-Aspen's historic May 5, 2009 IRV election audited as single ballots- 5/5/09 Aspen CO held an instant runoff election (IRV) for mayor and 2 council members. Interpreted contents of each ballot, scanned by True Ballot, were publicly released. Open records requests for a CD of image scans were denied. Aspen has been sued to protect records from destruction and to allow inspection of the scanned ballot files. A Court of Appeals ruling holds that unidentifiable ballots are public records.

Search this and related blogs

Friday, January 29, 2010

Report prepared for Aspen by Scott Adler on Release of Cast Election Ballots for Public Inspection

The following is a Disclosure of Testimony by the City of Aspen for Marks v. Koch for expert witness E. Scott Adler.  This document was partially converted for the web by H. Branscomb and annotated with yellow background on some key phrases.  The original and complete pdf is located here: Disclosure of Testimony. Please report any transcription errors to H. Branscomb as a comment to this blog.

Assessing the Impact of the Release of Cast Election Ballots for Public Inspection

Report prepared for the matter of Marks v. Koch, 09 CV 294

E. Scott Adler
Associate Professor of Political Science
University of Colorado, Boulder

January 21, 2010


      It is my contention that the release of election ballots or ballot images for public inspection would cause substantial injury to the public interest.  I base this position on my knowledge of research on the history of American elections and studies of contemporary voter behavior, election administration and public opinion.

Curbing Voter Fraud as a Motivation for the Adoption of the Secret Ballot

      Ballot secrecy is now considered synonymous with democracy. Its use dates back to Athens in the sixth century B.C.E. (Albright 1942) and Roman assemblies (Yakobsen 1995), and is even codified by the United Nations as a crucial components of a free and fair electoral process (Franck 1992).

      What led to the adoption of the Australian ballot?  The standard view of Progressive era reforms – the Australian ballot among them – is that such efforts were the culmination of a backlash to widespread partisan corruption and electoral fraud (Evans 1917; Freedman 1968). Political corruption at the time included such practices as the payment of voters, candidates required to pay parties in order to have their names placed on lists of preferred candidates, stuffing ballot boxes with votes by non-existent voters (Schaffer 2002; Ware 2000).  As a means of eliminating the “vote market,” politicians, under public pressure fostered by reform proponents, adopted various changes to electoral processes.  Among the most important was the institution of government printed ballots administered by election officials which could not be removed from polling places and would be cast in secret.  The primary purpose of such balloting systems was to disconnect the ballot’s contents from the identity of the voter.  This type of ballot, would eliminate the ability of political parties to “control” votes, and therefore the outcome of elections through the payment of voters to cast ballots in favor of their preferred candidate(s). Government controlled secret ballots removed the ability of parties to print distinctive ballots that were used as a means of confirming that votes were cast in accordance with the conditions of some kind of financial transaction.  Australian ballots were also a means of protecting vulnerable voters from intimidation (Schaffer 2002; Graves 1967)

      In the United States, it is generally considered that the widespread adoption of the secret ballot began in Louisville, Kentucky, which initially adopted a secret ballot to curb citywide political corruption in 1888.  Louisville was followed in the same year by the state of Massachusetts. By 1889 ten states had adopted the Australian ballot, and subsequently 31 states had adopted this form of secret voting by 1892 (Fredman 1968).  Ranney argues that such reforms, were a means of regulating the internal affairs of parties and converting them from private clubs to public entities (Ranney 1975).  Keller et al. (2006) assert that “Today, all fifty states have provisions in their constitutions for either election by ‘secret ballot’ or elections in which ‘secrecy shall be preserved’ which has been interpreted by the courts as an implied requirement for secret balloting” (314-15,2006).

      While the received wisdom that the adoption of the secret ballot was motivated by an effort to curb political corruption, there is a revisionist literature that argues that there might have been a secondary motive to the rapid adoption of this reform by established party officials: secret ballots were a means of disenfranchising likely supporters of opposition parties and suppressing their votes (Kousser 1974; Walker 2005; Ware 2000).  Decreases in voter turnout that resulted from the introduction of these reforms is the evidence often pointed to that confirms this “strategic disenfranchisement” hypothesis.  The argument is that, particularly in the South, the adoption of votes cast in secret would hinder the ability of black and illiterate voters to cast ballots in favor of Republican candidates.

      Perhaps the best study to adjudicate these two position [sic] on the motive for adoption of the secret ballot was conducted by Heckelman (2000).  Heckelman examines how the adoption of Australian ballot differentially affects two populations – poor voters (the vote market hypothesis) and black and illetarate [sic] voters (the strategic disenfranchisement argument).  He finds that there is no evidence to support the claim that the drop in turnout that resulted from the adoption of the secret ballot disproportionately affected illiterate or black populations.  However, he does demonstrate that it had a very significant effect on poor voters; the secret ballot took away the ability of parties to bribe voters. Thus low-interest voters who had previously only voted in elections to engage in voter fraud no longer had an incentive to participate.

Substantial Public Injury Could Result from Public Inspection of Ballots

      This then takes us to the question before us today: Will making cast ballots, or copies of cast ballots, available for public inspection cause substantial injury to the public interest?  I will contend that public release of ballots, even when no name is officially attached to those ballots, could still have three separate deleterious effects on voting and voting turnout.  1) To dissuade potential voters from participating.  2) For those who do vote, it sharply decreases the likelihood of voting their actual preferences if such preferences do not conform to the views of those in their social or economic groups.  3) To increase the likelihood of voter fraud or intimidation.

Dissuading Voters

      Voting is already thought to be an equivocal behavior for most individuals in the United States.  Voter turnout rates in the United States rank fairly low among advanced democracies world wide – which is the case for a number of reasons: the high costs of voting (registration requirements, elections conducted on weekdays, etc.), single-member districts/elected offices (no proportional representation), voter fatigue due to the frequency of elections, etc. (Teixeira 1992; Wayne 2003).  Political scientists have long known that for most voters the costs of voting, especially with the structure of our electoral system in the United States, significantly outweighs the benefits for individual voters (Riker and Ordeshook 1968). Disruptions to the process by which elections are conducted have been shown to have a dampening effect on already low turnout.  For instance, the simple act of consolidating polling places has been shown to significantly reduce turnout even among highly motivated voters in local elections (McNulty, Dowling, and Ariotti 2009).

      While there is no good time series data to demonstrate that changes in election administration alter public opinion regarding confidence in the electoral process, it is known that diminished confidence in election administration decreases political participation. Using public opinion polls, Alvarez et al. (2008) find a strong correlation between voter confidence in election administration and likelihood [sic] future turnout:  Voters who lack confidence in how elections are conducted are significantly less likely to vote in the future.

      Local elections, such as that for the office of Aspen mayor, are already low turnout affairs.  One recent study of local elections in California found that only 44 percent of registered voters cast ballots in mayoral elections, and the percent that participate decreases even further when city council races are examined (Hajnal and Lewis 2003). Driving more voters away from the polls because of suspicions about the secrecy of cast ballots will decrease citizen involvement and participation in public affairs and governance and increase the likelihood that election outcomes are determined by ever-smaller proportions of the citizenry.

Altering Vote Choice

      Releasing all ballots to the public creates the perception that individual vote choices may be identifiable.  In nearly all municipalities in the United States ballots or ballot images are not publically accessible.  For the most part, voters cast ballots confident that their vote choice, if they desire, will be kept secret.  Recent polling shows that the vast majority of voters believe that their vote choice is kept secret, unless revealed voluntarily (Gerber et al. 2009).

      However, voters who believe that the secrecy of their voting decisions are compromised by election administration exhibit significantly different voting patterns than those who do not.  Gerber et al. (2009) test the hypothesis that the votes of individuals who believe that the protection of their secret ballot is undermined by election administration reflect not just their own personal preferences but also their fears about defying the wishes of those who would learn of their choices.  Specifically, they examine whether the voting patterns of union members who are skeptical of the secrecy of their vote against the voting patterns of union members confident in their vote secrecy.  Gerber and coauthors find a sizable and significant decrease in the likelihood of union members voting for McCain in 2008 if they believe that the secrecy of their vote is compromised.  This conforms to what we already know about the effect of the introduction of the Australian ballot in the late-Nineteenth Century – that as voters perceived their electoral choices were free of outside pressures we saw a significant rise in split-ticket voting (Rusk 1970).

Increasing the Likelihood of Voter Fraud or Intimidation

      Public release of ballots reintroduces the possibility of voter fraud or selling votes.  For more than a century the secrecy of ballots in the U.S. has effectively eliminated the ability of individuals to sell their votes since, absent the record of a counted ballot, there is no way to confirm that the paid individual executed their end of the transaction.  Public release of ballots or ballot images can act as a means of verifying that the sale of a vote was carried out correctly.  Individuals wishing to confirm that they voted for a particular candidate or issue position may write-in a clearly distinctive candidate name (“Oscar Mayer,” for example), cast votes in a distinctive and very unlikely pattern, or inconspicuously mark their ballot in such a way that uniquely identifies it.
      Exploring the means by which voters may collude in such nefarious ways, Keller et al. state:

Elections with many contests, with write-ins allowed, or with information-rich ranked preference contests, implicitly contain extra fields in which to encode voter identity.  For example, if an election contains eight judicial retention questions, there are at least 6561 possible ways to complete a ballot, assuming Yes, No and No Preference are all options for each question.  Very few precincts will have over 6561 votes cast within them, so a systematic vote buyer could demand that every voter cast a uniquely identifying vote pattern on judicial retentions.  That unique pattern, plus the precinct marked on a ballot, in turn, could be correlated with a desired vote for a contested office. (Keller, Mertz, Hall, and Urken 2006)

Similarly, public release of ballots or ballot images facilitates the possibility of voter intimidation or efforts to reveal the electoral choices of voters by individuals with access to ballots prior to votes being cast (for example, spouses or family members under main-in ballots or poll workers).  Ballots could easily be marked in a revealing manner or voters may possess distinctive handwriting, such that the identity of the voter is discernable from the ballot or ballot image.  The result of public accessibility to ballots, as discussed above, is a decreased sense of secrecy, and a greater likelihood that potential voters will opt not to participate or cast votes that incorporate more than their personal preferences.

      Keller et al. prescription to avoid problems like vote collusion is precisely the recommendation to severely limit access to ballots. They state:

The best we can do is to control the disclosure of full ballots to mandated authorities, and maintain the chain of custody over the ballots, including the EBIs (Electronic Ballot Images).  A full ballot must be maintained, but only aggregations of votes, per contest, are disclosed to the general public.  The number of people who have access to full ballots should be as limited as is feasible, and even people with access to some full ballots should not necessarily be granted general access to all full ballots” (Keller, Mertz, Hall, and Urken 2006)

      The concern of stray marks on ballots is compounded by the fact that many states and municipalities have laws necessitating the disqualification of ballots with such markings.  Under mandates ro release ballots publicly, election officials are faced with the matter of adjudicating what marks may or may not distinguish the ballot.  This is very likely to lead to the disqualification of a high percentages [sic] or ballots due to stray and possibly unintentional marks out of concern that such ballots, a) could possibly identify the individual voters, or b) could be used as confirmation as part of an agreement to sell one’s vote.  Again, under elections with low voter turnout, this leaves the outcome of the race determined by an even smaller subset of eligible voters.

Works Cited

(this portion was not extracted for the purposes of this site- see the original document here: Disclosure of Testimony )

No comments: