Multiplex Network Analysis for Complex Governance Systems Using Surveys and Online Behavior

Abstract

Text analysis, web scraping, and other computational techniques enable policy network researchers to efficiently obtain objective measures of network connections. However, the extent to which these observational methods differ from traditional survey instrument‐based measures remains an open question. Focusing on a large regional policy network of 221 organizations, this study compares a measure of collaboration generated via survey instrument to two different measures based upon internet hyperlinks and Twitter interactions between network actors. We address two questions: (1) To what extent do objective network measures based upon observed online interactions and subjective measures based upon self‐reported relationships reveal the same inter‐organizational partnerships and structural network dynamics? and (2) How useful are online network measures for supplementing survey‐based network measures? We find a significant, but substantively small, correlation between survey‐based measures and online interactions. Thus, online network measures may complement survey‐based measures, but likely reflect different aspects of the overall policy network. We conclude by discussing the potential for multiplex measures of policy networks that draw upon multiple measures to more fully understand policy network landscapes. These results bridge and help to contextualize prior work on policy network measures and virtual policy networks within the broader context of complex governance systems.

Reconceptualizing the Policy Subsystem: Integration with Complexity Theory and Social Network Analysis

“…Hayes and Scott () take this reality to task by comparing traditional survey instruments to Twitter interactions (and hyperlinks) for constructing policy networks. They point out that there is likely a cap in size for mapping a policy network via survey instruments because as the network size increase so does the number of possible ties, which increases the burden on the survey respondent’s recall (Hayes & Scott, , p. 328). This point is particularly concerning for the mapping of complex systems, which can be exceptionally large.…”

Editorial Note

“…Hayes and Scott () pursue a topic first raised in the Policy Studies Journal by Yi and Scholz (), media reports, hyperlinks, and coparticipation in collaborative venues provide alternative sources of data that perform as well as surveys for measuring policy networks. They answered in the affirmative.…”

“…Hayes and Scott () revisit this question using hyperlinks, twitter mentions, and surveys of organizations who participate in one or more collaborative groups focused on ecosystem recovery in the Puget Sound region of Washington state. Applying a range of methods (QAP test, ERGMs, and a PR test), Hayes and Scott () find that the online ties capture dimensions of the survey network, thus providing complementary data to that of surveys, but the online data are a not a substitute. The authors advise, “Although using online network methods alone to measure these policy networks would likely be erroneous in many contexts, future research that focuses on how best to use these methods in concert could open up many new opportunities to investigate how networks change in response to policy or how different network characteristics affect policy outcomes across many networks.” (p. 19).…”

“…The authors advise, “Although using online network methods alone to measure these policy networks would likely be erroneous in many contexts, future research that focuses on how best to use these methods in concert could open up many new opportunities to investigate how networks change in response to policy or how different network characteristics affect policy outcomes across many networks.” (p. 19). Developing diverse means of measuring policy networks is a critical area of research that requires much further development, and Hayes and Scott () are at the cutting edge of this effort.…”

“…(p. 19). Developing diverse means of measuring policy networks is a critical area of research that requires much further development, and Hayes and Scott (2018) are at the cutting edge of this effort.…”