What Happens at the Polling Place: Using Administrative Data to Look Inside Elections

• Author: Kenneth R. Mayer,Donald P. Moynihan,Barry C. Burden,Jacob R. Neiheisel,David T. Canon
• DOI: http://doi.org/10.1111/puar.12592
• Published date: 01 May 2017
• Date: 01 May 2017

354 Public Administration Review • May | June 2017

Public Administration Review,

Vol. 77, Iss. 3, pp. 354–364. © 2016 by

The American Society for Public Administration.

DOI: 10.1111/puar.12592.

Kenneth R. Mayer is professor of political

science and faculty affiliate in the La Follette School

of Public Affairs at the University of Wisconsin–

Madison. His research interests are campaign

finance, election administration, and presidential

decision making. He is author of

With the Stroke of

a Pen: Executive Orders and Presidential Power

and

The Dysfunctional Congress?

(with David T. Canon).

He is active as an expert witness in campaign

finance, voting rights, and redistricting litigation.

E-mail: krmayer@wisc.edu

David T. Canon is professor and chair of the

political science department at the University of

Wisconsin–Madison. His research interests include

election administration, racial representation,

partisan realignments, political careers, and the

historical study of Congress (especially congressional

committees). He is author of

Race, Redistricting, and

Representation, The Dysfunctional Congress? The

Individual Roots of an Institutional Dilemma

(with

Kenneth R. Mayer),

Actors, Athletes, and Astronauts,

American Politics Today,

several edited books, and

various articles and book chapters.

E-mail: dcanon@polisci.wisc.edu

Barry C. Burden is professor of political science

and director of the Elections Research Center at the

University of Wisconsin–Madison. He is coeditor

of

The Measure of American Elections

(2014) and

author of

Personal Roots of Representation

(2007),

as well as numerous journal articles. His research

addresses various aspects of American politics, with

a focus on election administration.

E-mail: bcburden@wisc.edu

Abstract : Tremendous attention has been paid to local election administration since the 2000 presidential election

meltdown, yet policy makers still lack basic information about what happens at the polling place. One strategy to

understand the interactions between citizens and street-level election bureaucrats is to turn to administrative data.

Using logs collected by polling place workers, the authors analyze more than 66,000 individual incidents recorded

from four different statewide elections. Such data provide novel insights and guidance for the administration of

elections. Findings indicate that task scale (in terms of the number of ballots) and complexity (in terms of absentee

ballots) increase the incident rate. Managerial choices about how polling places are run also matter: the use of

electronic voting machines and central count processing of ballots reduce the incident rate, while splitting poll worker

shifts increases it. Operator capacity, measured in terms of experience, also reduces the number of incidents.

Practitioner Points

• Administrative data reveal how poll workers administer elections.

• Roughly 1 per 100 voters experiences a polling place incident, but this incident rate varies across types of

elections.

• The incident rate is affected by task scale, task complexity, operator capacity, and managerial choices.

• Local election officials face trade-offs between reducing incidents and meeting other administrative goals.

Barry C. Burden

David T. Canon

Kenneth R. Mayer

Donald P. Moynihan

University of Wisconsin , Madison

Jacob R. Neiheisel

University at Buffalo , SUNY

What Happens at the Polling Place:

Using Administrative Data to Look Inside Elections

T he polling place is a classic venue for citizen–

state interactions. Street-level bureaucrats

wield considerable power and discretion

to help or hinder citizens in their pursuit of a

fundamental right. But what happens at the polling

place remains opaque, despite intense political interest

and regulatory efforts. In the aftermath of the 2000

presidential election, policy makers at the federal,

state, and local levels implemented reforms designed

to improve how U.S. elections are administered.

Contemporary ideological debates and policy

actions about elections—such as the degree to which

ballot security and voter access are compromised or

maintained—center on polling place issues (Stewart

2013b). While we know something about voter

experiences at the polls (e.g., Hall and Stewart 2013),

scholars, administrators, and policy makers still lack

basic information on the functioning of polling places

from an administrative perspective.

How can we uncover such basic citizen–state

interactions? In many cases they leave a paper or

digital trail because street-level bureaucrats are

required to record their activities. Such information

can be converted to administrative data. For elections,

we draw on a previously untapped database of polling

place activities: incident logs. Incident logs offer

“an underutilized source of important information

and data that can be used in comprehensive election

performance analysis” (Alvarez, Atkeson, and Hall

2012 , 45). These administrative data are collected

across the United States and offer perhaps the best

way to get “under the hood” of what happens at the

polls. 1 Whereas most election data are examined

at the state or county level, incident logs provide

insights into the lowest level of administration

where poll workers interact with voters. Our use

of administrative data allows the first large-scale

systematic documentation of polling place events

from an administrative perspective.

Given the novel nature of the data, we devote a

good deal of attention to explaining incident logs

as an example of administrative data that can help

to gauge the performance of election systems.

Using a form provided by the state, poll workers

are instructed to write down a description of each

incident, indicate what time it occurred, and assign a

code that corresponds to the event. We merge these

incident data with census information and data from

the Election Assistance Commission (EAC). The

resulting data set covers four Wisconsin elections,

Donald P. Moynihan is professor in the La

Follette School of Public Affairs at the University

of Wisconsin–Madison. His research examines

performance management, administrative burdens,

election administration, and employee behavior.

He is a fellow of the National Academy of Public

Administration and prior winner of the Kershaw

Award for outstanding contributions to the study of

public policy and management.

E-mail: dmoynihan@lafollette.wisc.edu

Jacob R. Neiheisel is assistant professor of

political science at the University at Buffalo, SUNY.

His research interests include election administration,

political parties, religion and politics, and political

communication. His work has appeared in such

outlets as

Legislative Studies Quarterly, Political

Research Quarterly, American Politics Research,

Political Communication,

and

State Politics & Policy

Quarterly.

E-mail: jacobnei@buffalo.edu