Introducing Neural Computing in Governance Research: Applying Self‐Organizing Maps to Configurational Studies

• Date: 01 November 2017
• Author: Mark John Somers,Jose Casal
• DOI: http://doi.org/10.1111/corg.12173
• Published date: 01 November 2017

SPECIAL ISSUE ARTICLE

Introducing Neural Computing in Governance Research:

Applying Self‐Organizing Maps to Configurational Studies

Mark John Somers |Jose Casal

School of Management, New Jersey Institute

of Technology, Newark, New Jersey 07102,

USA

Correspondence

Mark John Somers, School of Management,

New Jersey Institute of Technology, Newark,

NJ 07102 USA.

Tel: 973‐596‐3279

Email: mark.somers@njit.edu

Abstract

Manuscript Type: Empirical

Research Issue/Question: Self‐organizing maps (SOMs), a neural computing paradigm,

were introduced as a methodology to enhance and extend configurational governance research.

The capabilities of SOMs include assessment of nonlinear relationships among study variables

and projection of firms and clusters in two‐dimensional space based on their relative similarity.

Research Findings/Insights: To demonstrate their application to governance research,

SOMs were used to study patterns of immunity to institutional governance logics in the financial

services industry. Firm sensemaking and governance logics were assessed by analyzing the lan-

guage and meaning of corporate codes of conduct. Content analysis was guided by the DICTION

software program. DICTION uses data dictionaries to analyze the meaning of text documents

based on word usage. Our results supported a configurational model characterized by distinct

groupings of firms with varying degrees of acceptance of prevailing institutional governance logics.

Practitioner/Policy Implications: SOM analysis demonstrated that context influences firm

governance logics. Specifically, different interpretations of environmental pressures led to different

adaptive responses suggesting reconsideration of the notion of universal or best governance practices.

KEYWORDS

Corporate Governance, Self‐Organizing Maps, Neural Computing, Governance Logics, Financial

Services Industry

1|INTRODUCTION

Self‐organizing maps (SOMs) are a neural computing paradigm that

identifies orthogonal clusters within a dataset (Kohonen, 1990). They

have been used in variety of academic disciplines to study a wide range

of problems. Specifically, SOMs have been used in climate science

research (Hewitson & Crane, 2002), medical research (Villmann,

Hermann, & Geyer, 2000), the geographic information sciences

(Agarwal & Skupin, 2008), image classification research (Richardson,

Risien, & Shillington, 2003), the environmental sciences (Cartwright,

2002), engineering (Bossio, De Angelo, & Bossio, 2013), and physics

(Camplani, Cannas, Fanni, Pautasso, & Sias, 2011).

SOMs are relevant to governance research because they have the

potential to open up new lines of inquiry in configurational studies.

Configurational studies represent a new direction in governance

research in which contextual variables are hypothesized to influence

governance models and practices (Filatotchev & Boyd, 2009). This

influence takes the form of differential adaptive patterns expressed

as distinct bundles of firms, thereby raising the possibility that the

assumption of universalism in governance research needs to be

reconsidered (Garcia‐Castro, Aguilera, & Arino, 2013; Schiehll,

Ahmadjian, & Filatotchev, 2014; Ward, Brown, & Rodriguez, 2009).

Although configurational governance studies area promisingdevel-

opment, they require appropriate research methodologies for their

potential to berealized. Specifically, because configurational research is

based on the premise that there are distinct, meaningful groupings of

firms withina larger entity, clustering methodsare central to this avenue

of inquiry(Garcia‐Castroet al., 2013; Judge, Fainshmidt,& Brown, 2014).

In this paper, SOMs are introduced as a new methodology that can

advance and extend configurational governance studies. Given that

neural computing is likely to be new to many governance scholars,

we explain the SOM algorithm and compare it with traditional cluster-

ing methods. An application of an SOM based on firms' adaptation to

environmental pressures in the financial services industry following

Revised: 10 June 2016

DOI: 10.1111/corg.12173

440 © 2016 John Wiley & Sons Ltd Corp Govern Int Rev. 2017;25:440–453.wileyonlinelibrary.com/journal/corg