Observations of 2007 Scottish Elections

Last year, Minneapolis voters elected to use Single Transferable Vote (more commonly called Instant Runoff Voting) for municipal elections.

As planning gets underway to implement the new system, a delegation from Minnesota had the opportunity to observe the first-time use of Single Transferable Vote (STV) in Scotland for its municipal elections on May 3rd.

The delegation included myself, Minneapolis Council Member Elizabeth Glidden, Minneapolis Elections Director Cindy Reichert, Hennepin County Elections Director Michelle DesJardin, Secretary of State Election Director Gary Poser and FairVote Minnesota Board and Minneapolis Charter Commission member, Tryone Bujold.

We were part of a larger national delegation, which included voting reform advocates, election administrators and reporters from places across the country that have adopted or are implementing Single Transferable Vote. Observers from other countries were there as well to observe the integrity and other aspects of the elections.

The delegation was sponsored by the Electoral Reform Society (ERS) of England, a leading expert on election methods and advocate for alternative voting methods, in particular the Single Transferable Vote form of proportional representation. Washington D.C. based FairVote helped coordinate the national delegation.

The Electoral Reform Society has published a comprehensive report of the elections and there has been extensive analysis from other sources as well to which I provide links here.

For the purpose of this write up, I provide a summary of key observations and lessons learned regarding the Scottish STV elections that have some relevance to the implementation of this voting method in Minneapolis and potentially other jurisdictions in Minnesota in the future, including Saint Paul where a ballot measure campaign is currently underway.

About the Elections

On May 3rd, Scottish voters went to the polls to elect their representatives to the Scottish Parliament (called Holyrood) and their local city councils. Parliamentary and local elections are held every four years and have been held on the same election day since 2003. Voters elected a total 129 members to Parliament and 1,222 councilors to 32 local councils. There were also local government elections in England and National Assembly elections in Wales but only Scotland currently uses STV for local elections.

For the first time, members of city councils were elected by proportional representation using Single Transferable Vote. Formerly, and like in most U.S. and other British cities still, single representatives for each local ward were elected by the standard First Past the Post method. This changed in 2003, when the Scottish Parliament adopted STV for the election of local representatives.

Members to the Scottish Parliament are also elected by proportional representation, but using a different method called the Additional Member System (AMS). This method was used in the first two parliamentary elections in 1999 and 2003.

Key Observations and Considerations for Minneapolis Elections

STV Elections

Most excitingly, Single Transferable Vote achieved its goal of providing proportional representation in local wards and, despite using it for the first time, voters appeared to understand the Single Transferable Vote ballot and the value of ranking candidates, with most ranking three or more candidates on the ballot. See election results and commentary analysis for additional information on election analysis.

This outcome is consistent with exit polls that demonstrate voter understanding and preference for ranked choice voting where it's in use in the United States (San Francisco, Cambridge MA, Burlington VT and Takoma Park MD).

Single Transferable Vote is now being recommended to replace Additional Member System for the next Scottish Parliamentary election which, if adopted, would mean that that both local and parliament elections would be using the same style method of election. This recommendation is supported by the Electoral Reform Society as well as the Association of Electoral Administrators and the Scotland Branch of the Scottish Assessors Association, which would like to see consistent voting methods used for Scottish parliament and local elections.

Ballot Design

As reporters waited for results to come in on election night, news was coming out of the vote counting centers that a large number of ballots were being invalidated because voters improperly completed their ballots. The problem was with the Additional Member System ballot used for the Scottish Parliament election, not the Single Transferable Vote ballot used for the local elections. On the Additional Member System ballot, there were two votes - the first was the regional vote for a party or individual and the second was the constituency vote for the local candidate. This is the same voting method that was used in the previous two elections of the Scottish parliament (1999 and 2003) so the method was familiar to voters. In the previous two elections, however, there were separate ballot papers for each vote and a lower rate of spoiled ballots, 1.3% combined regional vote and constituency vote ballots in 2003, in comparison to the 7% for the current elections - 4.07% spoiled constituency ballots and 2.88% spoiled regional ballots.

The main problem was poor ballot instructions and design. The phrase You have two votes at the top of the page led to over voting, meaning putting more than one x in a single column instead of one in each; and under voting, meaning voting in only one column (i.e. using only one vote).

Other contributing factors included:

  • Combining the two votes on one ballot paper, which was determined to be the best ballot design following extensive consultation and user testing
  • Modifying the ballot design in some areas after the sample ballot was mailed to voters
  • Inconsistency in "adjudicating" the ballots within and across the central counting locations so that a ballot found to be invalid in one location may have been valid in another
  • Problems with postal voting, with hundreds of people receiving their postal ballot too late
  • Variances in turnout and age among the constituency areas, with lower turnout rates and higher median ages experiencing higher than average spoiled ballot rates

Voters did not have the same problem filling out the Single Transferable Vote ballot, with a spoiled ballot rate of just 1.85 percent. The lower spoiled ballot rate suggests that voters were not confused by the ballot, despite the fact that they were using STV for the first time.

The spoiled ballot rate was higher than in the previous election, however, and this is likely due to the change from marking a traditional "x" to using numbers to rank candidates in preferential order. The most typical error made by voters on the STV ballot was marking an x instead of a number.

This illustrates the importance of considering traditional voting habits when designing and providing instructions for a ballot. But more fundamental to successful elections is solid ballot design and extensive user testing of alternative designs.

Some concern exists that the spoiled ballot problem was partially attributable to the combination of Parliamentary and local elections, which was a decision made by the Arbuthnott Commission following debate on the issue. But this does not appear to be a major factor. Not only was it the second time in which the two elections were combined, the spoiled ballot rate for both elections was low in the first combined election in 2003 and it was low for the STV election in this election as well. The spoiled ballot problem this election was mainly attributable to the poor design of the Additional Member System ballot, not the Single Transferable Vote ballot and decoupling the elections would not have solved this problem.

The upcoming 2008 Greater London Assembly (GLA) election also includes a combination of different voting methods for mayor (using the Supplementary Vote a modified form of Instant Runoff Voting) and the GLA (using the Additional Member System). The last GLA election in 2004, which included European Parliament elections, also experienced a high spoiled ballot rate due to poor ballot design. In preparation for the election, a detailed study, redesign and testing of the ballot has been done, in consultation with marketing and graphic design experts.

Voting Machines

Electronic counting in England is relatively new and was used for the first time in Scotland this election. The scanning machines were provided by Data and Research Services (DRS), which partnered with the election services division of the Electoral Reform Society to electronically count the ballots.

Testing and training of the new system took place before the election, but the unforeseen problems of processing folded postal ballots and adjudicating a large number of ballots caused problems and delays in reporting the results in some counting centers. See BBC news report and Computer World report.

Unlike in Minnesota and many other states, where ballots are scanned and tallied at precinct polling locations and provide voters with the opportunity to correct an improperly completed ballot, Scotland used a central counting system. All ballots were taken to regional counting centers, scanned, adjudicated (as needed) and tallied. Each center was responsible for reporting election results for their region.

Overall, the DRS equipment appeared to have performed well, aside from the delay caused by the need to adjudicate a very large number of ballots. Once all of the ballots were adjudicated, the DRS voting system computed the election results virtually instantaneously. A few centers delayed the count and tally of the local vote until the following day.

The immediate and accurate reporting of results demonstrates that ranked ballot voting does not create delays in reporting results, unless there other factors contributing to the delay such as delay in receiving mailed ballots.

The experience in Scotland also demonstrates the value of polling-place machines to scan paper ballots, a system that should be maintained for the implementation of Single Transferable Vote in Minneapolis. The next generation of precinct-based scanning machines in the U.S. will come equipped to scan and tally ranked ballots and are scheduled to be available by 2009, when Minneapolis is due to implement Single Transferable Vote. If the City is not ready or able to replace its existing equipment with new machines by then, it has the authority to delay implementation until the next election in 2013 or consider interim solutions, including, for example, the use of existing precinct-based scanners in combination with central scanners. This solution, which would require temporary certification from the Secretary of State, would use existing equipment to provide error notification (allowing voters to correct their ballots) and to count first choice votes at the precinct level and then use central off-the-shelf scanners and STV software to scan and tally the ballots the following day for races that require a runoff. The use of nonpropriety central scanners is an option being considered by the Greater London Authority for the next Greater London Assembly elections in 2008.

Election Integrity

Despite the Florida-type ballot design problem, the overall reliability of the DRS equipment, the ballot adjudication process and the observation component designed into the election day process helped ensure an apparent high degree of election integrity and transparency.

Given the nascent use of electronic counting in England, there are no national certification standards for equipment. Instead, vendors go through an extenstive proposal review and testing process with the jurisdiction in charge of elections. Concerns regarding the integrity of voting equipment and vote counting do not have the history they do in the United States. The greatest concerns related to potential electoral fraud in the postal voting system and the disparity in applying the rules for adjudicating the ballots.

A feature of integrity designed into the process is the use of bar codes, which while essentially tied a ballot to a voter, ensured that the scanning and counting of each ballot could be traced back to the ballot and that each ballot could be counted only once.

Voter Education

With the introduction a of new voting system, new ballot papers and a new method of counting all on the same day, election authorities planned for significant voter education prior to election day, with extensive education campaigns and training targeted to voters, candidates and election staff. The public bodies in charge of elections - the Electoral Commission, Vote Scotland and the Scottish Executive - in partnership with other pubic bodies, has primary responsibility for voter education and partners with several other organizations, including local councils, to carry out its education campaign.

Information packets, including sample ballots, were sent to all households. Very user-friendly educational materials were accessible on-line and "how to vote " posters and information officers were available at polling locations.

The cost of the public voter education for the Scottish elections was about 3 million pounds (2 million for advertising campaigns and 1 million for information officers) or about 1.43 pounds per voter.

The Electoral Reform Society played a significant role in educating voters, campaigns and the media about STV. A nonprofit group called Capability Scotland provided additional voter education materials and guidance to people with disabilities. Community radio helped get election information out as well.

Below are links to some very good educational tools, which will be helpful in designing voter educational materials for the implementation of Single Transferable Vote in Minneapolis:

Despite their breadth and scope, these voter education measures were insufficient to fully counter poor ballot design of the Scottish Parliament ballot. Even the information officers placed at each polling location did not anticipate the potential confusion of the ballot design and generally provided instructions only to voters who asked for help. When implementing change of this significance in an election, consideration should be given to the provision of on-site instructions to each voter.

Campaining under Single Transferable Vote

Campaigning under Single Transferable Vote was new to candidates and their campaigns. Guides were provided by both the Electoral Commission and Electoral Reform Society and these provide some excellent examples for the design of materials in Minneapolis.

Campaigning in the United States is different from England, where comparatively little Get Out The Vote effort is made by campaigns running up to the election. Voter turnout is more reliant on the public education campaign than the candidate campaigns. It may be even more critical to educate campaigns here about Single Transferable Votes and how they can, in turn, educate their voters.


Scotland introduced multiple changes in the 2007 election. It changed the method of electing local council members, redesigned the ballot papers, and used machines for the first time to count the ballots. Considerable and careful planning was undertaken to implement each of these changes.

Despite this, the elections faced a major problem with spoiled ballots due not to the multiplicity of changes themselves, but to a flawed ballot design in the Scottish Parliament election, which would have occurred regardless of new voting counting machines and the introduction of STV for local elections. This demonstrates the fundamental role good ballot design plays in the successful outcome of elections.

This and several other lessons have been learned from the trip to Scotland. The Electoral Reform Society has provided a unique opportunity to observe STV first hand and learn the approaches of another country. Comparative observations provide invaluable insight and what was learned on this journey will help inform the Secretary of State's "IRV Taskforce", which will launch in June, and the implementation process of STV as it moves forward in Minneapolis and is considered in Saint Paul and potentially elsewhere in Minnesota.

For additional information about the Scottish elections or questions about this report, contact:

Jeanne Massey
Executive Director, FairVote Minnesota









Become a Volunteer


Contribute and help us

Build a better democracy

Join the Movement

Join the Movement