SPICe briefing REJECTED BALLOT PAPERS. 26 June /36

Transcription

1 REJECTED BALLOT PAPERS STEPHEN HERBERT AND TOM EDWARDS This paper summarises what constituted a ballot paper at the elections in May and then details the level of ballot at the 2007 Scottish Parliament election and also at the local government elections held on the same day. SPICe briefing 26 June /36 The authors of the paper wish to express their thanks to Dr Christopher Carman and Professor James Mitchell for permission to draw upon research they have conducted. Scottish Parliament Information Centre (SPICe) Briefings are compiled for the benefit of the Members of the Parliament and their personal staff. Authors are available to discuss the contents of these with MSPs and their staff who should contact Stephen Herbert on extension or stephen.herbert@scottish.parliament.uk. Members of the public or external organisations may comment on this briefing by ing us at spice@scottish.parliament.uk. However, researchers are unable to enter into personal discussion in relation to SPICe Briefing Papers. If you have any general questions about the work of the Parliament you can the Parliament s Public Information Service at sp.info@scottish.parliament.uk. Every effort is made to ensure that the information contained in SPICe briefings is correct at the time of publication. Readers should be aware however that briefings are not necessarily updated or otherwise amended to reflect subsequent changes. 1

2 KEY POINTS The term ballot refers to a ballot paper that cannot be counted for one or more of a range of reasons. These are: the ballot does not have an official mark; the voter has cast more votes than they are entitled to (termed over-voting ); the voter has made writing or marks by which they can be identified; the voter has left the ballot paper blank or has marked it in such a way that it is not clear for whom they intended to vote There were 146,097 ballots at the 2007 Scottish Parliament elections or 3.47% of all votes cast There were 85,643 ballots on the constituency vote or 4.07% of votes cast. The level of ballots ranged from 7.9% of votes cast in Glasgow to 3% of ballots in Mid Scotland and Fife Of the ten constituencies with the highest proportion of ballots, 8 were in Glasgow and 2 in Lothian There were 60,454 ballots on the regional vote or 2.88% of votes cast. The level of ballots ranged from 4.2% of votes cast in Glasgow to 2.1% in North East Scotland There were sixteen constituencies where the number of ballot exceeded the majority of the winning candidate Statistical analysis of the ballot by academics from Strathclyde University resulted in two principal conclusions. Firstly that there is a statistically significant relationship between the relative level of social deprivation in a constituency and the level of ballot. In other words the higher the relative level of deprivation the higher the level of ballot. Secondly, regions with a larger number of parties on the regional list had higher levels of ballot rejection At the local government elections in 2007, using the Single Transferable Vote system of election, there were 38,319 ballots or 1.85% of votes cast. The level of ballots ranged from 1.11% in East Dunbartonshire to 2.77% in West Dunbartonshire The Electoral Commission announced on 4 May 2007 that an external review of the Scottish elections would take place. The remit for the review has been announced and the review is expected to be complete by the end of August 2

3 CONTENTS KEY POINTS...2 INTRODUCTION...4 WHAT IS A REJECTED BALLOT PAPER?...4 LACK OF OFFICIAL MARK...4 OVER VOTING...4 WRITING OR MARKS BY WHICH THE VOTER CAN BE IDENTIFIED...5 UNMARKED...5 VOID FOR UNCERTAINTY...5 REJECTED BALLOT PAPERS...5 ANALYSIS OF REJECTED BALLOT PAPERS...12 CORRELATION...13 REGRESSION...14 LOCAL GOVERNMENT ELECTION...15 INDEPENDENT REVIEW...18 ANNEX ONE REJECTED BALLOT PAPERS ON CONSTITUENCY AND REGIONAL VOTE BY PARLIAMENTARY REGION...19 SOURCES

4 INTRODUCTION The SPICe briefing Election 2007 (Herbert et al 2007) provided provisional data on the level of ballot at the Scottish Parliament election in 2007 due to the limited data that was available at the time the briefing went to print. Subsequently a complete set of data on ballot for both the Scottish Parliament and local government elections has been published by the Electoral Commission (Electoral Commission 2007a). This paper describes what constituted a ballot paper at both elections in May and then details the level of ballot at the 2007 Scottish Parliament election and also at the local government elections held on the same day. WHAT IS A REJECTED BALLOT PAPER? The term ballot refers to a ballot paper that cannot be counted for one or more of the following reasons: It does not have the official mark The voter has cast more votes than they are entitled to, called over voting The voter has made writing or marks by which they can be identified The voter has left the ballot paper blank or has marked it in such a way that it is not clear for whom they intended to vote The decision as to whether a ballot paper should be classed as can only be made by the Returning Officer. Candidates and agents have a right to witness, and to object to, decisions made. Each of the reasons for a ballot paper being are considered in more detail below. LACK OF OFFICIAL MARK All ballot carry an official mark. At the 2007 Scottish elections this was incorporated in each ballot paper when it was printed. A ballot paper would be for the lack of an official mark only if the mark had been removed or defaced by the voter. OVER VOTING Each voter at a Scottish Parliamentary election has one vote for a party or individual candidate contesting their region and one vote for a candidate contesting a constituency. The voter marks their vote using a cross (X) or similar mark. A ballot paper would be considered as constituting over-voting if: At a Scottish Parliamentary regional election, the voter marks more than one vote on the regional side of their ballot paper, then their regional vote cannot be counted. At a Scottish Parliamentary constituency election if votes are cast for more than one constituency candidate, then the constituency vote cannot be counted. 4

5 The voter marks more than one vote on either the regional or the constituency side of their ballot paper then that side of the ballot paper is classed as spoiled. A single vote cast on the other side of the ballot paper however would be counted. However, if a voter marks either their regional or constituency ballot paper using numbers (1, 2, 3, 4, etc), then that vote is not as voting for more than one candidate if there is a single number 1 against one candidate. The ballot paper is considered valid for that candidate. At a Scottish local government election, each voter should rank candidates in order of preference (1, 2, 3, 4 etc) rather than by making a single choice with a cross (X). At a Scottish local government election, if the ballot paper is marked, for example, 1, 1, 2, 3, 4, etc. (i.e. there are two first preferences cast), it should be classified as voting for more than one candidate and marked as spoiled. If, for example, the ballot paper is marked 1, 2, 3, 5, 6, 7, etc. (i.e. there is no fourth preference indicated), the first three preferences should be accepted after which the ballot paper becomes an exhausted ballot. Ballot such as these are described as partially accepted and are not classified as spoiled. WRITING OR MARKS BY WHICH THE VOTER CAN BE IDENTIFIED If it can be reasonably maintained that a voter has written their name, signature, initials or address on their ballot paper, then that paper must not be counted. UNMARKED Unmarked ballot cannot be counted. If at a Scottish Parliamentary election the voter marks one side of the ballot paper but not the other, then the marked side would be counted and the unmarked side would be classed as spoiled. VOID FOR UNCERTAINTY Voters must mark a clear vote at a Scottish Parliamentary election or a clear first preference at a local government election in order for their vote to be valid. If it is not clear for whom a vote has been cast (for example, if the centre of a cross or preference sits on the line between two candidates) then the ballot paper is classed as spoiled. REJECTED BALLOT PAPERS At the 2007 Scottish Parliament election there were 4,205,262 votes cast (including ballots) of which 146,097 were ballots or 3.47% of all votes cast. On the constituency vote there were 2,102,631 votes cast (including ballots) of which 85,643 were ballots or 4.07% of constituency votes cast. On the regional vote there were 2,102, 631 votes cast (including ballots) of which 60,454 were ballots or 2.88% of votes cast. 5

6 Table One details the number of ballots on the constituency vote by Parliamentary region at the Scottish Parliament election. It is important to note that ballots are included in the number of votes cast. Annex One details the number of ballot for each constituency. Table One clearly indicates considerable variation in the number of ballot between Parliamentary region with ballots accounting for 7.9% of votes cast in Glasgow and 5.2% in Lothians with Mid Scotland and Fife having the lowest proportion of ballots (3%). Table One - Rejected Ballot Papers on the Constituency Vote by Parliamentary Region Scottish Parliament Election 2007 votes cast % of Central Scotland 293,326 10, % Glasgow 215,667 16, % Highlands and Islands 191,279 6, % Lothians 296,132 15, % Mid Scotland and Fife 279,684 8, % North East Scotland 265,592 9, % South of Scotland 284,810 9, % West of Scotland 276,141 9, % Scotland 2,102,631 85, % Source: Electoral Commission (2007a) Of the ten constituencies with the highest proportion of ballots as a proportion of votes cast, 8 were in Glasgow and 2 in Lothians. The ten constituencies were as follows: Glasgow Shettleston (12.09%) Glasgow Maryhill (10.18%) Glasgow Pollok (9.79%) Glasgow Baillieston (9.67%) Glasgow Springburn (8.81%) Edinburgh East and Musselburgh (7.76%) Glasgow Cathcart (7.35%) Glasgow Anniesland (7.27%) Midlothian (6.16%) Glasgow Rutherglen (5.81%) The maps overleaf depict constituency ballots by constituency as a percentage of votes cast and constituency ballots by Parliamentary region as a percentage of votes cast. 6

7 7

8 8

9 Table Two details the number of ballot by Parliamentary region at the 2007 Scottish Parliament election. Annex One details the number of ballot on the regional vote for each constituency. There were 25,189 fewer ballot on the regional vote than was the case for the constituency vote across Scotland as a whole. In every Parliamentary region there were more ballot on the constituency vote than for the regional vote. The difference between the number of ballot on the constituency and regional vote varied from 7,933 in Glasgow to 972 in the Highlands and Islands. On average there were 3,149 more ballots on the constituency than the regional vote in each Parliamentary region. The proportion of ballot on the regional vote was again highest in Glasgow (4.2%) and lowest in North East Scotland (2.1%). The lower number of ballot on the regional vote is reflected in ballots accounting for 2.9% of all regional votes cast compared to ballots accounting for 4.1% of constituency votes cast. Again it is important to note that ballots are included in the number of votes cast. Table Two - Rejected Ballot Papers on the Regional Vote by Parliamentary Region Scottish Parliament Election 2007 votes cast % of Central Scotland 293,326 8, % Glasgow 215,667 9, % Highlands and Islands 191,279 5, % Lothians 296,132 9, % Mid Scotland and Fife 279,684 6, % North East Scotland 265,592 5, % South of Scotland 284,810 7, % West of Scotland 276,141 7, % Scotland 2,102,631 60, % Source: Electoral Commission (2007a) Table Three details the ballots on the constituency vote as a percentage of total votes cast ( ballots are included in total votes cast) for the Scottish Parliament elections in 1999, 2003 and The proportion of ballots increased between 2003 and 1999 by 0.3%. However there was a considerable increase in the proportion of spoilt ballots in 2007 compared to either of the two previous Scottish Parliament elections. In 2007 ballots accounted for 3.7% more votes cast than was the case in In comparison to 2003 there were 3.4% more ballots in Table Three - Rejected Ballot Papers as a % of Total Votes Cast on the Constituency Vote by Parliamentary Region Scottish Parliament Elections 1999, 2003 and Central Scotland 0.33% 0.67% 3.71% Glasgow 0.45% 0.68% 7.85% Highlands and Islands 0.40% 0.62% 3.39% Lothians 0.30% 0.72% 5.20% Mid Scotland and Fife 0.26% 0.57% 3.02% North East Scotland 0.28% 0.61% 3.44% South of Scotland 0.35% 0.57% 3.17% West of Scotland 0.33% 0.66% 3.39% Scotland 0.33% 0.64% 4.07% Source: Scottish Parliament (2007); Electoral Commission (2007a) 9

10 The maps below depict regional ballots by constituency as a percentage of votes cast and regional ballots by region as a percentage of votes cast. 10

11 11

12 Table Four details the ballots on the regional vote as a percentage of total votes cast ( ballots are included in total votes cast) for the Scottish Parliament elections in 1999, 2003 and As with the constituency vote the proportion of regional votes also increased between 1999 and 2003 by 0.3%. The only Parliamentary region which was an exception to this trend was Glasgow where the proportion of ballots fell by 0.06%. Comparing 2007 to the 1999 Scottish Parliament election the proportion of ballots increased by 2.57% in 2007 across Scotland as a whole. In comparison to the 2003 election there were 2.26% more ballots in 2007 across Scotland as a whole. Table Four - Rejected Ballot Papers as a % of Total Votes Cast on the Regional Vote by Parliamentary Region Scottish Parliament Elections 1999, 2003 and Central Scotland 0.33% 0.64% 3.00% Glasgow 0.38% 0.32% 4.17% Highlands and Islands 0.51% 0.86% 2.88% Lothians 0.28% 0.57% 3.07% Mid Scotland and Fife 0.22% 0.59% 2.36% North East Scotland 0.26% 0.62% 2.10% South of Scotland 0.32% 0.62% 2.77% West of Scotland 0.26% 0.75% 2.88% Scotland 0.31% 0.62% 2.88% Source: Scottish Parliament (2007); Electoral Commission (2007a) There were sixteen constituencies where the number of ballot exceeded the majority of the winning candidate. The constituencies were as follows (with the number of ballot minus the winning majority - in brackets): Airdrie and Shotts (90) Falkirk West (381) Glasgow Govan (476) Argyll and Bute (85) Edinburgh Central (308) Edinburgh East and Musselburgh (1,139) Linlithgow (572) Livingston (764) Ochil (526) Central Fife (90) Dunfermline West (281) Stirling (13) Aberdeen Central (394) Tweedale, Ettrick and Lauderdale (216) Eastwood (99) Cunninghame North (967) ANALYSIS OF REJECTED BALLOT PAPERS Preliminary statistical analysis of the pattern of ballot has been conducted by Dr Christopher Carman and Professor James Mitchell of Strathclyde University. This section of the paper provides an overview of their preliminary findings (Carman and Mitchell, 2007). Carman and Mitchell have conducted two forms of statistical analysis on the ballot paper data, 12

13 correlation and regression analysis, in order to test whether a relationship exists between a range of social and institutional factors and the level of ballot. The remainder of this section summarises the principal findings of the analysis conducted by Carman and Mitchell. CORRELATION 1 Carman and Mitchell carried out statistical analysis to determine whether there was a correlation between a number of social and institutional indicators and constituencies with a higher percentage of ballot. In terms of the social indicators used these were: Percentage of adults without academic qualifications by constituency Percentage of the public that are unemployed claimants by constituency Percentage of the public reporting that they are not in good health by constituency These indicators were adopted in order to test whether constituencies with higher levels of social deprivation and lower average levels of educational qualifications would have higher relative rates of ballot paper rejection. Carman and Mitchell found a significant relationship between each of these three variables above and higher levels of ballot 2. This leads them to conclude that: Clearly there is a strong relationship between constituency social context and the relative level of ballots generated across constituencies (Carman and Mitchell 2007, p 3) Carman and Mitchell also tested for correlation against two institutional variables, namely: The rate of ballot paper rejection at previous Scottish Parliament elections by constituency The number of parties listed on the regional ballot paper They found that rates of ballot paper rejection at previous Scottish Parliament elections were not related to the level of ballot in However in terms of the number of parties on the regional ballot a significant relationship between variables is evident 3. This leads the authors to conclude that: The problems associated with ballots increased with the number of parties on the regional lists. We take this as evidence that changes to the ballots made to accommodate the large number of parties registered for the regional list in certain regions had a strong and significant relationship with the percentage of ballots (Carman and Mitchell 2007). 1 A correlation is a basic statistical measure which seeks to ascertain whether there is an association between two variables. It is important to note that a correlation does not imply that there is a causal relationship between two variables i.e. that one variable influences or affects the other. 2 Academic qualifications was significant at.54 on the constituency vote and.57 on the regional vote. Percentage claiming unemployment benefits was significant at.49 on the constituency vote and.41 on the regional vote. Percentage reporting not in good health was significant at.78 on the constituency vote and.68 on the regional vote. All statistics significant at p< Significant at.72 (p<.001). 13

14 However the authors highlight that it is unclear whether the statistical relationship they found is due to the increased number of parties on the ballot paper or a consequence of the truncation of instructions on the ballot paper as a result of the increased number of parties. REGRESSION 4 In terms of the regression analysis carried out the authors sought to explain the influence of a range of indicators on the relative level of ballots on the constituency and regional vote. The indicators used were: An indicator of social deprivation (this was a combined indicator consisting of the three social indicators outlined above under correlation ) The number of parties on the regional list The percentage of ballots at the 2003 election The constituency turnout rate in 2007 The winning candidate s majority by constituency in 2007 (not used for regression on the regional list) The number of constituencies at the count centre that counted each constituency Using this model Carman and Mitchell are able to account for 85% of the variance in levels of ballot paper rejection. The authors obtained the following results: The rate of spoilt ballot in 2003 and the majority of the winning candidate in 2007 did not have a significant influence on the percentage of ballot in 2007 The higher the turnout in a constituency in 2007, the fewer ballot there were in that constituency Constituencies counted at the larger count centres tended to have higher rates of ballot rejection (although this result is heavily influenced by the Glasgow and Lothian counts) The greater the degree of social deprivation in a constituency the higher the rate of ballot rejections The number of parties on the regional list proved to be the strongest predictor in the model. In others words, as the number of parties on the regional list increased so did the rate of ballot In terms of analysis on the regional ballot the following indicators were found not to be significant: rate of ballots in 2003, turnout rate in 2007 and the number of constituencies counted at a constituency s count centre. However on the regional vote there were two significant relationships: The number of parties on the regional list was a significant predictor of regional ballots in constituencies The measure of social deprivation was also a significant predictor of regional ballots and proved to be the strongest predictor in the model dealing with regional ballots 4 Regression is a statistical procedure which seeks to estimate the relative influences of different variables upon a particular outcome, for example, the influence of the number of parties on the ballot paper upon the level of ballot. Regression analysis is a statistical test which seeks to ascertain whether there is a causal relationship between variables. Regression analysis operates by assessing the relationship between a dependent variable (in this case percentage of ballot ) against a set of specified independent variables termed predictors (such as, the number of parties on the ballot paper). Regression holds all other independent variables in a model constant while examining the influence of any one predictor upon the dependent variable. 14

15 Lastly the authors undertook a direct comparison between constituencies with 23 parties on the regional list (i.e. Glasgow and Lothian) which therefore had altered ballot instructions 5 on the ballot paper and all other constituencies. On this basis Carman and Mitchell find highly significant differences between the constituencies based on ballot design. These results lead the authors to conclude: There are two very clear findings from these analyses. First, there is a glaring and distinct relationship between the relative level of social deprivation in a constituency and that constituency s relative level of ballots. Given past research this is to be expected (though perhaps not to the degree that we find in Scotland in 2007). Second, there is a clear difference in the relative rates of ballot rejection that is related to the ballots given to voters in different regions. Voters in the Glasgow and Lothians regions were given ballots that engendered higher rates of ballot rejection. This second finding is more troubling as it represents a systemic failure in the design of the ballots used in Scotland in 2007 (Carman and Mitchell 2007, p 12). LOCAL GOVERNMENT ELECTION The local government elections in 2007 were held on the same day at the election in The Single Transferable Vote replaced First Past the Post as the electoral system at the election. Electronic counting was used at all the counting centres for the local government elections. Table Five details the number of ballot at the 2007 local government elections. In total there were 38,319 ballots at the 2007 local government elections. The level of ballots across all Scottish local authorities was 1.85%, considerably below that for the Scottish Parliament elections (4.07% on the constituency vote and 2.88% on the regional vote) clearly indicating that there were features of the Scottish Parliament electoral process which accentuated the level of ballot. The level of ballot varied from 1.11% in East Dunbartonshire to 2.77% in West Dunbartonshire. 5 Due to the large number of parties standing on the regional list in Glasgow and the Lothians (23 in each case) it appears that the instructions at the top of the ballot paper for voters in these regions was changed in order to ensure that all the parties could be accommodated on the ballot paper. 15

16 Table Five Rejected Ballot Papers at the 2007 Local Government Elections Local Authority votes cast % of Aberdeen City 77,369 1, % Aberdeenshire 95,944 1, % Angus 32, % Argyll & Bute 29, % City of Edinburgh 196,489 2, % Clackmannanshire 20, % Comhairle nan Eilean Siar 14, % Dumfries & Galloway 65,691 1, % Dundee City 53,385 1, % East Ayrshire 50,634 1, % East Dunbartonshire 50, % East Lothian 42, % East Renfrewshire 43, % Falkirk 61,270 1, % Fife 141,506 2, % Glasgow City 192,881 4, % Highland 97,151 1, % Inverclyde 32, % Midlothian 34, % Moray 34, % North Ayrshire 56,769 1, % North Lanarkshire 125,802 2, % Orkney Islands 8, % Perth & Kinross 63,663 1, % Renfrewshire 71,986 1, % Scottish Borders 49, % Shetland Islands 10, % South Ayrshire 49, % South Lanarkshire 126,683 2, % Stirling 40, % West Dunbartonshire 37,111 1, % West Lothian 65,041 1, % Scotland 2,072,530 38, % Source: Electoral Commission (2007a) The map overleaf depicts ballots in the local government election by local authority as a percentage of votes cast. 16

17 17

18 INDEPENDENT REVIEW The Electoral Commission announced on 4 May 2007 that it would expand its statutory report on the Scottish elections to include a full, independent review of the elections in Scotland (Electoral Commission 2007b, p 1). As part of this process the Electoral Commission has appointed Mr Ron Gould 6 to lead an external review of the Scottish elections. The remit for the review, which the Electoral Commission has asked Mr Gould to examine, is as follows: the reasons for the high number of ballots the electronic counting process and its impact on the final results the arrangements for the production and dispatch of postal ballot packs the decision to combine the Parliamentary and local government elections the decision to electronically count both the local and the Parliamentary ballot where the decision making in relation to these issues did take place and where it should have, either according to law or to responsibility the role of the Commission itself in the preparations for the elections Source: Electoral Commission (2007b) p 2. It is expected that the review will be completed by the end of August Background information regarding Mr Gould can be accessed at: 18

19 ANNEX ONE REJECTED BALLOT PAPERS ON CONSTITUENCY AND REGIONAL VOTE BY PARLIAMENTARY REGION Kilmarnock Loudoun and Constituency votes cast Central Scotland Regional % of votes cast % of Constituency election administered by the Returning Officer for 34,453 1, % 34,453 1, % East Ayrshire Falkirk East 31,352 1, % 31, % Falkirk Falkirk West 29,942 1, % 29, % Falkirk Airdrie and Shotts 28,696 1, % 28, % North Lanarkshire Coatbridge and 27,010 1, % 27, % North Chryston Lanarkshire Cumbernauld and 27, % 27,185 1, % North Kilsyth Lanarkshire Hamilton North and Bellshill Motherwell Wishaw and 26,711 1, % 26, % North Lanarkshire 27, % 27, % North Lanarkshire East Kilbride 36,935 1, % 36, % South Lanarkshire Hamilton South 23, % 23, % South Lanarkshire Central Scotland 293,326 10, % 293,326 8, % 19

20 Constituency votes cast Glasgow Regional % of votes cast % of Constituency election administered by the Returning Officer for Glasgow Anniesland 23,875 1, % 23, % Glasgow City Glasgow Baillieston 19,122 1, % 19, % Glasgow City Glasgow Cathcart 23,374 1, % 23, % Glasgow City Glasgow Govan 22,741 1, % 22,741 1, % Glasgow City Glasgow Kelvin 24,695 1, % 24, % Glasgow City Glasgow Maryhill 18,441 1, % 18, % Glasgow City Glasgow Pollok 21,522 2, % 21, % Glasgow City Glasgow Shettleston 16,836 2, % 16, % Glasgow City Glasgow Springburn 19,314 1, % 19,314 1, % Glasgow City Glasgow Rutherglen 25,747 1, % 25, % South Lanarkshire Glasgow Region 215,667 16, % 215,667 9, % 20

21 Constituency votes cast Highland Regional % of votes cast % of Constituency election administered by the Returning Officer for Argyll and Bute 29, % 29, % Argyll and Western (Eilean Siar) Isles Bute 14, % 14, % Comhairle nan Eilean Siar 23, % 23, % Highland Caithness, Sutherland and Easter Ross Inverness East, 40,858 1, % 40, % Highland Nairn and Lochaber Ross, Skye and 33,138 1, % 33, % Highland Inverness West Moray 31,292 1, % 31, % Moray Orkney 8, % 8, % Orkney Islands Shetland 10, % 10, % Shetland Islands Highlands and Islands Region 191,279 6, % 191,279 5, % 21

22 Constituency votes cast Lothians Regional % of votes cast % of Constituency election administered by the Returning Officer for Edinburgh Central 30,897 1, % 30, % Edinburgh City Edinburgh East 32,489 2, % 32,489 1, % Edinburgh and Musselburgh City Edinburgh North 33,221 1, % 33,221 1, % Edinburgh and Leith City Edinburgh Pentlands 35,905 1, % 35,905 1, % Edinburgh City Edinburgh South 33,976 1, % 33,976 1, % Edinburgh City Edinburgh West 36,657 1, % 36, % Edinburgh City Midlothian 26,760 1, % 26, % Midlothian Linlithgow 31,369 1, % 31,369 1, % West Lothian Livingston 34,858 1, % 34, % West Lothian Lothians Region 296,132 15, % 296,132 9, % 22

23 Constituency votes cast Mid Scotland and Fife Regional % of votes cast % of Constituency election administered by the Returning Officer for Ochil 32,569 1, % 32, % Clackmanan shire Central Fife 28,221 1, % 28, % Fife Dunfermline East 25,730 1, % 25, % Fife Dunfermline West 30, % 30, % Fife Kirkcaldy 25, % 25, % Fife North East Fife 32, % 32, % Fife Perth 35, % 35, % Perth and Kinross Stirling 33, % 33, % Stirling North Tayside 36,421 1, % 36,421 1, % Perth and Kinross Mid-Scotland and Fife Region 279,684 8, % 279,684 6, % Constituency votes cast North East Scotland Regional % of votes cast % of Constituency election administered by the Returning Officer for Aberdeen Central 21, % 21, % Aberdeen City Aberdeen North 26,064 1, % 26, % Aberdeen City Aberdeen South 31,024 1, % 31, % Aberdeen City Banff and Buchan 28,728 1, % 28, % Aberdeenshire Gordon 36, % 36, % Aberdeenshire West Aberdeenshire and Kincardine 35, % 35, % Aberdeenshire Angus 32, % 32, % Angus Dundee East 27, % 27, % Dundee City Dundee West 25, % 25, % Dundee City North East 265,592 9, % 265,592 5, % Scotland Region 23

24 Constituency votes cast South of Scotland Regional % of votes cast % of Constituency election administered by the Returning Officer for Dumfries 34,425 1, % 34,425 1, % Dumfries & Galloway Galloway and 31, % 31, % Dumfries & Upper Nithside Galloway East Lothian 35,540 1, % 35, % East Lothian Cunninghame South 24,477 1, % 24, % North Ayrshire Roxburgh and 26, % 26, % Scottish Berwickshire Borders Tweeddale, Ettrick and Lauderdale 31, % 31, % Scottish Borders Ayr 31, % 31, % South Ayrshire Carrick, Cumnock 35,217 1, % 35,217 1, % South and Doon Valley Ayrshire Clydesdale 34,538 1, % 34, % South Lanarkshire South of Scotland Region 284,810 9, % 284,810 7, % 24

25 Strathkelvin Bearsden and Constituency votes cast West of Scotland Regional % of votes cast % of rejecte d Constituency election administered by the Returning Officer for 37,693 1, % 37, % East Dunbartonshire Eastwood 43, % 43,155 1, % East Renfrewshire Greenock and 24,204 1, % 24, % Inverclyde Inverclyde Cunninghame 31,256 1, % 31, % North Ayrshire North Paisley North 23, % 23, % Renfrewshire Paisley South 26,642 1, % 26, % Renfrewshire West Renfrewshire 30, % 30, % Renfrewshire Clydebank and 28,165 1, % 28, % West Milngavie Dunbartonshire Dumbarton 31,068 1, % 31, % West Dunbartonshire West of Scotland Region 276,141 9, % 276,141 7, % Constituency votes cast Scotland Regional % of votes cast % of Scotland 2,102,631 85, % 2,102,631 60, % - Constituency election administered by the Returning Officer for 25

26 SOURCES Carman, C. and Mitchell, J. (2007) An Examination of Ballot Rejection in the Scottish Parliamentary Election of Unpublished Manuscript. Glasgow: University of Strathclyde. Electoral Commission. (2007a) Rejected ballot at the 3 May Scottish Elections. Edinburgh: Electoral Commission. Available at: Electoral Commission. (2007b) Review of the 2007 elections to the Scottish Parliament and local government by the Electoral Commission. Edinburgh: Electoral Commission. Available at: Herbert, S. et al. (2007) Election 2007 Edinburgh: Scottish Parliament Information Centre. Available at: Scottish Parliament. (2007) Election Results Spreadsheet 2007 Source Data. Edinburgh: Scottish Parliament. Available at: 03May2007Master_001.xls 26