Network Effects Research: A Systematic Review of Theoretical Mechanisms and Measures
| Author | Alejandra Medina,Michael D. Siciliano,Weijie Wang,Qian Hu |
| DOI | http://doi.org/10.1177/02750740221118825 |
| Published date | 01 October 2022 |
| Date | 01 October 2022 |
| Subject Matter | Articles |
Network Effects Research: A Systematic
Review of Theoretical Mechanisms and
Measures
Alejandra Medina
1
, Michael D. Siciliano
1
,
Weijie Wang
2
and Qian Hu
3
Abstract
This article contributes to the network effectiveness literature by identifying the theoretical mechanisms and network mea-
sures scholars in public administration and policy use to draw inferences between network structures and network effects.
We conducted a systematic review of empirical network effects research in 40 public administration and policy journals from
1998 to 2019. We reviewed and coded 89 articles and described the main social theories used in the network effectiveness
literature and the associated mechanisms that translate network structures to network effects. We also explain how scholars
operationalize those theoretical mechanisms through network measures. Overall, our findings reflect that there is limited use
of social theories for the explanation of network effects and in some cases, inconsistent use of network measures associated
with theories. Moreover, we identify several challenges confronting network effects research. These challenges include the
difficulty of isolating specific mechanisms related to a particular social theory, the use of network structures both as a mech-
anism and as a measure, and the lack of data to examine network dynamics and coevolution.
Keywords
systematic review, network effects, theoretical mechanisms, network structures, network measures
Introduction
Networks have long been viewed as a suitable form of orga-
nizing to address complex policy problems and coordinate
public service delivery (Provan & Milward, 2001).
However, much of the empirical research on the effectiveness
of networks in public settings did not begin until the 1990s.
Since that time, network research has grown tremendously.
There exist relevant frameworks to assist researchers in
assessing network performance (Herranz, 2010; Provan &
Milward, 1995; Raab et al., 2015) along with syntheses of
the extant network effectiveness literature (Cristofoli &
Markovic, 2016; Kenis & Provan, 2009; Turrini et al.,
2010). Despite this growth, there remains a lack of under-
standing as to how network structures determine outcomes.
Therefore, it is important to identify the specific mechanisms
associated with different relational configurations that
produce outcomes for the network members or the whole
network (Hedström, 2008).
Provan and Milward (2001) identified two critical chal-
lenges confronting research on network effectiveness: (i)
how to reach consensus on network goals and performance
metrics and (ii) how to identify the primary aspects of
networks that may influence performance. Our study is con-
cerned with the latter challenge. We conducted a systematic
review of the network literature in public administration
and policy to identify the theoretical mechanisms and
network measures scholars use to draw inferences
between network structures and network effects. Our sys-
tematic review covers network research for 21 years
(1998–2019) in 40 public administration and policy jour-
nals. In this research, we focus on network effects in a
broad sense and define them as the desired impacts or con-
sequences that a network may have on participating actors
or the system itself. Besides “network effects,”in this man-
uscript, we also use the terms “network effectiveness”and
“network performance”as they are used in the reviewed
articles.
1
Department of Public Administration, University of Illinois at Chicago
College of Urban Planning and Public Affairs, Chicago, IL, USA
2
Truman School of Government and Public Affairs, University of Missouri,
Columbia, MO, USA
3
School of Public Administration, University of Central Florida, Orlando, FL,
USA
Corresponding Author:
Alejandra Medina, Department of Public Administration, University of
Illinois at Chicago College of Urban Planning and Public Affairs, 400 S
Peoria, Chicago, IL, USA.
Email: amedin54@uic.edu
Article
American Review of Public Administration
2022, Vol. 52(7) 513–528
© The Author(s) 2022
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/02750740221118825
journals.sagepub.com/home/arp
This systematic review includes empirical network effects
research at the node and system levels. Network effects
research at the node level analyzes the extent to which the
characteristics of an individual’s network affect the desired
outcome (Raider & Krackhardt, 2017). For instance,
Maroulis (2017) analyzed if teachers with more connections
among different subgroups within their network are more
likely to present innovative behaviors. Network effects
research at the systems level considers aggregate and collec-
tive outcomes. Thus, system-level effects would include both
network-level and community-level outcomes as defined by
Provan and Milward (2001). Network effects research at
the system-level studies how the characteristics and structure
of the network affect aggregate outcomes for the community
served by the network or for the functioning of the network
itself (Provan & Milward, 2001). For example, Lee (2013)
compared two local governments and found that dense net-
works lead to a greater perception of e-government effective-
ness because information and knowledge flow faster in
denser structures.
This article contributes to the network effects literature in
three ways. First, we provide a concise description of the the-
ories and the associated mechanisms that translate network
structures into network effects. This description and review
distill the primary mechanisms by which our field connects
network structure to performance at the node and system
levels. Overall, we find that there is limited use of social the-
ories for the explanation of network effects. Second, we
examine how scholars operationalize those mechanisms
through different nodal and network measures. We find an
inconsistent use of measures associated with specific social
theories as scholars often use the same network measure to
operationalize different theoretical mechanisms. We further
analyze which network measures are most closely associated
with particular mechanisms and offer suggestions for future
research. Third, we identify the main challenges related to
the extant research on network effects. These challenges
include the difficulty of isolating particular mechanisms
when multiple mechanisms are associated with a given
theory or outcome and the lack of suitable data to study the
coevolution of network structure and effects. In the following
section, we explain the methodology used for the systematic
review and then discuss the main theories and network mea-
sures used in the network effects literature.
Systematic Review: Methodology
To examine the theoretical mechanisms explaining network
effects in public administration and policy, we reviewed
empirical articles about public sector networks published
from January 1998 to May 2019. To identify relevant
network effects articles, we followed the slightly modified
PRISMA protocol (Moher et al., 2009) used by Siciliano
et al. (2021) and Kapucu et al. (2017). We conducted a
general search in 40 main journals of public administration
and policy.
1
Our search and exclusion process consisted of
five steps. First, we searched on the websites of each
journal for articles that included in the title, abstract, or key-
words the following terms “network,”“network analysis,”
“collaboration,”or “collaborative.”This first step resulted
in a total of 2,402 articles. Then, we reviewed the abstracts
of these articles to ensure they were empirical network
studies, including descriptive articles, comparative case
studies, and articles that use inferential methods. We
excluded purely conceptual articles and network studies in
which no network data were collected. This process excluded
1,340 articles and included 1,062 articles. Third, we reviewed
the methodology sections of the articles to verify that the
authors use social network analysis methods or included
network measures. A total of 282 articles met this criterion,
and 780 articles were removed. Fourth, we confirmed the arti-
cles were about public sector networks and removed studies
concerning private networks (e.g., articles about engineering
firms, the airline industry, or high-tech companies); 196 arti-
cles met this criterion. Finally, since we were interested in
understanding the effects of networks, we removed 107 arti-
cles that use the network as the dependent variable. Articles
that treat the network as the dependent variable focus on ana-
lyzing network processes like tie formation and dissolution or
how a specific structural property is formed. Thus, we kept 89
articles that included networks as independent variables and
analyzed how different network structural properties impact
performance or outcomes at the node or system levels.
For the analysis of the 89 articles, we designed a compre-
hensive coding protocol to extract specific information on the
type and number of nodes in the network, type of ties, method
of network data collection, primary level of analysis, area of
study, research method, each network-related hypothesis and
its associated use of theory and measures, and the different
factors identified by the authors as drivers for performance.
We also reviewed how authors conceptualized and measured
(i) the dependent variable related to network effects, (ii) the
use of perceptual or objective measures, (iii) whether the
network effect variable was an output or a longer-term
outcome, and (iv) the level of analysis of the dependent var-
iable. Based on this coding we remove 15 articles from
further analysis due to a lack of clear conceptualization and
measurement of the network effects. Therefore, the final
number of articles included in the analysis was 74.
Prior to coding, we initiated a pilot test where each of the
four authors coded ten articles to compare and develop con-
sistency in coding and adjust the coding protocol where
needed. Once we concluded the pilot test and finalized the
coding protocol, we created two teams with two coders
each and coded all articles. Since we were interested in iden-
tifying the use of theories and network measures in the
network effects literature, we conducted open coding of
each hypothesis, the theoretical mechanism used by the
authors to frame and support the hypothesis, authors’descrip-
tion of the theoretical mechanisms illustrating the
514 American Review of Public Administration 52(7)
To continue reading
Request your trial