Vote buying allegations in Argentina

There was an article in this morning’s New York Times on allegations of vote buying in Argentina. The story recounted a number of stories of possible vote buying schemes, in particular in Tucuman.

This brought to mind the fact that social scientists (including my colleague Julia Pomares who is quoted in the NY Times story, who I have written a number of papers with regarding election integrity and the use of voting technologies to increase voter confidence) have written extensively in recent years on the topic of election fraud, in particular focusing on the development of tool that can be used to detect potential fraud in elections data.

Much of the earlier work on election forensics is presented in the book that I edited with Thad Hall and Susan Hyde, “Election Fraud: Detecting and Deterring Electoral Manipulation.” More recently, Ines Levin and I co-edited a Political Analysis virtual issue that summarized a number of papers that have been published in that journal about this topic, “Election Fraud and Electoral Integrity.” The current issue of Political Analysis has two new papers in it on election forensics, by Montgomery, Olivella, Potter and Crisp, and another by Medzihorsky.

A number of colleagues who follow Argentine elections, and those who study election fraud and integrity, will be following the current elections closely. There’s no doubt that the tools and techniques that social scientists have developed in recent years will be applied to the current elections in Argentina, and elections throughout the world in coming years, to help improve our understanding of elections and their integrity.

Automatic voter registration coming to California

California’s Governor Brown over the weekend signed into law new steps to automatically register eligible voters when they get their driver’s license in the state. Here’s a link to a NY Times story on the legislation.

The new automatic registration process won’t take effect until sometime next year, after California’s new voter registration system is fully functional. However, once the automatic registration procedures are enacted, though, this will mean that many eligible voters will be added to the state’s voter registration database and thus they won’t have to take any extra steps in order to register to vote. By being in the voter registration system, these new voters will start to get information from state and local election officials about upcoming elections, and no doubt they will also start to get communications from campaigns urging them to vote.

The longer-term effects of this new automatic voter registration procedure will no doubt be the subject of lots of research in the future. We’ll of course want to study changes in voter registration and turnout in the state, but also it will be important to determine what this new automatic registration system might imply for the workflow for local election administrators in California.

Two new research articles on elections

This morning two new research articles on elections were published electronically by Political Analysis, one on election forensics and the other on measuring the competitiveness of elections. Both should be of interest to Election Updates readers.

The first is by Jacob Montgomery, Santiago Olivella, Joshua Potter and Brian Crisp, “An Informed Forensics Approach to Detecting Vote Irregularies.” Here’s the abstract of their paper:

Electoral forensics involves examining election results for anomalies to efficiently identify patterns indicative of electoral irregularities. However, there is disagreement about which, if any, forensics tool is most effective at identifying fraud, and there is no method for integrating multiple tools. Moreover, forensic efforts have failed to systematically take advantage of country-specific details that might aid in diagnosing fraud. We deploy a Bayesian additive regression trees (BART) model—a machine-learning technique—on a large cross-national data set to explore the dense network of potential relationships between various forensic indicators of anomalies and electoral fraud risk factors, on the one hand, and the likelihood of fraud, on the other. This approach allows us to arbitrate between the relative importance of different forensic and contextual features for identifying electoral fraud and results in a diagnostic tool that can be relatively easily implemented in cross-national research.

This paper contributes to a series of papers published in PA that develop and text new methodologies for the detection of election irregularities and potential fraud.

The second paper is by Kai Quek and Michael Sances, “Closeness Counts: Increasing Precision and Reducing Errors in Mass Election Predictions.” Measuring the closeness of election contests is important for those who study elections, so this is a paper that readers of this blog should find of considerable interest. Here is the paper’s abstract:

Mass election predictions are increasingly used by election forecasters and public opinion scholars. While they are potentially powerful tools for answering a variety of social science questions, existing measures are limited in that they ask about victors rather than voteshares. We show that asking survey respondents to predict voteshares is a viable and superior alternative to asking them to predict winners. After showing respondents can make sensible quantitative predictions, we demonstrate how traditional qualitative forecasts lead to mistaken inferences. In particular, qualitative predictions vastly overstate the degree of partisan bias in election forecasts, and lead to wrong conclusions regarding how political knowledge exacerbates this bias. We also show how election predictions can aid in the use of elections as natural experiments, using the effect of the 2012 election on partisan economic perceptions as an example. Our results have implications for multiple constituencies, from methodologists and pollsters to political scientists and interdisciplinary scholars of collective intelligence.

The more the merrier, polling-place division

At the risk of becoming Doug Chapin’s Mini Me, I’m prompted to pile on Doug’s post today about the controversy in Summit County, Ohio over whether a state senator can serve as an Election Day poll worker.

As a college professor who has now worked for 15 years to bridge the gap between academics and election officials, I can’t help but cheer on state Sen Frank LaRose, who has applied to be a poll worker, but is being opposed by half the Summit County elections board, on the theory that only “regular citizens” should be poll workers.

Leaving aside the odd spectacle of a county elections board turning down a perfectly good poll worker applicant, the perspective that only a select tribe of individuals — ordinary citizens or highly trained professionals — can acquire real hands-on experience helping to run a polling place only hurts the cause of better election administration in the long run.  If we’ve learned anything from the constant attention over the past two decades to how elections are conducted, it’s that the world of election administration is often too insular.

My own personal experience has been focused on how this insularity oftentimes means that election administration is cut off from advances in the academic, non-profit, and business worlds.  The Summit County case highlights the insularity of election administration from the legislators who fund elections and write the laws that govern them.  How many times have I heard complaints from state and local election officials about county commissioners or state legislators making decisions that just make no sense, from the perspective of the trenches?

Poll workers are called on to make myriad decisions that affect the experience of voters, including whether they get to vote at all.  What better way is there for a state legislator to understand how election laws actually get implemented than to have him go through poll worker training and then to spend the day implementing election law in a polling place?

Count me as another voice in favor of granting Sen. LaRose his request to live out next Election Day working a precinct.

New research on the effects of voter registration deadlines

Recently in Political Analysis, a journal that I co-edit with my Caltech colleague Jonathan Katz, we published an interesting paper that studies the effects of voter registration deadlines using a novel approach. This paper, authored by Alex Street, Thomas A. Murray, John Blitzer, and Rajan S. Patel, is titled “Estimating Voter Registration Deadline Effects with Web Search Data.” The paper is available open access online, and here’s the abstract. It’s worth a read, both the results and the approach are quite interesting.

“Electoral rules have the potential to affect the size and composition of the voting public. Yet scholars disagree over whether requiring voters to register well in advance of Election Day reduces turnout. We present a new approach, using web searches for “voter registration” to measure interest in registering, both before and after registration deadlines for the 2012 U.S. presidential election. Many Americans sought information on “voter registration” even after the deadline in their state had passed. Combining web search data with evidence on the timing of registration for 80 million Americans, we model the relationship between search and registration. Extrapolating this relationship to the post-deadline period, we estimate that an additional 3–4 million Americans would have registered in time to vote, if deadlines had been extended to Election Day. We test our approach by predicting out of sample and with historical data. Web search data provide new opportunities to measure and study information-seeking behavior.”

“Ballot selfies”

A recent federal court decision in New Hampshire struck down that state’s ban on taking selfies in a voting booth, see stories in the Washington Post and this editorial in the LA Times. This is an interesting, and potentially important ruling, because if it stands it might open the door to a wave of new voting technologies. If a voter has a right to leave the polling place with a digital image of their ballot, that potentially opens the door to the development and implementation of new approaches to voter verification.

However, this ruling also means that election administrators are going to need to clarify or develop rules and regulations to insure that “ballot selfies” don’t become a distraction in the polling place. While some voters may want their “ballot selfie”, others may want privacy — and balancing the two might be a delicate task in a crowded polling place.

Mail Ballot Drop-Off Patterns

Doug Chapin’s most recent post on his Election Academy blog tells the tale of the late delivery of 1,270 mail ballots in a recent election in Orem, Utah.  This post brings to mind a surprising result (at least to me) from the 2014 Survey of the Performance of American Elections (SPAE) about the return of mail ballots.

In 2014, for the first time, the SPAE asked respondents who voted by mail how they returned their ballots.  Nationwide, 2/3 of absentee and mail ballot voters returned their ballots by mail.  That’s not the surprising part.  This is what surprised me:   If we look only at respondents from the three “vote by mail” states — Colorado, Oregon, and Washington — only 1/2 of “vote by mail” voters report returning their ballots using the Postal Service.  Half the voters in these states took their ballot to an official elections site to be counted — 38% of these used a dedicated drop box, 29% went to the main elections office, and the rest went to a combination of places, including traditional neighborhood precincts and early voting centers.

Even those who used the Postal Service did not often use the convenience of front-door pick-up to return their ballot.  Only 40% of voters who used the Postal Service to return their ballot had their own mail carrier pick up the ballot.  An even larger fraction (46%) took their ballot down to the post office, while the rest deposited their ballot in the mailbox around the corner.

I have one additional point to add to this.  Ever since I started administering the SPAE in 2008, I have asked voters how confident they were that their vote was counted as cast.  Each time I have asked this, voters using the mails have expressed significantly less confidence than those who voted in-person, either on Election Day or through early voting.  In 2014, for instance, 67% of those who voted by mail (or absentee) said they were very confident their vote was counted as intended, compared to 76% of Election Day voters and 73% of early voters.

The results from 2014 help to show that this lower confidence in postal voting is related to how the ballot is returned.  The following graph illustrates the relationship.  The dots illustrate the fraction of “mail” voters who answered they were very confident their vote was counted as cast, broken down by how they returned their ballot.  The “whiskers” around the dots are the 95% confidence interval around those estimates.  The dashed vertical line shows the fraction of in-person voters who were very confident.  Note that the “mail” voters who used the various Postal Service delivery routes were all less confident than those who voted in person.  Those who returned their ballots at the mail election office, or who used a vote center, were just as confident, and maybe even more confident, in the case of vote centers.  (The lower confidence among those who left their mail ballots at a neighborhood precinct is a little puzzling, but it might be related to the fact that very few people actually leave their mail ballots at Election Day precincts, which means that precinct workers may not always know what to do with them.)

absentee_confidenceThe Orem situation shared by Doug helps to illustrate the reality behind these national statistics.  In the aggregate, voters seem to recognize that if they leave it to the Postal Service to deliver their ballot, there is some risk involved.  By their behavior, vote-by-mail voters appear to like getting their ballot in the mail, but return it by mail?  Not as much.



Last day of FL early voting is big

The last day of in-person early voter in Florida ended the period with a bang — total turnout on the last Sunday was about triple four years ago, and the Democratic share was even greater than four years ago.

For the entire period, the relative Democratic share of the in-person early vote has been greater than in 2010. As Michael McDonald notes, the Democratic share of the absentee/mail vote is also greater than 2010. What remains to be seen is whether this is just shifting around when partisans vote, or if it reflects a shift in partisan electoral fortunes from four years ago.  It’s obviously a mix of both; we’ll know soon enough what the mix is.

Here are the graphs.  Click on a graph to see the full picture.

Day-to-day in-person early voting turnout:




Cumulative early voting turnout:day_to_day_cumulative_gross_20141103



Partisan composition of early voting turnout, compared to 2010:day_to_day_pct_gross_20131103