IRV
news
videos
costs
turnout
majorityfailure
minorities
Disabled
currentuse
sanfrancisco
Scotland
votingsystems
vendors
rawvotedata
fairvote
runoff elections
3rdparty
tallying
About
studies
Download
e-mail me



 

 The Scottish Elections in 2007

An Electoral Train-wreck.  Scotland switched from hand counted paper ballots to computerized voting machines for the first time when implementing STV in May, 2007.  They decided to use different electoral systems and voting procedures. The results of so many changes and a complex ballot resulted in a loss of 100,000 ballots and voter confidence.  It is important to note that Scotland ranks better in literacy than the USA. Spoiled ballots impacted the poorer and lesser educated, but even some well educated voters had problems. On this page are news items about the May 2007 election, and at the bottom of the page some information on Scotland's literacy rate.

Scottish poll probe: e-counting gets 'hold off until safe' verdict  The Register. By Lucy Sherriff Posted in Government October 26, 2007 ..."The report strongly recommends against introducing electronic voting for the 2011 elections, until the electronic counting problems from the 2007 elections are resolved....In May this year, seven constituencies had to abandon the electronic counters, amid accusations of computer crashes and scanners failing to cope with folded ballot papers.

It later emerged that more than 140,000 ballots were declared spoiled. Some 70,000 of these
were rejected by the counting machines with no human oversight, effectively disenfranchising
almost two per cent of the voting public. First Minister Alex Salmond described the news as
'astonishing', and deeply disturbing."

Not so much an election as a national humiliation Scotland’s voters were treated with arrogance and contempt

The voting machines:  Scotland switches to computerized voting -  Scottish officials followed the advice of  "experts" including the non profit "The Electoral Reform Society" and switched from hand counted paper ballots to computerized voting in May '07 to support STV, a form of Instant Runoff:

3.3.1 Experience of counting STV elections elsewhere led the Scottish Executive to examine the feasibility of counting the ballot papers electronically. Following a procurement exercise in late 2005 DRS Data Services Limited (DRS) was selected as the preferred supplier to provide the equipment and support necessary to support such a process. DRS provided the e-counting services in the Greater London Authority / London Mayoral Elections in 2000 and 2004. DRS also involved Electoral Reform Services (ERS) in their proposals – ERS having extensive experience of STV elections. http://www.pkc.gov.uk/NR/rdonlyres/F26CB0C8-EE76-46CD-953A-43B39495261B/0/07271.pdf ("Electoral Reform Services" is related to the Electoral Reform Society, a non profit whose goal is to spread IRV).

The Scotsman provides more background  on the voting machines used in the 2007 election:

Hitches in the electronic counting system delayed several declarations. Why was it brought in and how much did it cost?
With a complex council voting system being used for the first time, ministers decided a computerised system would be quicker. DRS landed an £8.9 million contract to "e-count" elections in Scotland.

What is DRS? Based in Milton Keynes, DRS was set up nearly 50 years ago and has run electronic vote-counting systems in London, Norway and Mali. Former Labour leader Neil Kinnock is a non-executive director. The company says he was not involved in the awarding of the contract.

An Analysis

Dominion was just one of several voting machines tested in Scotland. Here are excerpts of a report created by the Open Rights Group regarding problems with the machines and the election.

May 2007 Election Report

Findings of the Open Rights Group Election Observation Mission in Scotland and England main page: 
 
Dominion Voting Systems - pilot in Stratford and Warwick
...
Election Report: May 2007
33
At midnight a power surge disabled four of the five scanners, though the servers storing votes were protected with
uninterruptible power supplies. After four minutes of systems checks by staff from subcontractor Dominion Voting, scanning
resumed with the batches in progress being started again from scratch. Immediately after scanning resumed, the RO
explained what had happened and reassured attendees that votes could not be counted twice.
 
As scanning continued, candidates and agents increasingly expressed concerns to ORG’s observers. Shortly before 1am the
RO declared that e-counting was taking too long and that, in response to requests from agents, some wards would be hand
counted. Manual counting continued until 2.15am at which point, with a few wards declared, the RO agreed to continue the
count at 10am. Subsequently a manager from lead supplier Software AG had discussions with the RO who, the manager
stated to ORG’s observers, had decided to switch to manual counting without consulting them. The manager also asserted
that staying with the e-counting would have been quicker than switching to a manual count. At approaching 3am, the RO
announced that all wards for the district council elections would be manually counted, though three multiple-seat parish
council elections would be electronically counted as they would still be slower to count manually.
 
At 10am the manual counts began and were completed without incident, followed by the e-counted parish elections which
were also declared without further problems.
 
Staff from the local authority and suppliers raised several reasons for the problems that occurred. Firstly the print quality, as
previously mentioned. Secondly the scanners were flagging up more ballots than expected for adjudication. The RO
estimated that 25% of votes cast were being sent for adjudication. Not only was the adjudication process slow but any
batch with doubtful votes was completely re-scanned, further delaying the process.
 
It isn’t clear why re-scanning was required or why so many votes were sent for adjudication. However the RO has ordered a
report to be written into the conduct of the election and has suggested to ORG observers that he feels the problems
encountered warrant withholding at least some of the payment due to the suppliers.
 
The pilot in Stratford was run jointly with Warwick District Council. A submission to ORG from the Green Party concerning
the count in Warwick39 recounts similar problems which resulted in the count there extending to a second attempt on the
afternoon of 4th May. Problems reported were system freezes, vote images being replaced onscreen with large red crosses
and a large amount of files needing to be manually moved and deleted after an ‘operator accidentally hit the wrong key’.
Furthermore scanning seemed to be very sensitive, causing ballots with small crosses, large crosses and non-centrally
aligned crosses as well as folds to require adjudication. Finally the submission raised concerns that attendees, including the
submission’s authors, had not felt clear how errors, double-counting of votes and under-counting of ballots stuck together
were avoided.
 
Page 42 & 43
 
Usability Observations
 
Difficult Adjudication Interfaces
 
Observers in Bedford and Stratford, as well as Green Party submissions from Breckland and Warwick, noted numerous
problems with the adjudication interfaces provided by e-counting suppliers. It was frequently reported that there was poor
alignment between software boxes to receive vote marks and the positioning of boxes on ballot papers making visual
inspection of how the system had interpreted a ballot—particularly with a large number of candidates—a difficult process.
The adjudication systems ORG observed also required a considerable amount of mouse operation, which proved to be slow
and tiring for operators. Some common operations, such as selecting the reason for rejecting a ballot in Stratford, always
took a minimum of two mouse clicks to select. The interfaces should have been streamlined for heavy use while being clear
for party observers to verify and query if necessary. According to a Green Party submission52 one ballot adjudication queried
in Breckland took thirty minutes to retrieve and re-code from the system, after having been stored before an agent had a
chance to protest.
 
Examples
 
(see link for picture of the machines/ballot boxes)
 
Page 44
 
Poor Process Design
...
ORG also questions the process design in Stratford and Warwick, which required entire batches to be re-scanned when a
single ballot was rejected by the scanner. Given that scanners observed showed an extreme sensitivity to tears, folds and
marks, ORG believes better planning should have gone into how this was to be handled during the counts. Because the
DCA preferred standard high-capacity Commercial Off-The Shelf (COTS) scanners over more specialised scanners
dedicated to e-counting, options for more intelligent ballot-handling during scanning were limited.
 
Page 59
 
ORG observed a large number of problems and potential vulnerabilities, and heard doubts expressed by candidates and
agents. The large number of counts observed which experienced significant delays and problems leads ORG to believe that
neither the technologies nor the processes used were sufficiently mature. The considerable number of procedural, technical
and usability oversights also leads ORG to the conclusion that the procurement of technologies for elections in Scotland and
England were flawed. The Scottish process was, overall, the better of the two, with some published verification of systems
and procedures at counts much more clearly defined but neither was satisfactory.
 
Page 60
 
Scottish Elections
 
...Parties which only stood candidates on the regional ballot of the Parliamentary elections were unfairly penalised by the poor ballot
design which probably lead voters to mark their constituency choice in the regional column.
 
In Conclusion
 
The technologies used at the May 2007 election—and the processes by which they were implemented—caused significant
problems that raise concerns about the accuracy of the results declared. The nature of the technologies and the problems
they caused have served to seriously undermine the faith that candidates, agents and voters have in the integrity of British
elections. Much of the responsibility for this lies with the Government, which has shown a naïve and insufficiently robust
approach to managing technologies and their suppliers. ORG concludes that, given the problems observed and the questions remaining unanswered, it cannot express confidence in the results declared in areas observed. Given these findings, ORG remains opposed to the introduction of e-voting and ecounting in the United Kingdom.
 
Open Rights Group
Election
 
 

Literacy -Mean scores in reading, mathematical and scientific literacy (2000)  

Scotland outperforms the UK, the USA and Germany in reading literacy