Voting intentions In Scotland 2005-2006

Since the 2005 General Election, there has been a dearth of information published on voting intentions in Scotland, with only one published poll in April 2006. This article details the pattern of stated voting intentions in Scotland measured in Ipsos's Social Policy Monitor survey in 2005 and the first half of 2006. This survey is a face-to-face in-home 'omnibus' survey that uses random pre-selected sampling rather than quota sampling, and is therefore a unique source of polling data for Scotland.

Table 1 shows the voting intentions for the Scottish constituency vote in the five waves of the SPM conducted between the beginning of 2005 to June 2006. Table 2 shows the voting intentions for the list vote. Over this period, four strong themes emerge from the pattern of voting intentions for Holyrood.

Labour loses ground. The figures show a loss of support for Labour with regard to Holyrood and Westminster voting intentions since their peak around the General Election of 2005. This loss of support was gradual during the latter half of 2005 and more dramatic in 2006.

In the constituency voting intentions, Labour support peaked at 40% in the Apr-Jun 2005 wave and the Aug-Sept 05 wave, before slipping back to 38% in the first quarter of 2006. In the list voting intentions a similar pattern is seen, with Labour peaking on 37% in Apr-June 05, before slipping back to 33% in the first quarter of 2006, a position very similar to where they were in the first quarter of 2005.

However, between the first and second quarters of 2006, Labour suffered a dramatic fall in support for the constituency vote (from 38% to 28%) and for the list vote (from 33% to 26%).

The Conservatives and Liberal Democrats are flat lining. Support for the Conservatives and the Liberal Democrats has been relatively stable over this period, with the elections of Cameron and Campbell (and Goldie and Stephen), and the loss of support for Labour having little effect in Scotland.

In the first quarter of 2005, the Conservatives were standing on 16% for the constituency voting intention measure, and 14% for the list vote. Apart from the slight dip that they suffered around the time of the General Election, their share of the vote has always been within a couple of percentage points of these levels since then. Data from the most recent wave of the SPM puts them on 15% for the Constituency vote and 16% for the List vote.

Although the amount of fluctuation in the level of Liberal Democrat support has been slightly greater than that of the Conservatives, their level of support has still been relatively stable. Having stood on 21% for the Constituency vote, and 18% of the List vote in the first quarter of 2005, they now stand on 19% for both the Constituency vote and the List vote.

Recent surge for the SNP. In contrast to the Conservatives and the Liberal Democrats, the SNP enjoyed a small increase in support during 2005, and a sizeable increase in support in the second quarter of 2006.

In the first quarter of 2005, the SNP stood in third place behind Labour and the Liberal Democrats with regards to the constituency vote (20%) and second behind Labour on the list vote (24%). By the first quarter of 2005, they had moved to second place on the Constituency vote (24%) and maintained a relatively stable position with regards the List vote (23%).

However, in the most recent survey, conducted in the second quarter of 2006, their share of the Constituency vote increased 6% to 30%, and their share of the List vote increased 5% to 28%. This boost means that the SNP now narrowly leads from Labour on both of these measures.

Fluctuation in the proportion of undecided electorate. There has been a decrease in the number of people who were undecided on who to vote for during 2005, but a significant rise in the second quarter of 2006.

The proportion of those would said they were certain to vote, but undecided as to who they would vote for in the constituency ballot, dropped steadily from 11% in the first quarter of 2005 to 6% in the first quarter of 2006, but rose sharply again in the second quarter of 2006. A similar pattern is seen with the list vote, with a steady drop from 14% in the first quarter of 2005 to 7% in the first quarter of 2006, followed by a rise to 12% in the second quarter of 2006.

Table 1: Scottish constituency voting intentions (aged 18+, certain to vote)

Table 2: Scottish list voting intentions (aged 18+, certain to vote)

Similar patterns are seen in the figures for Westminster voting intentions in Scotland. (Table 3). Support for Labour in the first quarter of 2006 was at a similar level to their position in the first quarter of 2005. The Conservatives level of support has seen a decline since the beginning of 2005. Support for the Liberal Democrats has been reasonably steady over the period. Meanwhile, the SNP slowly increased in support throughout 2005. Again, sizeable changes are seen in the second quarter of 2006, with Labour dropping 8% to 36%, the SNP increasing 8%, and the proportion of people undecided almost doubling from 5% to 9%.

Table 3: Westminster voting intentions (aged 18+, certain to vote)

The survey details

While most voting intention polls are conducted using a telephone survey or face-to-face quota survey, or in the case of YouGov online surveys, Ipsos has been collecting voting intention data for Scotland on our quarterly random pre-selected omnibus survey. This survey, the Social Policy Monitor (SPM), is a face-to-face in-home 'omnibus' survey that uses random pre-selected sampling rather than quota sampling. It is therefore a new and unique source of polling data in Scotland.

Random pre-selected surveys tend to be used for large-scale government surveys, such as the Crime Survey, the House Condition Survey etc, in which prevalence measures are required, and by using statistical theory, it is possible to produce estimates of their level of accuracy, normally through the use of Confidence Intervals: e.g. "results accurate to within +/- 3% 95 times out of 100". Although empirical evidence indicates that quota samples -- when properly conducted -- can mimic random probability surveys, they cannot base their measures of 'margins of error' on the same solid grounding of statistical theory.

With face-to-face random pre-selected surveys, interviewers have to return to the same addresses time and again or order to try and interview a household member. This is time-consuming (and relatively costly). As quota sampling relies on getting a sample of interviews that match the target population on certain characteristics (e.g. sex, age) by filling quotas of these characteristics, interviewers can simply move on to the next address to try and find a suitable respondent. So with quota surveys, there is no requirement to visit addresses on several occasions to achieve an interview

While the fact that this is a random probability survey is important in terms of statistical theory, it is also important to note that the requirement for interviewers to make repeat visits to addresses means that the fieldwork period is substantially longer than for a 'standard' poll.

For each wave of the SPM, a two stage random sampling process was used in which 90-100 clusters of addresses were drawn from the Post Code address file. Around 1,750 randomly selected addresses are allocated to interviewers each wave, and around 1,000 interviews are completed each wave.

MORI's measurement of voting intention uses two questions for each measure. These are usually the first questions asked on the questionnaire with the randomly selected adult at each address. For the Scottish voting intention measures, the first question is:

Q In elections for the Scottish Parliament you have two votes. Your first vote is for a person to represent the constituency, like in elections for the Westminster Parliament. If there were elections to the Scottish Parliament tomorrow, how would you use your first vote?

Our interviewers record responses to this question as one of the following:

• Conservative
• Labour
• Liberal Democrats
• Scottish National Party (SNP)
• Scottish Socialist Party
• Green Party
• UK Independence Party
• An independent candidate
• Other
• Would not vote
• Undecided
• Refused

All the respondents who say they are undecided how they will vote, or refuse to answer (although not those who directly say that they will not vote) are then asked the second question:

Q Who would you be most inclined to support?

This follow-up question, often called the "squeeze" question, tends to find that many of those whose initial reaction is not to plump for any party have, nevertheless, a clear inclination or party preference. The voting figures that are reported, therefore, the "combined voting" figures calculated by taking together the answers from Q1 and Q2. (Past experience at elections has shown that the combined figures are a better guide to the electorate's voting behaviour than the figures from Q1 alone -- at most elections most of the "incliners" still vote.)

For the combined voting figures we simply add together for each party those who named it in response to either question. Those who said they "would not vote" at either question are similarly added together, while the residual "undecided" and "refused" respondents are those who still gave this response having been asked both Q1 and Q2.

The next stage is weighting the data. Weighting is a statistical procedure, used to ensure that the achieved sample is representative of the population. For example, women make up 52% of the British adult population; if we were to find that in one poll they made up 55% of the sample, the computer would down-weight their answers slightly to ensure that in the final figures the opinions of women contributed the correct 52% of the numbers and men the other 48%. In practice the effect of weighting is almost always very small, and represents fine-tuning of the data rather than anything more drastic. Data in the SPM have been weighted by local authority, sex and age, to match the profile of the population.

For analysis of voting intentions, the data is filtered on those aged 18 and over who said that they would be certain to vote. We ask respondents to rate their likelihood to vote on a scale of 1-10 where 10 means absolutely certain to vote. The data is then filtered so that only the responses of those who answer 10 out of 10, absolutely certain to vote make up the final figures.

Finally, voting intention figures exclude those who say they would not vote or refuse to name a party.