndrome. It's clear why the Leavers campaigned the way they did. Already in 2015 it should have been obvious that a populist, anti-immigration, anti-elite, pro" hard-working families" campaign would be very appealing. It didn't take great political skill to link all these things together. The audience already believed it. All that was needed were the populists who were willing to exploit it and a Remain campaign that confirmed its elite and 'out of touch' image by trying to avoid mentioning those particular inconvenient facts.  The Leavers also, of course, ignored many other sorts of inconvenient facts, but these were not the ones the punters cared about. Ideologically they had an incumbent advantage.

Wednesday, 6 July 2016

The Social Structure of Brexit

The high politics of the fall-out from Brexit are so fascinating that it is difficult to take one's eyes off the rapidly evolving tragedy/farce. Even so, just to get things clear in my own mind, I took a little time out to look at some data. My idea is nothing more profound than to illustrate what my old LSE tutor Eileen Barker used to call "what goes with what".  I suppose the definitive source for this sort of thing will be the ESRC referendum study that is being directed by my colleague Geoff Evans. We also already know quite a bit from the polls  and from the aggregate level picture. What I'm going to ask here is what can we have learn from the 2015 British Election Survey cross-section? The BES data were collected between May and September 2015, so long before campaigning began in earnest, but my guess is that for our purposes this doesn't matter too much.

One caveat before I go further: this is a blog post not an academic article so I'm going to cut to the chase rather quickly without going into all the details about how I've processed the data.

The first thing I want to look at is support for the EU and how it relates to other issues and dispositions that are likely to be of  relevance to how people voted on June 23rd. I look simultaneously at four things: 1) support for the EU; 2) immigration; 3) feeling alienated from politics in the sense that you don't believe politics reflects your concerns; 4) believing that welfare benefits are a bad thing or are abused. I take 3 items on each of these things from the BES and estimate four latent class models. In each case the model fits very well and divides the population into three groups: 1) those who are positive about the EU/immigration/feel politics reflects their concerns/think welfare is a good thing; 2) those who are ambivalent or don't know what they think about these issues; 3) those who are negative about the EU/ concerned about immigration/feel alienated from politics/dislike welfare.

If you put all four of these latent classifications into a multiple correspondence analyses (MCA) you get a pretty clear picture of what goes with what (click to enlarge).
People who don't like the EU (-) also don't like immigration (-), feel politically alienated (yes) and think that something has gone wrong with welfare (-). Conversely if you like the EU (+) you don't mind immigration (+) , don't feel politically ignored (no) and feel positive about welfare (+). Nothing particularly surprising here, apart from perhaps quite how closely the four domains map onto each other.

Next step. How do these attitudes relate to party politics? We can add to the MCA plot information about how people voted in the 2015 election.

Again, no particular surprises. UKippers are very decided in their views, Labour, Nationalist and other voters are strongly for the EU and the other things that go with it. LibDems are also pro EU but are also pulled towards ambivalence. The Conservatives and those that didn't vote tend towards an anti EU position but also feel the gravitational pull of ambivalence.

OK, what about the social structural stuff then? Let's start with social class. Yup, that seems to work.

The routine and semi-routine employees (labelled here "working class") and the self-employed dislike the EU, immigration etc. What about if we do class in a different way by looking at household income?
 That's pretty clear too. The rich like the EU, the poor don't. And age?

Pretty clear relationship there as well.  I could on. I've got similar plots for graduates versus non graduates, students versus non students and of course for geographical regions.

If you do something fancy like estimating a multinomial logit with the latent pro, ambivalent and anti EU groups as the response variable, all these relationships stand up as partial "effects" when you include all the others as predictors. It's probably worth emphasizing though that it is difficult to know what sense to give to the coefficients from that sort of exercise. What you really want to know is the discriminating power of the predictors. An ROC plot would probably be revealing.

Punch line. Social structure is important, but then we knew that anyway didn't we?

Friday, 1 July 2016


It's Friday afternoon, everything's gone to Hell in a handcart, but at least we still have Chet Baker.

Tears before bedtime

Has anyone  noticed that our politicians seem to be rather lachrymose these days? When exactly did the trembling lower rather than the stiff upper lip become a public relations asset? Or are they all really regressing back to the emotional world of the nursery?