From the TUC

Polling nerdvana

17 Sep 2012, by in Society & Welfare

The British Social Attitudes Survey is polling as it should be done. It is annual, giving a chance to track opinion, and the questions are designed to find out what people think, rather than worded to try and get the highest possible level of support for the views of those commissioning them. 

Their latest results are out today on this rather well designed website and, even better, the Guardian datablog has put many of the results into time series which you can download here. It’s polling nerdvana.

The great danger with such wide-ranging surveys is that people look for the results that confirm what they already think – what psychologists call confirmation bias. We can see that in coverage today, with some highlighting support for higher public spending but others leading on ‘tough’ attitudes on welfare and immigration.

The other temptation is to try to construct a coherent world view and describe this as ‘what the public think’.

Politicians are sometimes prone to this when they talk about the centre-ground as if it produces a coherent set of policies and attitudes. (But that is another argument – and I wouldn’t argue that you can ignore public opinion either. Indeed that’s why I follow it rather closely.)

The truth is much more complicated – because people are much more complicated than these approaches suggest. Few of the interesting questions in the survey have such overwhelming majorities that there are not a significant number of dissenters. We do not know whether the people on the majority side of questions tend to be the same people, or whether they are more randomly distributed between winning or losing arguments.

And as anyone who has sat through a focus group knows, most people do not have a coherent set of beliefs about the world and every policy issue. Most people don’t find the need to think deeply about issues that do not affect them directly and often have very imperfect knowledge. If they did follow issues in detail the precise wording of poll questions woud not matter so much.

Trying to impose a pattern on lots of messy data is always going to be an inexact science, especially if you are not cherry-picking the results that suit.

One helpful concept is George Lakoff‘s theory of framing. He argues that people usually do not make decisions or strike attitudes purely through rational choices but by a more slippery process based on emotion, metaphor and language. Here’s a piece from the New York Times which explains this at a journalistic level. Lakoff argues that much of politics consists of competition between liberal and conservative frames (using these words in the US context, not as Clegg v Cameron).

Almost everyone uses a mix of these, although most people tend to use one more than the other. People in the centre are those therefore with the least fixed framing, rather than those who make the most rational choices.

I have observed people in focus-groups switch frames. One minute they echo conservative scrounger rhetoric, the next complain how tough it is for an unemployed relative to live on their benefits, but with the former completely influencing their political views.

I am not sure all of Lakoff’s work can be easily transferred to practical campaigning and some of it is too rooted in the US experience to easily transfer to the UK. (This free pdf ebook is a great introduction to his thought, though I find Drew Westen’s not dissimilar work a little more practical. )

But framing is still a very useful way to think about the British Social Attitudes Survey.  I read its results therefore as a competition between progressive and conservative frames for the issues it covers.

So we can see that conservatives have very successfully framed the welfare debate as about scroungers and abuse.

This chart tracks those who think unemployment benefits are too high less than those who think they are too low. (I’ve extrapolated data for two years when the question was not asked.) And don’t forget the real value of unemployment benefit has fallen considerably over the period covered by this graph.


The next graph (which comes from the BSAS website)  however shows how unsuccessful conservatives have been in the UK in arguing for a smaller state. The argument is between those arguing for more spending and those who say keep it the same – and there has been a slight shift in the progressive direction in the last year.


  F_Welfare _1.1

 I am not sure there are any flip conclusions from all of this for progresive campaigners. Shifting views on welfare is hard. Prejudice against claimants runs deep and is emotional. There are similar views on immigration (though wrong to see as racism, other than for a minority).

But while the government wants to frame all economic questions through the lens of deficit reduction, they are not succeeding. Otherwise the Chancellor’s personal poll rating and the government’s economic competence measures  would  be higher (though BSAS don’t track such issues as they have less long term relevance than the more fundamental attitudes they monitor.)

Yet thinking about these attitudes in terms of frames can be helpful for campaigners. The challenge is always to get more people to view more issues through the progressive frame – and a good test for any action or argument is whether it helps achieve this.

UPDATE: Here’s a piece by Lakoff about the UK on the Progress website.



2 Responses to Polling nerdvana

  1. Common Threads – September | Common Cause
    Sep 25th 2012, 11:26 am

    […] Polling nerdvana – Good analysis of the most recent British Social Attitudes Survey, and thoughts on how it represents a competition between progressive versus conservative framing of social issues. […]