Scottish independence / Appendix

Appendix
 

The data for Figure 7.1 are shown below:

undefined

 
The full question text for the 2011 questions reported in Table 7.1 is given below. In each case there were five answer options: in addition to the middle option of "no difference", respondents could choose "a lot" or "a little" to indicate the strength of their positive or negative view. In the table these two sets of responses have been combined.

Thinking now about what might happen if Scotland were to become an independent country, separate from the rest of the United Kingdom but part of the European Union.

Do you think that, as a result of independence, taxes in Scotland would become higher, lower or would it make no difference?

As a result of independence would Scotland's economy become better, worse, or would it make no difference?

As a result of independence would the standard of the health service in Scotland become better, worse, or would it make no difference?

As a result of independence would people in Scotland have more pride in their country, less pride or would it make no difference?

As a result of independence would the standard of living in Scotland be higher, lower, or would it make no difference?

As a result of independence would Scotland have a stronger voice in the world, a weaker voice, or would it make no difference?

Tables A.2 and A.3 report full details of the multivariate modelling, the results of which are summarised in Table 7.2.

This model is based on a stepwise regression with insignificant variables not entered. Dependent variable is support independence versus not doing so based on responses to the question detailed at Table A.1. Independent variables are shown in the order of entry into the model. All such variables apart from age are entered as interval level variables. In the case of the 'impact on independence' variables the higher the score, the more positive the evaluation. In the case of national identity, the higher the score the more Scottish and less British someone feels. In the case of age, the coefficients are simple contrast coefficients with the reference category shown in brackets.
 

undefined

 
This model is based on an ordinal regression of a five-point scale ranging from 'very confident' to 'very worried'. The higher the score on the dependent variable, the more confident and less worried someone was about independence. Independent variables are listed in rank order of their Wald scores, with variables that proved to be insignificant at the 5% level removed. In the case of the 'impact on independence' evaluations, all of which are entered as interval level variables, the higher the score the more positive the evaluation; in the case of impact on taxes, a high score represents a belief that taxes would be lower. In the case of age and sex coefficients are simple contrast coefficients with the reference category shown in brackets.
 

undefined



The data for Figure 7.2 are shown below:

undefined

Download chapter
Notes
  1. Such a commitment had also been included in the SNP's 2007 election manifesto. However, as a minority government between 2007 and 2011, the SNP lacked the votes in the Scottish Parliament needed to pass legislation authorising a ballot.
  2. In the case of the first two items the unweighted and weighted sample size in 1997 is 676. In the case of the remaining items the unweighted size is 657 and the weighted 659. The unweighted sample size for all items in 2011 is 1156 and the weighted 1167.
  3. Support for independence has been modelled using binary logistic regression in which the dependent variable is support for independence (either inside or outside the European Union) versus any other response. Confidence in independence has been modelled using ordinal logistic regression in which the dependent variable is a five-point scale ranging from "very confident" to "very worried".
  4. In contrast to binary logistic regression, the ordinal logistic procedure in SPSS does not provide a stepwise facility. This means we do not know the order in which the variables would be entered using such an approach. However, an alternative analysis of the data on confidence in independence using stepwise binary logistic regression revealed that the order in which the variables were entered using that approach was much the same as the order of the Wald scores reported by the ordinal regression.
  5. Bases for Table 7.3 are as follows:

    undefined
  6. The full question wording was as follows:

    Which of the following do you think has most influence over the way Scotland is run?

    And which do you think ought to have most influence over the way Scotland is run?

    [the Scottish Government, the UK government at Westminster, local councils in Scotland, the European Union]
  7. Note that, unlike the questions reported in Table 7.5, these questions did not offer the answer options "local councils in Scotland" or "European Union". As so few respondents chose these options when they were offered, their exclusion will have made little or no material difference to the pattern of response.
  8. We should note though that the balance in favour of decisions being made in Edinburgh rather than London is in both cases somewhat less than for the already devolved area of university tuition fees, where no less than 86 per cent think decisions should be taken by the Scottish Parliament and only 10 per cent say responsibility should lie with Westminster.
  9. Note that in contrast to the question reported at Figure 7.1, independence is not referred to here as involving 'separation' from the rest of the UK. In general, survey questions that include 'separation' in their description of independence have tended to elicit lower levels of support than those that do not.
  10. In the case of welfare benefits, the relevant figure in 2010 was 82 per cent while in the cass of taxes it was 83 per cent.
  11. The relevant 2010 figures are 61 per cent for welfare benefits and 54 per cent for taxes.
  12. Debarring Scottish MPs from voting on English laws would not necessarily prove unpopular with the general public north of the border either. When the question presented in Table 7.8 was last asked by Scottish Social Attitudes, in 2009, 47 per cent agreed that Scottish MPs should not vote on English laws, while only 22 per cent disagreed.
  13. In 2004-2006 the second option read "that makes decisions about the region's economy, planning and housing". The 2003 survey carried both versions of this option and demonstrated that the difference of wording did not make a material difference to 
the pattern of response. In Figure 7.2 the figures shown for 2003 are those for the two versions combined.
  14. It has also been suggested that the creation of directly elected mayors in the major cities of England might provide a focus for greater devolution in a manner that, along with the creation of the Greater London Assembly, it has already done. However, as the Constitutional reform chapter shows, public opinion towards directly elected mayors is somewhat equivocal and, in practice, when 10 of England's largest provincial cities were asked in May 2012 to vote in a referendum on whether they should have such a mayor, only one voted in favour.
  15. This increase would appear to be attributable to the increased concern about Scotland's share of public spending, albeit not wholly so. Those who think that Scotland secures more than its fair share are markedly more likely to support Scottish independence (33 per cent) than are those who do not think it secures more than its fair share (19 per cent). The increase in support for Scottish independence between 2007 and 2011 among those who say that Scotland secures more than its fair share is, at five points, a little less than the seven point increase in the population as a whole. At the same time, the equivalent figure among those who feel Scotland does not secure more than its fair share is, at two points, well below the general increase of seven. Some of that overall increase of seven points must therefore have arisen because of the rise between 2007 and 2011 in the proportion who think that Scotland secures more than its fair share.
  • Download chapter
  • Authors
  • Notes
    1. Such a commitment had also been included in the SNP's 2007 election manifesto. However, as a minority government between 2007 and 2011, the SNP lacked the votes in the Scottish Parliament needed to pass legislation authorising a ballot.
    2. In the case of the first two items the unweighted and weighted sample size in 1997 is 676. In the case of the remaining items the unweighted size is 657 and the weighted 659. The unweighted sample size for all items in 2011 is 1156 and the weighted 1167.
    3. Support for independence has been modelled using binary logistic regression in which the dependent variable is support for independence (either inside or outside the European Union) versus any other response. Confidence in independence has been modelled using ordinal logistic regression in which the dependent variable is a five-point scale ranging from "very confident" to "very worried".
    4. In contrast to binary logistic regression, the ordinal logistic procedure in SPSS does not provide a stepwise facility. This means we do not know the order in which the variables would be entered using such an approach. However, an alternative analysis of the data on confidence in independence using stepwise binary logistic regression revealed that the order in which the variables were entered using that approach was much the same as the order of the Wald scores reported by the ordinal regression.
    5. Bases for Table 7.3 are as follows:

      undefined
    6. The full question wording was as follows:

      Which of the following do you think has most influence over the way Scotland is run?

      And which do you think ought to have most influence over the way Scotland is run?

      [the Scottish Government, the UK government at Westminster, local councils in Scotland, the European Union]
    7. Note that, unlike the questions reported in Table 7.5, these questions did not offer the answer options "local councils in Scotland" or "European Union". As so few respondents chose these options when they were offered, their exclusion will have made little or no material difference to the pattern of response.
    8. We should note though that the balance in favour of decisions being made in Edinburgh rather than London is in both cases somewhat less than for the already devolved area of university tuition fees, where no less than 86 per cent think decisions should be taken by the Scottish Parliament and only 10 per cent say responsibility should lie with Westminster.
    9. Note that in contrast to the question reported at Figure 7.1, independence is not referred to here as involving 'separation' from the rest of the UK. In general, survey questions that include 'separation' in their description of independence have tended to elicit lower levels of support than those that do not.
    10. In the case of welfare benefits, the relevant figure in 2010 was 82 per cent while in the cass of taxes it was 83 per cent.
    11. The relevant 2010 figures are 61 per cent for welfare benefits and 54 per cent for taxes.
    12. Debarring Scottish MPs from voting on English laws would not necessarily prove unpopular with the general public north of the border either. When the question presented in Table 7.8 was last asked by Scottish Social Attitudes, in 2009, 47 per cent agreed that Scottish MPs should not vote on English laws, while only 22 per cent disagreed.
    13. In 2004-2006 the second option read "that makes decisions about the region's economy, planning and housing". The 2003 survey carried both versions of this option and demonstrated that the difference of wording did not make a material difference to 
the pattern of response. In Figure 7.2 the figures shown for 2003 are those for the two versions combined.
    14. It has also been suggested that the creation of directly elected mayors in the major cities of England might provide a focus for greater devolution in a manner that, along with the creation of the Greater London Assembly, it has already done. However, as the Constitutional reform chapter shows, public opinion towards directly elected mayors is somewhat equivocal and, in practice, when 10 of England's largest provincial cities were asked in May 2012 to vote in a referendum on whether they should have such a mayor, only one voted in favour.
    15. This increase would appear to be attributable to the increased concern about Scotland's share of public spending, albeit not wholly so. Those who think that Scotland secures more than its fair share are markedly more likely to support Scottish independence (33 per cent) than are those who do not think it secures more than its fair share (19 per cent). The increase in support for Scottish independence between 2007 and 2011 among those who say that Scotland secures more than its fair share is, at five points, a little less than the seven point increase in the population as a whole. At the same time, the equivalent figure among those who feel Scotland does not secure more than its fair share is, at two points, well below the general increase of seven. Some of that overall increase of seven points must therefore have arisen because of the rise between 2007 and 2011 in the proportion who think that Scotland secures more than its fair share.
  • Related links