A Systematic Review of Public Sector Innovation Studies : Combining Topic Modeling and Bibliometric Analysis

Research output: Conference PapersRGC 32 - Refereed conference paper (without host publication)peer-review

View graph of relations



Original languageEnglish
Publication statusPublished - Jan 2020


Title5th Innovation in Public Services and Public Policy Conference, PUBSIC 2020
LocationUniversity of Stavanger
Period29 - 31 January 2020


Since the 1960s, public sector innovation has received increasing attention from both scholars and practitioners, many of whom believe that it helps improve public service quality and enhance public organizations’ governance capacity (Damanpour and Schneider 2009; Osborne and Brown 2011; Walker 2014). After over five decades of development, the research field has accumulated a rich and heterogenous array of concepts, theoretical approaches, and empirical results. To advance knowledge accumulation and field development, we need to systematically review the existing literature and examine shared knowledge in this field(Cole 1983; Kuhn 1962). Specifically, what topics have been addressed in the innovation studies? To what extent do scholars have knowledge consensus, and what predicts the varying degree of knowledge consensus?

Existing studies have reviewed innovation studies either conceptually or empirically, seeking to understand public sector innovation and summarize the existing empirical research (Bason 2010; Osborne and Brown 2011; De Vries, Bekkers, and Tummers 2015; Walker 2014). While these reviews provide details and clarity of existing innovation literature, they mostly use preconceived taxonomy as a guiding framework for their analysis and pay limited attention to knowledge consensus and field identity.

In this study, we apply topic modelling and bibliometric analysis to analyse 2643 peer-reviewed articles in innovation (1969-2019) in political science and public administration disciplines from Web of Science. Using topic modelling technique (Blei and Lafferty 2007), we analyse the titles and abstracts of these innovation studies, identify 100 different topics, and illustrate their evolution over time. Following Kessler’s (1963) bibliometric coupling concept, we develop a knowledge consensus index for each article by calculating how many articles in the two disciplines share at least one reference with it. We will further examine how various factors, such as target audience of the articles (general or specific), innovation focuses, and other control variables (e.g., author collaboration, number of citations, number of references, etc.) are associated with the degree of knowledge consensus.

Preliminary findings indicate that articles targeting a broader audience are likely to have more knowledge consensus. That is, studies examining general policy and management issues have a more shared identity than those that look into specific contexts (policy areas or regional contexts). Overall, using novel analytical techniques, our study contributes to an understanding of what constitutes and affects the shared identity of the field of public sector innovation.

Bibliographic Note

Research Unit(s) information for this publication is provided by the author(s) concerned.

Citation Format(s)

A Systematic Review of Public Sector Innovation Studies: Combining Topic Modeling and Bibliometric Analysis. / Walker, Richard; Chandra, Yanto; Zhang, Jiasheng et al.
2020. Paper presented at 5th Innovation in Public Services and Public Policy Conference, PUBSIC 2020, Stavanger, Norway.

Research output: Conference PapersRGC 32 - Refereed conference paper (without host publication)peer-review