Prelims

Half Title Page

CASE STUDY RESEARCH:

Core Skill Sets in Using 15 Genres

Second Edition

Title Page

CASE STUDY RESEARCH:

Core Skill Sets in Using 15 Genres

Second Edition

by

Arch G. Woodside

Boston College, Chestnut Hill, MA, USA

United Kingdom – North America – Japan – India – Malaysia – China

Copyright Page

Howard House, Wagon Lane, Bingley BD16 1WA, UK

First edition 2017

Copyright © 2017 Emerald Group Publishing Limited

Reprints and permissions service

Contact: permissions@emeraldinsight.com

No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. Any opinions expressed in the chapters are those of the authors. Whilst Emerald makes every effort to ensure the quality and accuracy of its content, Emerald makes no representation implied or otherwise, as to the chapters’ suitability and application and disclaims any warranties, express or implied, to their use.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-1-78560-461-4 (Print)

ISBN: 978-1-78560-460-7 (Online)

Preface

This expanded edition of Case Study Research (CSR) includes three new chapters that offer additional depth of coverage in case-based theory, method, and applications. The three additional chapters supplement the book’s general theme of building and testing case-based theory in describing, explaining, predicting, and controlling outcomes. The titles of the three additional chapters indicate the additional coverage in this expanded edition:

• Chapter 10: Exchange (Talk) Behavior in Natural Settings: An Exposition of Variable Based Analysis of Case Studies

• Chapter 18: Visualizing·Matching·Generalizing: Case Identification Hypotheses and Case-Level Data Analysis

• Chapter 20: Constructing Cased-Based Macro Models: Cultures’ Consequences on Entrepreneurship, Innovation, and Quality-of-Life.

Chapter 10 deepens the theory and findings on personal exchanges appearing in Chapter 9. Reading Chapters 9 and 10 may stimulate your interest in implementing additional studies on real-life face-to-face exchanges in management/marketing contexts. A substantial literature in anthropology is available on the study of talk that is usually ignored — a particularly strange phenomenon given the fields of management and marketing build from the foundational premise that two or more persons interact before, during, and post engagements in actions. “DEPC” (describing, explaining, predicting, and controlling) naturally occurring talk in management and marketing is worthy of a lifetime of research.

Chapter 18 explains the need and enables the move away from the bad practice of null hypothesis statistical testing (NHST). NHST is the biggest bad practice that dominates research practice in management today (late second decade, 21st century). Hubbard (2016) for an in-depth coverage on just how bad NHST is in management research. Problems with using NHST include its focus on an issue of little relevance to the theory understudy; NHST findings focus on the probability (e.g., p < .05) of an observed finding given that the null hypothesis that the true finding is equal to zero. If the observation provides p < .05, the researcher rejects the prior belief that the null hypothesis is true. Such testing is faulty on several grounds. A study’s sample size influences the probability for an observed finding when the null is true; with very large sample sizes (n ≥ 1000), most NHSTs result in p < .05 findings. More importantly, researchers using NHSTs rarely focus on the accuracy of model predictions for additional samples. Almost all researchers report only fit validity (e.g., the multiple regression model comes with an adjusted R ² < 0.00, p < .05); they do not test the accuracy of the model in predicting scores for cases in new samples — samples not used in constructing the model in the first place. Testing regression models for predictive validity indicate that regression models do poorly in comparison to algorithm models (Gigerenzer & Brighton, 2009) in the accuracy they achieve in predicting outcome scores for cases in additional samples.

Another telling problem, nearly all researchers using regression models ignore cases in their data sets that are contrary to the directional relationships of terms that they report in their models. Contrarian cases are observable almost always in large data sets (n ≥ 100). For example, most very frequent (top-quintile by visits) casino gamblers have high incomes relative to middle-quintile casino visitors; however, some of the top-quintile casino visitors are in the bottom-quintile by income. Cross-tabbing casino gamblers by their income and visits (in quintiles for both income and visits) indicates that cases appear in all 25 resulting cells (Woodside & Zhang, 2012). Consequently, just finding and reporting a positive income-and-casino visit relationship (b coefficient > 0.00, p < .05 in a regression model) represents shallow analysis. Additional models are necessary and useful to identify and explain the low-income (bottom quintile) casino visitors who are in the top-quintile of cases by visits as well as to identify the contrarian cases in the top-quintile of income who are in the bottom quintile of visits. Moving away from variable-based regression and NHSTs to case-based algorithms and statistical sameness testing (SST) deepens and enriches explanation and model usefulness. Rather than asking if a significant relationship between an independent (X) and a dependent (Y) variables occurs in a study, case-based modeling asks what configurations of ingredients that include high (low) scoring cases for X consistently indicate high scores in Y. The researcher moves beyond variable-based theory-analytics to embrace case theory-analytics by reframing if questions to what and when questions. Chapter 19 offers details on how researchers can release the embrace of NHST and adopt statistical sameness testing (SMT). SMT indicates how consistently a case-based model achieves the same outcome (i.e., a case-based model should indicate high scores in an outcome condition for 9 of 10 cases with high scores in the complex model statement). SST is leap up-and-beyond NHST.

Chapter 20 describes how to use configural analysis in comparative cultural research. Rather than testing the singular impact of cultural values as done in 99 percent plus studies on culture (Hofstede, 2003), Chapter 21 describes how to write complex, configurational statements as empirical models that express cultures as complex wholes. “Culture” by definition refers to a complex whole and examining the impacts of individual cultural values is not the study of culture. Chapter 21 describes and applies complexity theory and SST analytics, tools useful in many research studies beyond comparative cultural research.

The need for a source offering broad and deep coverage of theory, methods, and practice in case study research is the central premise for the original and this expanded edition of CSR. See Figure 1 for a quick scan of methods and how to classify CSR theories-methods on two dimensions: within and between cases and interpretive versus empirical positivistic stances.

Opens in a new window.

Figure 1

Four quadrants mapping case study research methods.

CSR’s principal objectives include offering nitty–gritty details of processes (steps) in building theory and designing, implementing, and evaluating a broad range of case study research methods — coverage and depth that you will not find elsewhere in one source. CSR includes comparing the criticisms and strengths of case study research with theory and methods that rely on matrix-algebra-based hypotheses testing in NHSTs. CSR includes introductions and details within lengthy examples of using qualitative comparative analysis (QCA) with available software (fsqca.com) as well as an in-depth treatment on building in degrees-of-freedom (DOF) in case study research. QCA is a Boolean-algebra-based approach for formal testing of the accuracy of complex statements of contingent relationships among recipes of antecedent conditions in predicting outcome conditions. Building in DOF analysis is often discussed in case study research but not to the depth that you find in this book — formal tests of hypotheses are possible in case studies using DOF analysis as well as other methods (e.g., latency response methods and system dynamics modeling). Students and scholars — recognizing the limited ability of informants to accurately report their own (mostly unconscious) thinking and doing processes — and the problems relating to self-editing biases of informants while answering questions — and the biases in the questions that researchers frame — frequently seek alternatives to using fixed-point survey response instruments and collecting verbal-only responses. Case study researchers worry about (insist on) achieving high accuracy in understanding, explaining, and predicting thinking and doing processes. Substantial evidence supports the view:

• Most thinking occurs unconsciously

• Humans have limited access to their own thinking-doing processes

• More than one person affects the process under examination.

Consequently, case study researchers frequently find the use of one-shot interviews with one person for each of 50–300 organizations (firm, family, or government organization) to be too low grade in accuracy and information to be acceptable. Too much nuance is missing, too much reality remains unknown, in studies that rely on one-shot interviews. Chris Rock’s (American comedian and commentator) insight automatically comes-to-mind here, “When you meet someone for the first time, you are not meeting that person, you are meeting his representative.” Explicit consideration of four dimensions is relevant for theory building and testing of processes involving humans:

• Time: micro-seconds, minutes, hours, days, weeks, months, decades, centuries

• Thinking: unconscious only, unconscious and conscious, complex problem-solving

• People: one person, group; intergroup, nation, East–West, and native cultures

• Context: home, work, travel; first-time versus repeat; weather, noise level, odors.

Figure 1 shows two of these four dimensions: time and thinking. Figure 1 includes superimposing several prevalent case-study research methods in these methods most relevant time-thinking locations. For example, according to Jung’s (1916/1959) archetypal theory, human memory include genetic primal forces that affect automatic responses to different context without conscious thought — behavioral and thinking outcomes of responses learned over thousands of years. Gigerenzer (2008) may be the most insightful scholar currently working on examining meta-thinking issues. Note that Figure 1 attempts to show communication flows between the various levels of thinking. While recognizing that conscious and unconscious thinking and thoughts occur separately is useful theoretically, much thinking likely includes some bits of both conscious and unconscious thoughts (Evans, 2008). “Go deep!” Go deep by both going into the field — real-life contexts and by learning a variety of case study theories and methods is the advice and direction that this book directs you to follow. This suggestion is a corollary to Weick’s (1979) famous suggestion, “Complicate yourself!”

The complicated individual can sense variations in a larger environment, select what need not be attended to, what will not change imminently, what won’t happen, and by this selection the individual is able to amplify his control variety. He safely (that is, insightfully) ignores that which will not change, concentrates on that which will, and much like the neurotic psychiatrist is able to anticipate significant environmental variation when and where it occurs. Complicated observers take in more. They see patterns that less complicated people miss, and they exploit these subtle patterns by concentrating on them and ignoring everything else. (Weick, 1979, p. 193, italics in the original)

This expanded CSR book offers deep coverage of 16 case research methods; the closing chapter offers 12 specific principles to implement to increase accuracy of what is happening and what will happen in real-life contexts-processes involving thinking and behavior by humans. Useful tools for going deep and for complicating yourself!