Religiosity and Party Choice

It’s interesting to read Justin Parkinson, Political Reporter at the BBC, on Christianity and the election.

Clive Field has already posted extensively earlier this week here and here on the ComRes/CPanel survey of Christians’ political attitudes, conducted 30 March-12 April. The think-tank Theos also commissioned a poll of religion and politics on 17 and 18 February, which Clive has covered here.

It’s noteworthy that in the European elections of 2009, the Christian Party – established five years earlier – won 250,000 votes, or 1.6 per cent nationally. While this was not a significant result at the national level, it was nevertheless ahead of Arthur Scargill’s Socialist Labour Party. The party also polled 2.9 per cent in London. Andrew Brown has set out some interesting thoughts on the Christian Party here.

My personal guess, though, is that strong Christians won’t swing the election this time, although this is almost pure intuitivism.

First, it is incredibly difficult to predict what will happen now – whether the Lib Dem surge is going to last, and how that will affect the outcome. As far as I’m aware, psephologists had not expected the leaders’ debates to matter much, beyond a brief bounce. If there is a great deal of switching, particularly from none to Lib Dem (and particularly by the under-25s), and turnout is affected, it’s difficult to predict both shares of votes and how they translate to seats.

Second, my guess is that there are very few Christians who vote as Christians, rather than in line with their political ideology which in turn is more likely to be driven by socio-economic status, as well as perception of the apparent fitness for government of different parties and party leaders.

Third, the ‘Christian vote’ is not an enormous constituency. A quick look at the British Social Attitudes survey for 2008 suggests that about 9% of the population are Christians who attend church weekly or more (and there is other evidence to suggest that people over-report their church attendance – David Voas alluded to this earlir this week in his post on Easter church attendance).

I did a basic analysis of party choice and religiosity for this year’s British Social Attitudes report. This suggested that after controlling for other socio-demographic variables, there was a weak but positive correlation between religiosity (strength of religiousness, rather than just religious identity) and likelihood of supporting Tory or Labour compared to no party (you have to have a base category so in this case it was ‘none’). For the Lib Dems, and other parties (Greens, Nats and others combined), there was no difference comapred with the ‘nones’.

The next question was whether this was causal, or whether the causality might run the other way (from political alignment to religiosity). I’m doing further work at the moment which suggests it is causal.

But even if there were such a relationship at the national level, what does this mean for this election? Expressing a party choice in the 2008 BSA survey was cheap talk, almost two years ago. Furthermore, where are the ‘strong Christians’ living? I can imagine that many older, traditional, staunch Anglicans (for example) live in safe seats, where even if their vote is partly determined by their religiosity, this might have negligible impact.

It’s also difficult to test. We could gather data on congregations at the constituency level, or compile many case studies. One possibility might be the Taking Part survey which provides neighbourhood-level measures of religious practice in the 2007/8 survey (though it doesn’t collect data on party choice). This could be used to calculate constituency-level measures of practice using http://geoconvert.mimas.ac.uk/ and incorporated into a constituency-level study.

Ideally we would have a very large survey including measures of religious affiliation and strength of religiosity, with enough detail to look at how these correlate with party choice at the constituency level – or at least the marginal constituencies.

An interesting insight is the following work by Ed Fieldhouse and Dave Cutts on electoral turnout (rather than party choice) by South Asian communities in 2001, which found differences between Hindus, Sikhs and Muslims.

There is undoubtedly more work out there which looks at religion (particularly religious affiliation rather than religiosity) and voting at ward level – so please send links because I have much to learn here. In some areas parties may sponsor candidates who they think can help deliver part of a community vote. Other cases include the Respect alliance which in some areas has appealed to a Muslim anti-war vote. Ingrid Storm’s PhD research suggests a link between ‘ethnic Christianity’ and attitudes to immigration. If true, this may partly account for the increasing use of ‘Christian Britain’-type language of the BNP.

If you have more to add on Christianity and the forthcoming election – or religiosity and politics writ large – please add your thoughts below.