24 June 2011

Those Scottish Election Study headlines...

My thanks to the helpful Dr Christopher Carman for alerting me to the fact that some initial slides from the Scottish Election Study of 2011 have now been published.  Many of you will have seen the edition of BBC Newsnicht this week, which enjoyed an early sight of the Study's preliminary findings about May's Holyrood election. The research team of political scientists, based at the University of Strathclyde, describe their methodology thus:

"In 2011, the Scottish Election Study took the form of a two-wave internet panel survey, with data collection undertaken by YouGov. The aim of the study is to explain the decisions of Scottish voters on 5 May, both whether and for which party they voted. As well as voting behaviour, the survey questionnaires cover the following topics: attitudes to parties and leaders; issue opinions and evaluations; national identity; constitutional preferences; multilevel party identification; preferences for political compromise; socio-demographic characteristics. In addition to the pre/post-election panel, the 2011 SES also reinterviewed respondents to the 2007 Scottish Election Study."

Data collection took place in two waves, one before and one after election day. The first engaged with 2,046 respondents, with fieldwork being conducted in late April. The second wave occurred later in May, with 1,760 respondents. While analysis of the data is still ongoing, the research team have now made three sets of slides available online, laying out preliminary findings.

The first, from Dr Carman's presentation from a seminar at the University of Strathclyde held this week, looks at issues of turnout; the use of two ballot papers for constituency and regional votes in Holyrood 2011, and its comprehensibility; the abortive AV referendum and respondents understanding of the AV system - and finally, on the phenomenon of "multi-level party support" in Holyrood and Westminster, or to borrow a phrase from John McTernan, Scotland's "promiscuous", Bobbing John electorate.

Secondly, we have the University of Essex's Dr Rob Johns, who asks Why does ‘performance politics’ win Scottish elections? Johns submits that there are four key aspects of context: 1. The Scottish Parliament matters; 2. Class and party dealignment; 3. Ideological convergence and 4. Reshaping of the constitutional issue. The first limb focusses on how respondents envisage key policy areas, including law and order, health - and so on. Do they conceive of outcomes in these policy areas as being primarily due to the UK or the Scottish Government? Secondly, Johns looks at responses to the question “Do you usually think of yourself as being a supporter of one particular party?” While 44% of respondents said ‘no’, the study revealed that many said ‘yes’ then abstained or defected in the most recent Holyrood election. The third limb examines perceptions of "ideological convergence" between political parties. The fourth plucks out the issue of the constitution.“How do you think the return of an SNP minority government would affect the likelihood of independence?”, asked the research team. According to 1,784 respondents to the Study the return of an SNP minority administration would make independence...

  • Much more likely ~ 7% 
  • Bit more likely ~ 29%
  • No difference ~ 42% 
  • Bit less likely ~ 13% 
  • Much less likely ~ 9%
Johns then turns to look at credit and blame, party image - capable of strong government, united, in touch with ordinary people, keeps promises - followed by a comparative analysis of leader-party popularities. All interesting and very much worth a look.

Lastly, at least for now, we have Professor James Mitchell's slides, which deal with a couple of issues of particular interest to me - support for party by gender, and by social grading. In the interests of comparative brevity, I'll tease out the Study's findings on these two issues - and simply note the other findings, without much getting into them.  Those of you who stayed with me during the Holyrood campaign may remember my series of posts, looking at the disaggregated data in YouGov's pre-election polling, focussing throughout on gender and class. One striking feature of this series of polls was the thumping leads the SNP were recording amongst C2DE voters - those assessed to be working class based on the occupational criteria. The polls also tended to show a narrowing "gender gap" in the SNP support, with increasing percentages of women, minded to support the Nationalists. My interest in that subject goes back some time, with my first dedicated post on the topic dating from August 2010, filling out some of the context informing a column by former Salmond special advisor Jennifer Dempsie in the Scotland on Sunday, arguing that "winning over female voters" was "crucial to SNP ambitions". Unfortunately, the wider media didn't really pick up on these interesting trends in the course of the campaign.  If they had done so, the Study's conclusions would be much less surprising, extensively anticipated as they were.

Significantly, Professor Mitchell's study confirms YouGov's pre-election polling which suggested a narrowing gender gap and a significant lead amongst working class voters. According to the survey data collected by the Study, their respondents voted as follows, by gender (N.B. Mitchell is referring us to the Study data on regional voting in the 2011 Holyrood election).

Male respondents (SES)...
  • SNP ~ 46%
  • Labour ~ 24%
  • Tories ~ 12%
  • Liberals ~ 4%
  • Greens ~ 6%
  • Others ~ 8%

And women...

Female respondents (SES)...
  • SNP ~ 43%
  • Labour ~ 29%
  • Tories ~ 12%
  • Liberals ~ 6%
  • Greens ~ 4%
  • Others ~ 6%

On social class, the Study asked about (a) subjective social class, namely how respondents self-identify and (b) objective social class, based on the familiar ABC1/C2DE categorisations I've discussed previously.  Firstly, the Scottish Election Study's working class data, subjectively then objectively defined:

Respondents subjectively identifying as working class...
  • SNP ~ 47%
  • Labour ~ 33%
  • Tories ~ 7%
  • Liberals ~ 4%
  • Greens ~ 3%
  • Others ~ 6%
C2DE voters (working classes)...
  • SNP ~ 47%
  • Labour ~ 28%
  • Tories ~ 9%
  • Liberals ~ 4%
  • Greens ~ 4%
  • Others ~ 8%

In brief, amongst those subjectively identifying as working class in the Election Survey, the SNP beat Labour by 14%. According to social grading's objective criteria, the gap was even wider - with Labour lagging 19% behind the SNP.  Interestingly, this outcome echoes (and amplifies) the results of YouGov's pre-election polling, which recorded an SNP lead over Labour amongst C2DE voters of a magnitude varying from 4% to 15% in the constituency vote and -1% to 10% in the regional vote. Interestingly at the beginning of the campaign, the polls suggested that ABC1 voters remained to be convinced by the Nationalists, holding on the Labour allegiances more tenaciously than their C2DE fellow citizens. According to Mitchell's data, bourgeois participants in the Survey reported the following voting behaviour, with the same subjective then objective analysis...

Respondents subjectively identifying as middle class...
  • SNP ~ 37%
  • Labour ~ 16%
  • Tories ~ 22%
  • Liberals ~ 8%
  • Greens ~ 9%
  • Others ~ 8%

And according to social grading's objective criteria...

AB voters (upper middle & middle classes)...
  • SNP ~ 41%
  • Labour ~ 25%
  • Tories ~ 14%
  • Liberals ~ 5%
  • Greens ~ 8%
  • Others ~ 7%

And...

C1 voters (lower middle classes)
  • SNP ~ 41%
  • Labour ~ 25%
  • Tories ~ 17%
  • Liberals ~ 8%
  • Greens ~ 6%
  • Others ~ 3%

The rather bizarre looking results of the subjective social class findings - 22% of which is Tory - can probably be explained by factoring in a third option given to respondents - to subjectively identify with no social class at all.  Amongst those respondents, the results were striking - and as follows.

Respondents subjectively identifying as having no class...
  • SNP ~ 53%
  • Labour ~ 17%
  • Tories ~ 14%
  • Liberals ~ 4%
  • Greens ~ 4%
  • Others ~ 8%

Professor Mitchell's slides also lays out the voting data according to national identity - with an interesting and sensitive range of options being afforded to those questioned, allowing those answering to give priority to British or Scottish identities, declare an equivalence between them, or to deny either. The Professor also sets down voting by religious affiliation, albeit with a fairly limited range of categories. Three in fact: no religion; Church of Scotland or Catholic. The data on these issues can be found on his ninth slide. No doubt the Study will generate plenty of other eccentric pieces of data to keep amateur psephologists in the public cheered and distracted from their other labours, for some time to come.

6 comments :

  1. Scottish republic24 June 2011 22:24

    A rather different picture from Scottish voting patterns in the 1960s when I was a wee boy.

    5 years of an excellent SNP administration might well solidify these new trends.
    5 years of a poor administration would be unhelpful.

    As long as the SNP keeps smart then things should work if the Brit nats don't deliberately try and screw up the SNP government.

    ReplyDelete
  2. How do YouGov select their interviewees?

    ReplyDelete
  3. Scottish republic25 June 2011 12:36

    It's a mystery but it all seems to work out in the end.

    Though it's true, they got it wrong just prior to the SNP election victory.

    Let me add, it's a political victory for Scotland but it won't be an actual victory until we have independence. None of us will rest, ever till that goal is achieved.

    ReplyDelete
  4. Scottish Republic,

    Certainly, it is tempting to conclude that the polls conducted in the last days of the Holyrood campaign didn't capture the underlying movement across the country. One might try a manful defence, however, and argue that the polls had indeed narrowed in the final days and things conclusively shifted in the SNP's favour in the last days of the campaign. I wouldn't necessarily want to defend the argument - but one can at least make it.

    Anonymous,

    YouGov explain their methodology here.

    ReplyDelete
  5. LPW

    Thank you for the link.

    I've always been suspicious of the YouGov methodology which involves people putting themselves forward to be on the panel of interviewees.

    It is not beyond the wit of man to distort the balance of the panel by being, shall we say, disingenous when completing the 'detailed profiling questionnaire.'

    Obviously I'm not saying that SNP supporters would do such a thing, any more than they would bombard newspaper discussion forums, as some sort of HVP strategy. But they could, if they were dedicated enough.

    ReplyDelete
  6. Anonymous,

    It would seem like pretty excessive lengths to go to - just for the vague possibility of being contacted and polled. Especially when you consider that respondents have basically no control over how the data is weighted by YouGov, before publication. I also wonder what the nefarious cybernat you envisage might say they were in order to sneak in - to moonlight as an old Tory lady from Orkney perhaps?

    ReplyDelete