The Design and Practice of Integrating Evidence: Connecting Performance Management with Program Evaluation

• DOI: http://doi.org/10.1111/puar.12865
• Published date: 01 March 2018
• Date: 01 March 2018

The Design and Practice of Integrating Evidence: Connecting Performance Management with Program Evaluation 183

Public Administration Review,

Vol. 78, Iss. 2, pp. 183–194. © 2017 by

The American Society for Public Administration.

DOI: 10.1111/puar.12865.

Donald P. Moynihan is director of

the La Follette School of Public Affairs and

Epstein and Kellett Professor of Public

Affairs at the University of Wisconsin–

Madison. He is author, most recently, of

Toward Next-Generation Performance

Budgeting,

published by the World Bank

in 2016. He was the 2014 recipient of the

Kershaw Award, given every two years to a

scholar under the age of 40 for outstanding

contributions to the study of public policy

and management.

E-mail: dmoynihan@lafollette.wisc.edu

The Design and Practice of Integrating Evidence:

Connecting Performance Management with

Program Evaluation

Alexander Kroll is assistant professor

of public administration in the Steven J.

Green School of International and Public

Affairs at Florida International University.

His research interests are the management

of government organizations, the functions

and dysfunctions of performance systems,

and the role of organizational behavior in

improving public services. He has received

awards from the American Society for

Public Administration and the Academy of

Management.

E-mail: akroll@fiu.edu

Abstract : In recent decades, governments have invested in the creation of two forms of knowledge production

about government performance : program evaluations and performance management. Prior research has noted

tensions between these two approaches and the potential for complementarities when they are aligned. This article

offers empirical evidence on how program evaluations connect with performance management in the U.S. federal

government in 2000 and 2013. In the later time period, there is an interactive effect between the two approaches,

which, the authors argue, reflects deliberate efforts by the George W. Bush and Barack Obama administrations to

build closer connections between program evaluation and performance management. Drawing on the 2013 data,

the authors offer evidence that how evaluations are implemented matters and that evaluations facilitate performance

information use by reducing the causal uncertainty that managers face as they try to make sense of what performance

data mean .

Evidence for Practice

• Program evaluation and performance management produce different knowledge and are more effective when

they are complementary.

• Managers involved in program evaluations are more likely to use performance data because program

evaluations make causal inference easier, establishing synergies for performance information use.

• The connections between program evaluation and performance management do not arise naturally; rather,

they require a mixture of continuity and design.

A fundamental philosophical value of the study

of governance is that structural conditions

can be designed to improve governmental and

therefore societal outcomes. This design value is so

widely accepted that it saturates the language of public

administration. Consider, for example, the persistent

tendency by the advocates of any sort of change to

frame their ideas as “reform,” implying the need and

ability to engage in a corrective improvement.

The promise and perils of design also suffuse basic

tensions in the study of public administration

(Levine et al. 1975). Practitioners who criticize the

limited real-world relevance of scholarship are often

frustrated by the lack of clear design prescriptions:

what should we do? The long-running debate about

the role of rationality often centers on barriers

that undercut good design, such as the role of

politics (Moe 1989 ), the failure to account for local

conditions (Lindblom 1959 ), or the reliance on

simplistic assumptions of human behavior and the

functioning of administrative instruments (Andrews,

Pritchett, and Woolcock 2013 ; Heinrich and

Marschke 2010 ).

Another, less frequently considered, design issue

is how the different components of government

fit together. This issue is most often expressed in

concerns about how structural silos of resources and

expertise constrain the solving of “wicked problems”

that demand more collaborative designs both inside

and outside government (Emerson, Nabatchi,

and Balogh 2012 ; Skelcher, Mathur, and Smith

2005 ). A variation of this problem is how different

administrative policies or techniques fit together.

Light ( 1998 ) narrates how a variety of government

changes adopted in the United States over 50 years

piled atop one another. Each could be justified in

isolation, but the changes were not designed to work

together as part of a coherent framework.

In this article, we identify a related and common

design problem that is informed by these prior

critiques: the coordination of different forms

of knowledge production about government

performance. In particular, we examine how program

evaluation and performance management relate to

one another, using the U.S. federal government as

a case study. To the casual observer, these initiatives

Alexander Kroll

Florida International University

Donald P. Moynihan

University of Wisconsin– Madison