A Composite Index to Measure Integration of Global Business Citizenship

A Composite Index to Measure Integration of

Global Business Citizenship

Section 1: Foundation of the Study

Corporate Social Responsibility (CSR) is a broad concept consisting of the

economic, legal, ethical, and philanthropic responsibilities of corporate leaders (Carroll,

1999). Leaders of multinational or global corporations take CSR seriously. Business

Citizenship focuses on the ethical responsibly of corporate leaders as informed by CSR

(Carroll, 1999). Ninety-seven percent of the top 100 U.S. companies operate in multiple

nations and claim to be business citizens (Fifka, 2013). Wood and Logsdon developed

the theory of Global Business Citizenship (GBC) in 2002. GBC is defined as “a business

enterprise (and its managers) that responsibly exercises its rights and implements its

duties to individuals, stakeholders, and societies within and across national and cultural

borders” (p. 4). The definition implies that these rights fall within the ethical

responsibilities of CSR. By 2006, Wood, Logsdon, Lewellyn, and Davenport expanded

the GBC theory into a four-step implementation process. The first step is to develop a

companywide overarching code of ethical conduct consistent with the definition of GBC

(VALUE). The second step is to implement the overarching code of conduct throughout

the organization and adapt it to local customs, norms, and ethical standards (IMPLE).

The third step is to analyze problem areas and experiment with solutions to remediate the

conflicts (ANALY). The fourth step is to systemize learnings from the IMPLE and

ANALY steps and institutionalize the best policies, practices, and behaviors throughout

the organization. (LEARN). These four steps are the principles of the theory of GBC. As

leaders implement each step, their companies are maturing as global business citizens.

Companies that demonstrate these four steps are following the GBC principles.

Numerous research articles referenced and built upon the GBC theory. Current

researchers continue to cite the theory (Aguinis & Glavas, 2012; Font, Walmsley,

Cogotti, McCombes, & Hausler, 2012; Godos-Díez, Fernández-Gago, &

MartínezCampillo, 2011; Hart & Sharfman, 2012; Hemphill & Lillevik, 2011). Research

citing and building upon the GBC theory indicates that researchers consider it a valid

theory. The challenge facing researchers and business practitioners is the absence of a

rating system to determine whether a company meets the definition of a business citizen

and a composite index to measure the maturity level a company has achieved toward

becoming a business citizen. The key lies in the theory of GBC. The theory simply

provided the framework, or constructs needed to assess the maturity level a company has

achieved toward becoming a business citizen.

The ultimate objective of this research was to develop a composite index to

evaluate the maturity level that a company has attained in implementing the four steps of

the GBC framework. From a methodological viewpoint, the first step was to develop a

survey with a superset of questions that capture the multiple dimensions, or complexities,

of each step of the four-step framework developed by Wood et al. (2006). In quantitative

statistics, the steps of implementing the GBC theory equate to constructs. Donna J. Wood

was the lead researcher of the GBC theory. Wood agreed that VALUE, IMPLE,

ANALY, and LEARN could be considered the four constructs of the GBC theory (D. J.

Wood, personal communication, August 14, 2014; Appendix E).The second step was to

administer the survey to a qualified sample of business executives with knowledge and

understanding of GBC. The third step was to feed the collected sample into a powerful

statistical technique called exploratory factor analysis (EFA). The EFA method

rationalized the superset of questions into a minimal subset of questions necessary to

capture each construct. EFA generates so-called eigenvalues, or weights, on each

question. The result was a consensus survey with weights on each question. The final

step was to use the eigenvalues to determine an overall score or representative index of

global business citizenship to measure each company.

Researchers have made substantial progress in developing instruments to measure

various aspects within the broad field of corporate social responsibility. The extensive

literature review for this study and a review of the literature performed by Wood in 2010

did not reveal a measure for business citizenship, corporate citizenship, or any derivation

or definition of the term. Without a measure, it is difficult to know if a company meets

the societal expectations of a global business citizen or is merely making the claim

(Shinkle & Spencer, 2012).

Background of the Problem

For 60 years, researchers have studied the social responsibilities of business

leaders (Carroll, 1999). Research shows that companies that responsibly exercise their

rights and implement their social duties may obtain legitimacy from society and increase

competitive performance (Menck & Oliveir, 2014). Legitimacy and increased

competitive performance contribute to maximizing shareholder value. As such, there was

no longer a question of whether leaders should integrate social responsibility into their

business strategies, but how (Crittenden, Crittenden, Piney, & Pitt, 2011; Shepherd,

2014). The theory of GBC contends that the answer is for companies to become global

business citizens (Wood et al., 2006).

The concept of global business, or corporate, citizenship has rapidly gained

popularity in the corporate, academic, and political arenas (Crittenden et al., 2011). In the

corporate world, major companies, such as Boeing, Dow, IBM, and Microsoft all claim

that they are business citizens (Crittenden et al., 2011). Universities are incorporating

global business citizenship into their curricula (Lilley, Barker, & Harris, 2014). The

Clinton Global Initiative recognizes companies with the Global Citizenship Award

(Clinton Global Initiative, 2013). Despite the popularity of business citizenship to

describe the ethical component of social responsibility, at the start of this study there was

no published and publicly available rating system to measure the level of global business

citizenship for a given company.

Problem Statement

In 2014, over 8,000 business leaders across 145 countries had signed the United

Nations Global Compact to demonstrate their corporations’ commitment to ethical values

and responsibilities within society (Ortas, Alvarez, & Garayar, 2015). Business

citizenship has emerged as the preeminent term to describe the ethically responsible roles

of corporations within society (Crittenden et al., 2011). Of the top 100 U.S. companies,

97% claim to be business citizens (Fifka, 2013). The general business problem was that

there was no self-administered rating system available for business leaders to report to

stakeholders the steps they had achieved toward becoming an ethically responsible

business citizen (Milne & Gray, 2013). The specific business problem was that there was

no published self-administered survey instrument, or composite index derived from such

a survey, to assess the maturity level a company had achieved toward becoming a

business citizen as defined by the GBC theory (Wood, 2010).

Purpose Statement

The purpose of this cross-sectional quantitative study was to develop Likert

survey questions (independent variables) and apply EFA to reveal factors (dependent

variables) and assign weights to questions to develop a self-administered rating system to

measure the GBC theory, which assesses the maturity level a company has attained

toward becoming a global business citizen. Development of this rating system required

four methodological steps. First, creating a survey consisting of 1 qualifying question and

22 Likert questions that are operationalizing the GBC principles of VALUE, IMPLE,

ANALY, and LEARN. Second, administering the survey to members of professional

associations who were senior executives of U.S. corporations with an understanding of

GBC. Third, applying the EFA statistical method to the data. EFA revealed the

relationship between the Likert survey questions and the factors that emerged, reduced

the questions, and assigned weights to the remaining questions. Fourth, use the EFA

assigned weights to develop a composite index. The result of this study provides a rating

system to measure a company’s GBC maturity level. This study contributes to social

change by providing practitioners, academics, and stakeholders with a rating system to

evaluate the maturity level that corporate leaders have attained toward becoming a global

business citizen.

Nature of the Study

The goal of this research was to operationalize the four high-level GBC theory

steps, or constructs, elaborated by Wood et al. (2006), into a useful survey instrument

and weighted index. These artifacts allow practitioners, academics, and stakeholders to

selfadminister a quantitatively validated survey to corporate leaders to evaluate the

maturity level they have attained in implementing the four steps of the GBC theory. The

results of the survey provide a rating system in the form of a composite index to measure

the maturity level corporate leaders have achieved toward becoming a global business

citizen.

Applying a quantitative, qualitative or even a mixed methodology could have

produced an instrument and weighted index. There are advantages and disadvantages to

each methodology. The critical decision in selecting a quantitative method for this

research was the fact that the method must be appropriate for construct operationalization

to render a weighted index based on the scores attained by the survey instrument.

Researchers often use the qualitative methodology to describe and explain a phenomenon

and discover relevant concepts to propose a theory (Gioia, Corley, & Hamilton, 2012).

Once researchers identify concepts and propose a theory, they use the quantitative

methodology to formulate constructs. These constructs are qualities that researchers can

operationalize and quantify as variables for the purpose of measurement (Gioia et al.,

2012). Because the goal of the research was to operationalize and measure the constructs

of the previously developed theory of GBC, a quantitative methodology was most

appropriate.

Of the numerous quantitative methods available, EFA was appropriate for this

study. Researchers commonly use EFA to develop and validate self-reporting assessment

instruments, especially when there is little or no a priori knowledge of the structural

model (Ruscio & Roche, 2012). The EFA quantitative method was appropriate because it

is a rigorous statistical approach. EFA provided an unbiased method for reducing the

number of factors, examining relationships between factors, and evaluating the construct

validity of a measurement scale (Williams, Brown, & Onsman, 2012). Measurements

should repeatedly produce the same results, the measurement should be stable over time,

and the measurements should be similar within a given period (Hasson & Keeney, 2011).

Using EFA should ensure reliability by establishing construct, content, and statistical

validity. The EFA technique established weighted factors that loaded to the constructs,

which described the theoretical framework, thereby ensuring construct validity. The EFA

technique determined content validity. Applying EFA demonstrated that the measure

covered the range of meanings associated with the constructs (MacKenzie, Podsakoff, &

Podsakoff, 2011; Oluwatayo, 2012; Podsakoff, MacKenzie, & Podsakoff, 2012).

Computing the scree test to determine the optimal number of factors to retain ensured

statistical validity. To mitigate time and expense constraints, five experts with the

familiarity of corporate or business citizenship assessed the face validity of the survey.

The experts determined that the questions were reasonable, unambiguous, and clear,

(Oluwatayo, 2012; Synodinos, 2003). Figure 1 depicts the EFA method to evaluate GBC

integration.

Figure 1. Schematic of EFA method to evaluate GBC integration.

The four-step GBC framework provided the assumed constructs. The constructs

were the following. First, developing overarching corporate values (VALUE). Second,

implementing the values (IMPLE). Third, analyzing problems and experimenting to

revise the values or local implementations (ANALY). Fourth, learning from the previous

steps and institutionalizing best practices (LEARN) (Wood et al., 2006). In Figure 1, the

black arrows indicate the survey questions that may describe each construct. EFA applied

to the data collected by administering the survey removed redundant questions and

assigned scores (called eigenvalues) to each question. The result is the minimal set of

questions that are necessary for the final survey and weighted factors. At a minimum, one

question should load to each construct. The blue arrows indicate this minimal question,

but the final research may yield more than one. When company leaders complete the

resulting survey, the resulting data yields weighted scores, indicating the maturity level

that they have attained in implementing the four-step GBC framework. In summary, EFA

statistical analysis was not only appropriate but also necessary to determine the critical

factors and their weightings that should comprise a standard rating system in the form of

a composite index to measure GBC.

Researchers commonly use Likert-type survey instruments to collect data for the

EFA quantitative method (Harrison & Reilly, 2011; MacKenzie, Podsakoff, &

Popsakoff, 2011). A Likert scale is a 5- or 7- point ordinal scale used to measure the

degree to which participants agree or disagree with a statement (Sullivan & Artino,

2013). Because archival data to support this type of research was not available, a survey

design was appropriate. Furthermore, a cross-sectional survey design was best suited to

collect the significant amount of data necessary for EFA (Beavers, Lounsbury, Richards,

Huck, Skolits, & Esquivel, 2013). The actual survey consists of 23 questions. The first

question was a “yes” or “no” qualifying question. The 22 Likert-type questions were

designed to attempt to capture the four constructs of the GBC theory. The study was

cross-sectional, meaning the sample represented a cross-section of the population for

which the measure was designed (MacKenzie et al., 2011). In this case, executive leaders

of multinational business organizations in the United States were the intended population

for this instrument. This cross-sectional sample should represent the population so that

the results should generalize to the broader population (MacKenzie et al., 2011).

Research Question

The purpose of the research was to develop a rationalized survey consisting of the

minimal set of survey questions required to assess the step or steps corporate leaders

have achieved in implementing the four steps of becoming a global business citizen. A

subsequently weighted index based on such questions evaluated the maturity level that

corporate leaders have attained toward becoming a global business citizen as defined by

the theory of GBC. Fundamentally, the research aimed at answering the following key

research question:

RQ: How many and what factors (dependent variables) are needed to

characterize the Likert survey questions (independent variables) to assess a company’s

GBC maturity level?

For an EFA study, such an overarching research question can be broken down

into four sub-questions given the fact that the GBC theory stipulates four key steps or

constructs:

Sub-Research question1 (SRQ1): Does the survey adequately capture the VALUE

construct of the GBC theory?

Sub-Research question 2 (SRQ2): Does the survey adequately capture the IMPLE

construct of the GBC theory?

Sub-Research question 3 (SRQ3): Does the survey adequately capture the

ANALY construct of the GBC theory?

Sub-Research question 4 (SRQ4): Does the survey adequately capture the

LEARN construct of the GBC theory?

Hypotheses

Multiple hypotheses could have been stated to operationalize the overarching RQ

and sub-research questions SRQ1-SRQ4. However, EFA is not an inferential statistical

technique and therefore when using this technique, researchers cannot stipulate

inferential hypotheses (Beavers et al., 2013). Instead, when using EFA, researchers often

stipulate so-called propositions about the number of factors to retain to capture the

relevant constructs. Given this general practice, researchers can state the propositions as

how many factors are required to represent the survey and the nomenclature k-factors to

stipulate such propositions (Henson & Roberts, 2006).

The GBC theory detailed four steps that leaders must implement to become a

global business citizen. These four steps are the GBC principles and equate to the

constructs of VALUE, IMPLE, ANALY, and LEARN (D. J. Wood, personal

communication, August 14, 2014; Appendix E). The EFA analysis revealed how many

factors emerged to best answer the question of whether a company meets the definition

of a global business citizen. The EFA analysis also yielded a measure of the maturity

level a company has achieved toward becoming a global business citizen by identifying

the step or steps leaders have implemented. Before the EFA technique, it was unknown

how many factors would emerge to answer the question of whether a company meets the

definition of a global business citizen and measures the maturity level a company has

achieved. The correct hypothesis centered around how many k-factors characterize

companies following the GBC principles. Mathematically, the following proposition

equated to the required hypotheses:

H0k: k= number of factors needed to characterize companies following GBC

principles.

Theoretical Framework

The theory of Global Business Citizenship was the basis of the theoretical

framework of this quantitative study. Researchers Logsdon and Wood developed the

GBC theory in 2002 (Logsdon & Wood, 2002; Logsdon & Wood, 2005; Wood &

Logsdon, 2002). Wood and Logsdon used a deductive, descriptive typology to develop

their theory of GBC (Wood et al., 2006). In 2006, together with Lewellyn, and

Davenport, Wood and Logsdon further developed their theory into a framework (Wood

et al., 2006). The theory describes a four-step framework. Figure 2 depicts the four-step

framework for implementing GBC within an organization. The first step was to develop a

companywide overarching code of ethical conduct consistent with the definition of GBC

(VALUE). The second step was to implement the overarching code of conduct

throughout the organization and adapt it to local customs, norms, and ethical standards

(IMPLE). The third step was to analyze problem areas and experiment with solutions to

remediate the conflicts (ANALY). The fourth step was to learn from the previous steps,

systemize, and institutionalize best practices (LEARN). These four steps are the GBC

principles. Companies, academics, and politicians are using the term “corporate

citizenship” to describe the socially responsible role of businesses (Clinton Global

Initiative, 2013; Crittenden et al., 2011; Lilley et al., 2014). Because stakeholders are

using the term, the GBC theory fits to inform whether companies are business citizens.

Figure 2. The four-step framework for implementing GBC.

Operational Definitions

Corporate Social Responsibility. “Context-specific organizational actions and

policies that take into account stakeholders’ expectations and the triple bottom line of

economic, social, and environmental performance” (Aguinis & Glavas, 2012).

Global Business Citizen. “A global business citizen is a business enterprise

(including its managers) that responsibly exercises its rights and implements its duties to

individuals, stakeholders, and societies within and across national and cultural borders”

(Wood et al., 2006, p. 35).

Globalization. As a result of declining costs, the process of free trade across

borders connecting and transferring capital, goods, and people at distant locations (Bond

& O’Byrne, 2014). When referring to a “global business citizen” or “global corporation”,

globalization implies that the entity considers the entire world as a single space (Bond &

O’Byrne, 2014).

Stakeholder. “Any person, group, or organization who can affect or is affected by

the organization’s actions. Traditionally, a company’s stakeholders include investors,

employees, customers, suppliers, and the local communities. Others—governments,

NGOs, activists, the media—are also considered stakeholders today” (Wood et al., 2006,

p. 11).

Assumptions, Limitations, and Delimitations

The basis of the research for this study was the formation and administration of a

Likert-type scale instrument designed to identify the key questions that were significant

indicators of the four steps, or constructs, of GBC. The below assumptions, limitations,

and delimitations frame the study.

Assumptions

Assumptions are ideas that seem self-evident and are taken-for-granted as true

(Jansson, 2013). Researchers make assumptions about methods, design, and data (Leech,

Onwuegbuzie, & Combs, 2011). Several key assumptions underlie this research study.

The first was that the theoretical framework was measurable. The second was that

operationalization of the constructs captured the four-step GBC framework. The third

was that the instrument measured what it was intended to measure, also known as

construct validity. The fourth was that participants had sufficient knowledge of GBC to

complete the survey. The fifth was that the selected sample represented medium to large

national and multinational companies. The sixth assumption was that a survey approach

and EFA was appropriate for this study

The first assumption was that the theoretical framework was measurable. The

assumption was that a combination of some or all of the four key constructs might

measure the maturity level a company has achieved toward becoming a global business

citizen. The constructs were VALUE, IMPLE, ANALY, LEARN. This assumption may

have been limiting because this study did not consider constructs outside these four.

Mitigating the risk of this being a limiting assumption was the fact that these four

constructs are the basis of the GBC theory. Inherently, it was safe to assume that the

GBC theory included all of the constructs necessary to capture the stated definition of

business citizenship. The assumption was that the GBC theory was necessary and

sufficient for assessing an organization’s level of maturity as a global business citizen.

The second assumption was that the operationalization of the constructs into

survey questions accurately captured the four-step GBC framework. That is, the

assumption was that the survey’s 22 Likert questions adequately covered the majority of

issues that contribute to each of the four steps of the GBC framework. Three items

mitigated this potential limitation. First, the literature informed the constructs and the

specific questions. Second, the lead GBC theory author communicated her agreement.

Third, five subject matter experts (SME’s) evaluated the questions. Specifically, the

literature included the seminal studies in which Donna J. Wood and Jeannine M.

Logsdon developed the theory of GBC and the book published by the authors of the

theory of

GBC. In a personal email communication, Donna J. Wood agreed that VALUE, IMPLE,

ANALY, and LEARN captured the process of implementing the four-step GBC

framework (D. J. Wood, personal communication, August 14, 2014; Appendix E).

Additionally, five GBC subject matter experts (SMEs) reviewed the questions in the

newly developed instrument and agreed that the questions are clear and concise

(Appendix F). Applying EFA to the survey results reduced the number of questions. This

reduction identified the critical questions needed to evaluate the maturity level that a

company has attained in implementing the four-step GBC framework. The EFA process

also mapped the questions to their underlying factors, thus enabling development of a

composite index.

The third key assumption was that the instrument measured what it was intended

to measure, also known as construct validity (Cronbach & Meehl, 1955). The goal of the

research studies is for investigators to confirm or disconfirm that the instrument measures

what the investigator hypothesized it would measure (Cronbach & Meehl, 1955).

Cronbach and Meehl (1955) proposed three steps to evaluate construct validity. The

following is a summary of how instrument validity was determined. The Instrument

Validity subsection under the Data Collection section details these steps. In summary, the

first was to state the theoretical framework and assign meaning to each construct.

VALUE, IMPLE, ANALY, and LEARN were the constructs. These constructs were

informed by the literature and personal communication from Donna J. Wood, principal

researcher of the GBC theory (D. J. Wood, personal communication, August 14, 2014;

Appendix E). The survey questions were an attempt to assign meaning to each construct.

Five SME’s indicated that they felt the questions captured the meaning of the constructs

and were clear (Appendix F). The second step was to develop methods and empirically

measure how adequately the instrument substantiated the assigned construct meanings.

Some methods for examining construct validity exist. One was the multitraitmultimethod

matrix (MTMM), described by Campbell and Fiske's landmark paper

(1959). Others include factor analysis, and structural equation modeling (Marsh, Morin,

Parker, & Kaur, 2014). Using EFA satisfied this assumption. The third step was to

interpret correlations and present evidence and reasoning to show the reader why the

correlations confirmed or disconfirmed the hypothesis. EFA was the heart of this

research. Once the research was complete, the data was analyzed using the EFA

technique. Interpretation of the correlations provided evidence and reasoning to show

why the correlations confirmed or disconfirmed the hypothesis. In the case of applying

EFA, the hypothesis equates to propositions about the number of factors to retain to

capture the relevant constructs. In conclusion, the assumption of the construct validity of

the instrument was strongly satisfied via the original theory author’s confirmation, SME

confirmation, and EFA application.

The fourth assumption was that the participants possessed sufficient knowledge of

GBC to complete the survey. This assumption could have been a limiting factor because

surveys completed by participants with insufficient knowledge of GBC may not reflect

answers about which items suggest the four constructs of GBC. The first question was

designed to mitigate this potential limitation. The first question was, “I am familiar with

the concept of corporate citizenship, business citizenship, corporate social responsibility,

or the ethical responsibilities of corporations”. Participants indicated “yes” or “no” to this

question. The online survey directed participants answering “no” to this first qualifying

question to the end of the survey without answering the Likert questions. The online

survey continued to the Likert questions for participants answering “yes” to the question.

The fifth assumption was that the business leaders assembled at the Executive

Suite professional business society represent the business leaders of medium to large

national and multinational corporations. By definition, a cross-sectional sample implies

that it was representative of medium and large companies, but it may not have been true

given that the sample frame included executives belonging to this one professional

society. This study did not include businesses headquartered in countries outside the U.S.

The sixth assumption was that a survey approach and EFA was appropriate for

this study. A survey approach was best suited to collect the requisite data because

archival data to support this research was not available. This study was exploratory

versus conclusive in nature. Being exploratory allowed for exploring the dimensions of

the survey items and development of a measurement model (Williams et al., 2012).

Based on existing studies in which researchers applied exploratory surveys to develop a

composite measurement index, EFA appeared to be an appropriate methodology. In the

absence of any other quantitative studies in which researchers attempt to develop a model

to measure GBC, EFA was appropriate.

Limitations

To make research feasible, investigators limit what is under study (Bridges,

Hauber, Marshall, Lloyd, Prosser, Regier, Johnson, & Mauskopf, 2011). Limitations are

characteristics that are out of the researcher’s control but may influence the interpretation

of the findings of the study and establishment of external and construct validity (Brutus,

Aguinis, & Wassmer, 2013). Investigators may limit several areas. First, researchers may

limit the selection of elements under study. Researchers may make simplifications

necessary to produce a feasible study. Researchers may limit correlations among

elements. Alternatively, investigators may limit whether the participants generalize to the

population (Bridges et al., 2011). Several limitations to this study existed. As already

mentioned, the first was that sampling executives assembled at Executive Suite were a

subjective sample due to geography, demographics, and economic conditions.

Consequently, the results of this study may not generalize to businesses headquartered in

countries outside the United States. Replicating this study in other countries to determine

whether the results of this study generalize to other countries was an area for further

research.

A second limitation was the lack of existing academic studies specifically about

the theory of GBC and about developing a composite index to measure GBC. Leaders of

major corporations claim that their companies are business citizens (Crittenden et al.,

2011). Universities are incorporating business citizenship into their curriculums (Lilley et

al., 2014). Every year, the Clinton Global Initiative presents the International Global

Citizenship Award (Clinton Global Initiative, 2013). Despite these facts, the academic

literature was surprisingly lacking in the particular area of corporate or business

citizenship (Crittenden et al., 2011). There was abundant research going back 60 years in

the broad field of CSR, of which GBC is a subset (Carroll, 1999). While the lack of GBC

studies was a limitation, extrapolating relevant data from abundant CSR research

mitigates the limitation. This limitation also highlighted the need for this study.

A third limitation was that the operationalization of the constructs into survey

questions might have induced question redundancy or inclusion of weaker questions. The

EFA technique removed redundancy and retained the questions that were minimally

necessary to represent each factor. Applying EFA mitigated the limitation of question

redundancy or weak questions.

The EFA technique did not detect if there were missing questions in the original

survey. Thus, a fourth limitation was that in operationalizing the constructs into survey

questions, there might have been missing questions that jeopardized construct validity.

To mitigate this issue within the constraints of time and resource limits in concluding the

study, five SMEs reviewed the questions and offered suggestions for including missing

questions. The SMEs agreed that the questions were valid and did not offer suggestions

for adding additional questions.

Delimitations

Delimitations are intentional inclusionary and exclusionary boundaries that

researchers establish to guide their research and analysis process (Bartoska & Subrt,

2012). Delimitations identify the scope, or boundaries, of the study (Thomas & Magilvy,

2011). Establishing the boundaries allows future researchers to use the same data to

replicate or transfer the study (Thomas & Magilvy, 2011). The scope of this study was

limited to identifying critical factors required to measure GBC and determining the

minimal set of questions necessary to capture the factors. The four-step GBC framework

provided the assumed constructs. The constructs were the following. First, developing

overarching corporate values (VALUE). Second, implementing the values (IMPLE).

Third, analyzing problems and experimenting to revise the values or local

implementations (ANALY). Fourth, learning from the previous steps and

institutionalizing best practices (LEARN) (Wood et al., 2006). Operationalizing these

constructs, I developed a 23-question survey. Twenty-two Likert-type questions

described the four constructs. Using data from the survey, applying EFA reduced the 22

Likert questions to the minimum that were necessary, and mapped the essential questions

to an unknown number of subsets. In EFA, these subsets are known as “factors.” The

hypothesis section of this paper explains the factors. The resulting questions and factors

were those that best evaluated the maturity level that a company has attained in

implementing the four-step GBC framework. The resulting questions and factors

informed the creation of the composite index. The scope of this study was to create an

instrument that best captured the four constructs using EFA and using the EFA weights

(eigenvalues) to weight the instrument questions to compute an overall index of maturity.

Certain elements of GBC were out of the scope of this study. The study of why

business leaders do or do not implement the four-step framework of GBC was well

beyond the scope of this study. How or why stakeholders such as employees, consumers,

competitors, and non-governmental agencies pressure companies to practice business

citizenship behavior was beyond the scope of this study. Despite the importance of why

business leaders do or do not implement GBC or how or why stakeholders pressure

companies to practice GBC behavior, this study was limited to identifying factors that

described the four-step framework of GBC.

It was not within the scope of this study to assess the general views or opinions

corporate leaders had about the value of GBC. Soliciting the views about the value of the

GBC process was subjective. A qualitative study of such opinions may have value in

forwarding the academic research about the topic of GBC. Because this study was a

delimitated as described above, the study only focused on what items in the survey

described the implementation process.

This study did not cover how corporate leaders could move to implement the

four-step GBC framework. In their book, Global Business Citizenship, Wood et al.

(2006) clearly articulated how corporate leaders can apply the four-step GBC framework.

When complete, this study provided a way to measure at what point corporate leaders

have achieved in implementing the four-step framework of GBC. However, the study did

not touch upon how to apply the framework.

As already discussed, leaders at major companies such as American Electric

Power, Boeing, Dow, IBM, McAfee, McKesson, Microsoft Corp., and Nestle Waters all

claim that their companies are business citizens. The leaders claim that their companies

are business citizens because they are acting responsibly toward individuals,

stakeholders, and societies (Crittenden et al., 2011). This study did not examine how

corporate leaders that claim to be business citizens define the term or validate their claim.

Significance of the Study

CEOs at more than 80% of the U.S. Fortune 500 companies consider social

responsibility a mainstream component of their communications strategies (Taneja,

Taneja, & Gupta, 2011). As part of the communications strategies, 97% of the top 100

U.S. companies claim to be business citizens (Fifka, 2013). Research indicates that

business citizenship behavior may produce sustainable long-term economic, social, and

environmental benefits for the company and its stakeholders (Campbell, Eden & Miller,

2012; Menck & Oliveira, 2014; Wood et al., 2006). Business citizenship also provides a

moral, social, and political compass for business practice (Wood et al., 2006). Business

leaders are claiming that their companies are business citizens possibly to achieve

longterm economic, social, and environmental benefits. Business leaders are also making

the claim because stakeholders are demanding that they conduct their business as socially

responsible citizens of society (Park & Ghauri, 2015; Öberseder, Schlegelmilch, &

Gruber, 2011; Shum & Yam, 2011). The problem was that in searching the literature I

was not able to find a rating system for business citizenship, corporate citizenship, global

business citizenship, or any other derivative of the theme that business leaders can

selfadminister.

This study contributed to business practice and social change by introducing the

GBC index as a tool to quantitatively measure whether a company is a global business

citizen. Corporate leaders can use such a tool to demonstrate quantitatively to their

stakeholders the stage of maturity they have attained in becoming a global business

citizen. The following sections explain more fully the contribution to the business

practice and implications for social change.

Contribution to Business Practice

Measuring GBC was different from existing measures of CSR. The academic

literature provided several methods for measuring various aspects of CSR. The available

methods have limitations. One limitation was that measures rely on third-party data

(Glavas & Kelley, 2014). Another limitation was that reporting may not be conclusive for

comparisons between companies or countries (Panait, Voica, & Radulescu, 2014). A

final limitation was that reporting does not follow a standard format with defined

indicators that allow stakeholders to compare results within a firm over time (Berliner &

Prakash, 2014). Research by Calabrese, Costa, Menichini, Rosait, and Sanfelice (2013)

concluded that a method for assessing the stage a company had reached in overall CSR

cultural development was lacking. Most importantly, none of the CSR instruments

measured GBC. Measuring GBC is important for business leaders. Leaders who build the

reputation of being good business citizens may improve legitimacy, reduce the cost of

capital, improve access to capital, and experienced improved profitability.

Corporate leaders that build the reputation of being good business citizens

improve legitimacy in the eyes of their stakeholders (Wolf, 2014). When stakeholders

view a company as being legitimate, they provide access to essential resources (Wolf,

2014). Without essential resources, a company cannot operate (Wolf, 2014). Companies

must rely on stakeholders to obtain essential resources and maximize the value of the

corporation (Moura-Leite, Padgett, & Galán, 2014). Stakeholders are demanding that

corporate leaders conduct their business as though they are socially responsible citizens

of society (Park & Ghauri, 2015; Öberseder et al., 2011; Shum & Yam, 2011). To

conduct business and maximize corporate value, leaders must demonstrate to their

stakeholders that their companies are legitimate. Demonstrating that their companies are

business citizens indicates legitimacy (Park & Ghauri, 2015; Öberseder et al., 2011;

Shum & Yam, 2011).

Research suggested that corporate leaders that build the reputation of being good

business citizens reduced the cost of capital and improved access to capital by attracting

investors (Attig, El Ghoul, Guedhami, & Suh, 2013; Jiraporn, Jiraporn, Boeprasert, &

Chang, 2014; Oikonomou, Brooks & Pavelin, 2014). Credit ratings are an integral factor

in the rate companies pay for capital (Attig et al., 2013). Attig et al. (2013) provided

evidence that credit rating companies awarded relatively higher ratings to firms that

demonstrated business citizenship. Oikonomou et al. (2014) found that bond yield

spreads were higher for companies that demonstrated business citizenship. Oikonomou et

al. concluded that it appeared that investors perceived business citizens as having lower

credit risk and better credit quality.

Scherer and Palazzo (2011) compiled over 100 empirical studies that all

demonstrated a positive relationship between positive CSR and improved financial

performance. Researchers continue to demonstrate this positive relationship (Flammer,

2015; Gallardo-Vázqueza & Sanchez-Hernandez, 2014; Michelon, Besso, & Kumar,

2013; Torugsa, O’Donohue, & Hecker, 2013; Wang, Lu, Kweh, & Lai, 2014). As

discussed, GBC is a subset of CSR. Leaders who practice business citizenship may yield

improved profitability.

Because legitimacy, the cost of capital, and profitability are essential, it is

important for leaders to be able to demonstrate whether their companies demonstrate a

high, medium, or low level of business citizenship. Future researchers, practitioners, and

stakeholders will be able to use the GBC index as a tool to calculate the GBC score for

individual companies quantitatively. The score assesses the level of global business

citizenship for a given company. The score indicates whether individual companies

demonstrate a high, medium or low commitment to the four-step framework. The

fourstep GBC framework provided the assumed constructs. The constructs were the

following. First, developing overarching corporate values (VALUE). Second,

implementing the values (IMPLE). Third, analyzing problems and experimenting to

revise the values or local implementations (ANALY). Fourth, learning from the previous

steps and institutionalizing best practices (LEARN) (Wood et al., 2006).

Practitioners could use the GBC index to determine a baseline that would allow

them to establish goals for implementing the four-step GBC framework and chart

progress toward the goals (Wood et al., 2006). Practitioners could also use the index to

uncover areas for improvement or, conversely, identify areas of excellence to replicate

(Wood et al., 2006). By improving the GBC score, leaders would demonstrate their

commitment to GBC. A high GBC score could lead to improved legitimacy, the cost of

capital, and profitability.

Implications for Social Change

The concept of global business citizenship provides a moral, social, and political

compass (Wood et al., 2006). GBC is a process that allows companies to integrate

responsible and ethical business policies and actions that positively affect their economic,

social, and environmental performance (Wood et al., 2006). This compass may help

managers practice ethical conduct within the communities where they do business.

Practicing ethical conduct may produce sustainable long-term economic, social, and

environmental benefits for the company and its stakeholders (Wood et al., 2006). GBC is

an indicator of the level to which companies are maximizing shareholder value and

gaining a competitive advantage at the same time that they are incorporating laws, public

policies, political issues, and the interests of stakeholders. It also indicates that they are

acting ethically and responsibly for the benefit of individual managers, corporations,

industries, and society as a whole (Wood et al., 2006). The GBC index may identify

corporate leaders that apply the framework for being global business citizens.

A Review of the Professional and Academic Literature

This section focuses on a review of the literature needed to support the research.

First is an overview of the search method and instruments used to conduct the literature

review. The literature review covers the following four main themes: (a) explanation of

GBC; (b) the business and social case for GBC; (c) existing approaches to measuring

CSR; and (d) overview of EFA and its application.

This literature review begins with the seminal articles in which Wood and

Logsdon used deductive, descriptive typology to develop the theory and arrived at the

definition of GBC. Next is a presentation of the case for why corporate leaders should

integrate the GBC framework. With the definition and argument for implementing GBC,

the next section assessed current measurement approaches for CSR since none existed for

GBC. The final section presents an overview of the EFA technique and examples of how

researchers use EFA, to demonstrate the scope, strengths, and limitations of the

technique.

Literature Review, Search Methods and Search Instruments Employed

The extensive literature review began with the seminal articles published in 2002

by Donna J. Wood and Jeanne M. Logsdon, the authors of the theory of GBC. The

seminal work also included the subsequent book published by the authors in 2006. From

there, the search of the literature focused on the keywords corporate social

responsibility, CSR, corporate social performance, global business citizen, business

citizen, and corporate citizen. The search primarily included articles published in peer-

reviewed journals between 2011 and 2015 available from EBSCOhost, Emerald

Management Journals, ProQuest ABI/INFORM, SAGE Journals Online, and

ScienceDirect databases. Table 1 depicts the databases used. Automated Google Scholar

alerts for the search terms corporate social responsibility and global business citizen

yielded relevant articles appearing as advanced online publications in 2014 and 2015.

Updates to the reference section occurred as the journals publish the cited articles.

Table 1

Databases Used

EBSCOhost

Emerald Management Journals

ProQuest ABI/INFORM

SAGE Journals Online

Science Direct

The literature supported how and why the Wood and Logsdon developed the

theory of GBC. It also provided evidence for a solid case for the relevance of GBC

because when companies implement the GBC framework, they can obtain legitimacy

from society and increase competitive performance. Lastly, the literature provided

evidence that there was no rating system for GBC.

Database

Besides the core research topic, some reviewed literature provided an in-depth

understanding of the EFA research method. For this topic, the search of the literature

focused on the keywords exploratory factor analysis, EFA, and factor analysis. This

portion of the literature review provided information for an overview of factor analysis in

general and the EFA technique in particular. Researchers use EFA to test relationships,

generalize results to alternate populations, use existing scales, test the effectiveness of a

measurement instrument, or developed and/or refined a new measurement instrument.

The final section of the literature review covered EFA studies that demonstrated how

researchers used the EFA technique in these ways.

There are 175 peer-reviewed articles cited in this study. Of those, 91% of them, or

159, are references within five years of the anticipated graduation year of 2016 and 16

are older than five years. Of the 16 studies that are older than five years, two are seminal

articles published by Wood and Logsdon. Along with these two seminal studies, the book

Wood and Logsdon wrote based on their seminal studies was cited. The study referenced

two websites, and one was a government website. Overall, 89% of the references were

peer-reviewed and published within five years of the anticipated graduation year. Table 2

contains the numbers and percentages of the professional and academic literature

reviewed and all references used in the study.

Table 2

Reviewed Literature and All References Statistics

Literature 5 or fewer years Literature older Total Percentages <= 5

Literature Type old than 5 years years

Books 0 3 3 0

Dissertations 0 0 0 0

Peer-Reviewed Articles 159 16 175 91

Web Pages 2 1 3 67

Others (e.g., Gov.) 1 2 2 50

Total

Peer-Reviewed and

162 21 183 89

Dissertations <= 5 years 159 0 183 87

The Theory of Global Business Citizenship

Research into the role of business in society began in 1953 with the publication of

Howard R. Bowen’s book titled Social Responsibilities of the Businessman (Carroll,

1999). The term corporate social responsibility (CSR) emerged to describe the social

responsibility of businesses (Crittenden et al., 2011). As the society became concerned

with corporate practices such as depleting the environment, producing harmful consumer

products, and inhumane workplaces, there was a significant increase in CSR research

(Logsdon & Wood, 2002). By 1973, there was still no consensus about exactly what CSR

was, so the American Enterprise Institute sponsored a major debate about the meaning of

CSR (Carroll, 1999). In Carroll’s 1979 seminal study, he proposed a CSR model that

encompassed the economic, legal, ethical, and philanthropic responsibilities expected of

corporate leaders (Carroll, 1979). Through the 1990s, scholars introduced concepts

related to CSR. CSR concepts include corporate social performance (CSP), stakeholder

theory, business ethics theory, and corporate citizenship (Carroll, 1999; Crittenden et al.,

2011). Business leaders embraced the term “corporate citizenship” to define the ethical

component of their CSR efforts (Carroll, 1999). With this new practitioner focus, Carroll

revisited his model. He wrote, “the CSR firm should strive to make a profit, obey the

law, be ethical, and be a good corporate citizen” (Carroll, 1999, p. 289). However, there

was still no distinction or connection between CSR and corporate citizenship (Carroll,

1999; Logsdon & Wood, 2002; Wood & Logsdon, 2002). Global corporate citizenship

emerged as the prominent term on the ethically responsible role of business (Crittenden

et al., 2011). Wood and Logsdon seized the opportunity to develop a theory of corporate

citizenship that integrated CSR with the idea that businesses were citizens (Wood &

Logsdon, 2002). The researchers used the term “business” in place of “corporate” to

indicate that companies are engaged in business. Wood and Logsdon then extended the

theory to the global level. Wood and Logsdon linked global citizenship to global business

strategy by implementing broad universal principles and integrating legitimate cultural

norms, rules, and performance expectations (Logsdon & Wood, 2002).

Wood and Logsdon used a deductive, descriptive typology to develop their

theory of GBC (Wood et al., 2006). Deductive, descriptive typology is a well-established

analytical method used in social science research to reduce complexity, form concepts,

and explore dimensionality (Collier, LaPorte, & Seawright, 2012; Fiss, 2011). Wood and

Logsdon first presented the argument that the idea of citizenship translated from the

individual to a business organization (Wood & Logsdon, 2002). Wood and Logsdon then

presented strategic approaches for implementing GBC.

Wood and Logsdon (2002) began with the fact that in a globalized, free market

without a single government to establish and enforce rules, companies had new

opportunities to exploit people and the environment. Wood and Logsdon proposed that

the concept of global business citizenship would provide a moral, social, and political

compass. This compass would help managers practice prudent and ethical conduct within

the communities where they did business. Wood and Logsdon theorized that the ethical

conduct at the local level would produce sustainable long-term economic, social, and

environmental benefits for the company and stakeholders (Wood et al., 2006).

Wood and Logsdon began constructing their theory by demonstrating how the

idea of individual citizenship translated to companies. Over the past several thousand

years, philosophers have developed the notion that individuals are citizens with inherent

human rights (Logsdon and Wood, 2002). In their first seminal article, Wood and

Logsdon (2002) examined how the three most widely supported modern views of

citizenship for individuals translated to citizenship for companies. The three views of

citizenship for individuals they examined included the minimalist theory of civic

association, the communitarian model, and the universal human rights model. The

minimalist theory says that citizens are free agent residents of a common jurisdiction

striving to achieve their goals within the constraints of rules necessary to protect their

individual liberties (Wood & Logsdon, 2002). In the communitarian model, citizens unite

in a community having duties to participate in making and carrying out rules for the

welfare and preservation of the culture of the community. The universal human rights

model provides individuals the freedom to pursue their interests and has autonomy of

action while balancing that freedom with the welfare of the overall society. In this model,

the government secures and protects individualism, independence, and social welfare by

establishing rights and duties based on a set of shared values. Wood and Logsdon

analyzed whether the idea of business citizenship made sense from these three

perspectives of individual citizenship.

From the minimalist theory of civic association perspective, Wood and Logsdon

(2002) determined that companies could not be citizens. Wood and Logsdon resolved

that, within the minimalist framework, companies were a legal structure that allowed for

managers, as agents of shareholders, to negotiate contracts. Shareholders provide capital

and acquire property ownership through business activities to maximize their

selfinterests. Within this view, companies cannot be citizens because they do not act

independently of their shareholders.

Wood and Logsdon (2002) concluded that the communitarian model perspective

did provide for companies being business citizens. Wood and Logsdon determined that

because companies used the resources of the community and reflected the values of the

community, they were a part of the community yet distinct from individuals. Wood and

Logsdon argued that the business leaders tend to conform to local norms and contribute

to the welfare of the community to remain in good standing with the community. Wood

and Logsdon reasoned that the businesses thrived by helping the community thrive. In

the communitarian model, companies act as individual citizens, carrying out rules for the

welfare and preservation of the community; therefore, society considers them citizens.

Wood and Logsdon (2002) determined that the universal human rights model also

support companies as citizens, however, with weaker rights and duties than individuals.

Similar to the communitarian model, Wood and Logsdon argued that companies use the

resources of the community with the purpose of creating a surplus, thereby allowing

individuals and societies to do more with their limited resources. Companies must

become members of stakeholder networks that span multiple locales to use the resources

of local communities. Wood and Logsdon concluded that companies could be citizens

because they pursue their interests, have autonomy of action, and provide for social

welfare.

Furthering their argument for corporations as citizens, Wood and Logsdon (2002)

explained the legal and moral status of corporations. Wood and Logsdon began by

showing that corporations are legal entities with rights analogous to individuals. Wood

and Logsdon outlined how U.S. constitutional and case law recognized corporations as

artificial persons, subject to some identical criteria and protections as individuals. For

example, in the United States, both individuals and corporations enjoy the protections of

the First, Fourth, Fifth, and Fourteenth Amendments of the U.S. Constitution. Individuals

and corporations must abide by residency laws and have rights within the due process of

law to privacy, life, liberty, property, and exemption from double jeopardy. Corporations

also have the right to political participation, in the forms of lobbying and contributions to

political campaigns. Wood and Logsdon concluded that granting legal status to

companies allow them to, “better serve their human purposes and their human

constituents” (p. 83). Wood and Logsdon further concluded that since companies have

limited legal rights analogous to individuals, we should attribute limited citizenship

rights and duties to them as well, and being legal citizens comes with moral obligations.

Wood and Logsdon (2002) went on to determine the moral obligations of

companies. Wood and Logsdon outlined French’s 1979 viewpoint that since business can

behave rationally and with an intention, they have moral obligations to self-regulate,

according to the community’s moral rules. By contrast, they outlined the opposite

viewpoint put forth by Ladd in 1970. Ladd’s view was that companies lack intentionality

and autonomy. Therefore, companies are incapable of moral obligations, and regulatory

controls are necessary to achieve ethical business practices. Wood and Logsdon also

outlined middle-ground perspectives by Donaldson in 1982, Werhane in 1985, and

DeGeorge in 1999. These middle-ground perspectives included several ideas. First,

companies have a different moral agency to follow than individuals because the people

within companies collectively know more than an individual can know. Second, business

actions cannot be reduced to individual actions; therefore, businesses have some moral

obligations. Third, companies are moral actors but not moral persons because they are

not human beings with awareness of their actions. Fourth, stakeholders should not expect

companies to act morally, but should praise positive moral actions and place blame when

they violate the moral law. From the arguments about the legal and ethical obligations of

companies, Wood and Logsdon concluded that individuals form companies to further

societal satisfaction. Businesses add to societal satisfaction by creating jobs, growing the

economy, investing in research and development, and education. As such, corporations

should be subject to some of the same legal and moral rights and obligations as

individuals. However, Wood and Logsdon also concluded that it was inappropriate to

consider that companies are equivalent to individuals.

In a second seminal article, Logsdon and Wood (2002) used Aristotle’s idea that

individual citizens have duties to participate in the political process, pay taxes, and

participate in military service. Over the centuries, societies have adapted these

fundamental duties to custom and law. In their argument, Logsdon and Wood

demonstrated how these duties of individual citizens translated to corporate

responsibilities.

First in Logsdon and Wood’s (2002) argument was that individual citizens have a

responsibility to participate in the political process. Logsdon and Wood argued that

individuals participate in voting, engaging in political discourse and avoiding unfair

influences, such as bribery or coercion. Wood and Logsdon showed that while firms

cannot vote, they do participate in the political process by lobbying and avoiding unfair

influences such as bribery or coercion. Additionally, Logsdon and Wood argued that

investors and consumers sanction business activities by effectively voting with their

spending dollars.

Second in Logsdon and Wood’s (2002) argument, individual citizens are

obligated to pay taxes that benefit the collective good. Individuals also voluntarily

support social services through philanthropy. Likewise, firms are obligated to pay taxes.

Many corporations also voluntarily support social services through philanthropy.

In the final piece of Logsdon and Wood’s (2002) argument, individual citizens

have a duty to participate in the defense against common threats. Firms are required to

support their employees serving the military. Furthermore, Logsdon and Wood claimed

that corporations should “defend against enemies of the business institution whether

within or across political borders” (p. 174). Logsdon and Wood argued that capitalism

depend on laws and the moral ideas of private property and human rights. Without these

ideas, there is no ground for entering into contracts. Logsdon and Wood thus contended

that the enemies of business organizations are conditions that threaten human rights, and

firms must defend against such threats. With the understanding that companies do have

duties similar to the duties of individual citizens, Logsdon and Wood concluded that

companies meet these criteria of being citizens. Logsdon and Wood also found that

citizenship implies inherent natural human rights. The next step Wood and Logsdon took

was to define the four-step framework of the theory of GBC. The four steps were

developing corporate values (VALUE), implementation (IMPLE), problem analysis

(ANALY), and learning and institutionalizing best practices (LEARN).

Values. Logsdon and Wood (2002) began integrating the idea of global

citizenship into their theory by examining universal values of individual citizenship.

Logsdon and Wood cited the United Nations’ Universal Declaration of Human Rights as

the foundation that there was a uniform set of rights and obligations protecting the

individuals of the globe. Logsdon and Wood also cited the fact that most countries

incorporated civil and political rights into their constitutions. Logsdon and Wood “claim

that there is now a common awareness of cross-cultural conditions and a common

language of rights that help to shape the social, political, and economic forces of the

world” (p. 164). Logsdon and Wood argued that multinational companies profoundly

shape the economy, society, and politics. Therefore, companies directly affect human

rights. Articles 28–30 of the Universal Declaration of Human Rights tasks companies,

countries, other organizations, and individuals with promoting social welfare, protecting

human rights, and defending against any person or entity attempting to destroy any of the

rights outlined in the Universal Declaration of Human Rights.

Global companies have an obligation to act as citizens and uphold these universal

human rights. The theory of GBC states that the first step toward becoming a global

business citizen is to develop a companywide, overarching values in the form of a code

of ethical conduct (Wood et al., 2006). The theory of GBC suggests that corporate

leaders use the Universal Declaration of Human Rights as the foundation for developing

companywide overarching values (Wood et al., 2006). Developing overarching values is

the first step of the four-step framework of the theory of GBC (Wood et al., 2006). This

first step equates to the first construct of VALUE. The next question Logsdon and Wood

addressed was how to handle differences between overarching values and local norms.

Implementation. Logsdon and Wood (2002) argued that in today’s global

business environment, business leaders must uphold universal human rights. Logsdon

and Wood also argued that in upholding human rights, they must work within the

“norms, rules, and performance expectations” of the local communities in which they do

business

(p. 165). Logsdon and Wood first demonstrated that local norms vary among

communities, and some may even conflict with universal rights, but all have social

legitimacy within their community.

A study by Barkemeyer and Figge (2014) supported the view that corporate

leaders should incorporate local values into value codes. Barkemeyer and Figge found

that centralizing development and maintenance of CSR agendas at the corporate

headquarters did not result in optimal solutions for stakeholders outside the headquarters.

Barkemeyer and Figge called the phenomenon “headquartering” Barkemeyer & Figge

found that headquartering led to management problems at subsidiaries away from the

headquarters. Barkemeyer and Figge suggested leaders that control the CSR agenda and

do not allow for local implementation could expect one of three outcomes. First, when

there is a match of headquarter and local values, the local stakeholders will uphold the

local values. Second, when there is a mismatch of priorities, local stakeholders will

pursue their agendas. Third, when there is a mismatch of talk and action, local

subsidiaries may not implement the headquarters’ initiatives.

The language used in ethical codes is as important as the ethics themselves.

Winkler (2011) studied companies listed on the Frankfurt Stock Exchange to determine

how the language of ethical codes defined relationships between employees, line

managers, top managers, and compliance officers. The results showed that the ethical

codes placed management at the upper side of the formal hierarchy. The language

granted this group superior rights and knowledge. The codes depicted employees as

passive receivers of the codes. The language suggested that employees required ethical

monitoring and control. The language further indicated that employees were not

competent to interpret the codes nor empowered them to achieve higher morality.

However, Winkler’s results also exposed that management expected employees to

demonstrate the ethical behavior of the corporate leaders. This ambiguity leads to

indifference toward the ethical code. Winkler concluded that to have ownership for a

corporate ethical code; employees must have a stake in the code and perceive it as their

personal code of conduct. Winkler acknowledged that establishing ownership is difficult

when a company has thousands of employees spread across the globe, and the parent

company attempts to disseminate their fixed code throughout the organization.

International managers face pressure to adapt local norms if they wish to do

business in the community. Wood et al. (2006) proposed that GBC companies adopt a

“limited number of broad universal principles; a wide range of variant applications, some

acceptable, some not; and a “free space” in which norms and applications remain to be

developed” (p. 171). The free space allows international managers the ability to

incorporate local norms, rules, and performance expectations into their business

practices. Recent studies support this assumption.

Through an exploratory and descriptive case study, Proenca and Branco (2014)

demonstrated that local managers at companies in Portugal engaged in CSR activities

that aligned with their personal values and morals. Owen and Kemp (2014) argued that to

bring about positive CSR change within the mining industry, the moral sensibilities of

the local, in the trenches, employees should drive professional habits and organizational

strategies. Despite the importance of following local issues, Bondy and Starkey (2014)

found that the 37 multinational companies they researched did not incorporate local

culture into their CSR policies but adopted a unified international strategy. Bolton, Kim,

and Gorman (2011) conducted a case study to examine the initiation, implementation,

and maturation process of implementing a CSR strategy. In the end, Bolton et al.

identified a CSR strategy as a morally responsible business citizen. Bolton et al. selected

of one of the largest multinational companies with more than 102,000 employees as their

case study company. Bolton et al. found that implementing or sustaining any CSR-related

initiatives without employee engagement and an agreement was difficult. Through

analyzing the company’s “People Survey” of more than 100,000 employees around the

globe, Bolton et al. found that employees’ personal morals significantly affected the

company, both positively and negatively. It was mandatory that companies allowed

employees at the local levels to have input to creating the CSR strategy.

Logsdon and Wood (2002) argued that to uphold human rights; corporate leaders

must incorporate the norms, rules, and expectations of local communities.

Implementation of ethical codes is a long-term process that requires commitment,

communication, and integration into business practices for employees to integrate them

into organizational cultures (Erwin, 2011). Implementing the overarching values

throughout the organization and incorporating local customs, norms, and ethical

standards became the second step of the four-step framework of the theory of GBC. This

second step equates to the second construct of IMPLE.

Analyze. Wood et al. (2006) acknowledged that incorporating local norms, rules,

and performance expectations into business practices might lead to conflicts with the

overarching values or conflicts within local societies. Logsdon and Wood proposed that

international managers must experiment to determine which norms, rules, and

performance expectations to incorporate, and how. Through this experimentation, the

business adopts policies and procedures that fit the local culture, ethical tradition, and

law, and best serves the people and the firm. Analyzing problem areas and experimenting

with solutions to remediate conflicts became the third step of the four-step framework of

the theory of GBC (Wood et al., 2006). This third step equates to the third construct of

ANALY.

Learn. Logsdon and Wood further argued that, through experimentation, the

entire business must systematically learn from their successful implementation of local

business practices. Systemizing learnings from the IMPLE and ANALY steps, and

institutionalizing best practices throughout the organization became the fourth step of the

four-step framework of the theory of GBC (Wood et al., 2006). This fourth step equates

to the fourth construct of LEARN.

Wood et al. (2006) defined a global business citizen as, “a business enterprise

(including its managers) that responsibly exercises its rights and implements its duties to

individuals, stakeholders, and societies within and across national and cultural borders”

(p. 4). Within their concrete framework, the first step is to develop a companywide

overarching code of ethical conduct consistent with the definition of GBC. The second

step is to implement the overarching code of ethical conduct throughout the organization

and adapt it to local customs, norms, and local ethical standards that seem in conflict with

the overarching code of ethical conduct. The third step is to analyze areas where local

norms, rules, and performance expectations conflict with the code of ethical conduct and

experiment with creative and practical solutions to remediate the conflicts. The fourth

step is to learn from the previous steps, systemize, and institutionalize the best policies,

practices, and behaviors throughout the organization.

Grounded in an overarching code of conduct, implementing the four-step GBC

framework will orient companies in a way that enhances legitimacy through socially

acceptable behavior at the individual, organizational, and systemic levels. With their

seminal companion articles, Wood and Logsdon made a compelling case that companies

have limited legal and moral obligations analogous to individual citizens. Wood and

Logsdon showed how companies carry out rules for the welfare and preservation of the

community. Wood and Logsdon explained that corporations had autonomy to pursue

their interests and provide for social welfare. Wood and Logsdon showed that, like

individual citizens, corporations have duties to participate in the political process, pay

taxes, and participate in military service. Wood and Logsdon made the case that global

companies have an obligation to act as citizens and uphold universal human rights. When

corporations act like global business citizens, they not only meet these obligations, but

they also receive economic, social, and environmental performance benefits.

The Case for Global Business Citizenship

As we have seen, the theory of GBC integrates CSR with the concept that

companies have limited legal and moral rights and obligations as citizens (Wood &

Logsdon, 2002). The result of implementing the GBC framework’s responsible and

ethical business policies and actions positively affect a company’s economic, social, and

environmental performance. Research shows that CSR strategies have the potential to

obtain legitimacy from society and increase competitive performance (Menck & Oliveira,

2014). Wood et al. (2006) suggested that corporate leaders could integrate CSR into

business practices by identifying and implementing ethical codes of conduct, the first

step to becoming a GBC business. The benefits of CSR translate to the benefits of GBC,

therefore, in making the case for GBC it was appropriate to examine the literature related

to CSR.

Research and surveys indicate that by engaging in CSR behavior, companies have

the potential to obtain legitimacy from society and increase competitive performance

(Menck & Oliveira, 2014). Legitimacy and competitive performance directly affect

profitability. The below sections make the case for why business leaders would benefit

from implementing the GBC framework.

Legitimacy. With globalization came the loss of shared moral orientation,

widening governance gaps, and increased public awareness of the conduct of corporate

leaders (Voegtlin, Patzer, & Scherer, 2012). As such, it is more important for corporate

leaders to build their legitimacy and maintain trustful relationships with stakeholders

(Voegtlin et al., 2012). An important way for corporate leaders to obtain legitimacy from

society is by engaging in CSR activities (Menck & Oliveira, 2014; Zheng, Luo, &

Makisomov, 2014). Legitimacy is the assumption that the actions of a company are

consistent with societal norms, values, and beliefs (Du & Vieira, 2012). Legitimacy is

vital for corporations because stakeholders will only ensure a continuous flow of

essential resources to entities they perceive as legitimate and reputable (Du & Vieira,

2012). Fundamentally, a reason companies would benefit from implementing the GBC

framework is that it helps build reputation and legitimacy (Campbell, Eden & Miller,

2012).

The reputation and legitimacy of companies reside within the perceptions of

stakeholder (Bitektine, 2011). The theory of social judgments of organizations explains

that stakeholders build their perceptions of the reputation, legitimacy, and status of

corporations by five factors (Bitektine, 2011). The first is whether a firm belongs to a

category or industry. The second is whether an organization has the right to exist. The

third is whether an organization is beneficial or hazardous to individuals, social groups or

society as a whole. The fourth is the performance and behaviors exhibited by the

corporation. The fifth is how an organization ranks with similar organizations.

Stakeholder perceptions establish reputation and legitimacy (Bitektine, 2011) and CSR

strategy significantly influence stakeholder perceptions (Dilling, 2011).

Within the context of legitimacy, Parsons, Lacey, and Moffat (2014) studied how

managers conceptualized that society grants corporations a ‘social license to operate’.

Parsons et al. found that the participants conceptualized the notion of social license to

operate within the four themes of legitimacy, localization, process and continuum, and

manageability. Concerning legitimacy, the participants of the study indicated that they

focused on the community approving and accepting of the way corporate leaders do

business, rather than whether a corporation’s values aligned with societal or cultural

values. The participants evaluated corporate legitimacy.

The results of a study by Park, Lee, and Kim (2014) showed economic and legal

CSR initiatives had a significant impact on corporate reputation. Their study also showed

that ethical and philanthropic CSR initiatives did not directly affect corporate reputation.

However, these categories did indirectly affect the trust consumers had related to

corporate integrity and benevolence. The authors demonstrated that if ethical and

philanthropic CSR strategies instilled integrity and social benevolence in consumers, they

enhanced corporate reputation.

Stakeholders perceive the reputations of companies in the oil industry to be highly

controversial (Du & Vieira, 2012). As a result, one of the strategies of oil companies to

build a positive reputation is to engage in CSR initiatives (Du & Vieira, 2012). In a study

by Du and Vieira (2012), the researchers found that oil companies engaged in a broad

range of CSR initiatives and used those initiatives as public relation campaigns. Because

of the study, the researchers recommended that to build a reputation; oil companies

should stop using “CSR as public relations.” Instead, the researchers suggested that oil

companies should engage in long-term socially responsible initiatives, such as

developing renewable energy sources.

Stanaland, Lewin, and Murphy (2011) looked at the issue from the perspective of

how perceived financial performance, and ethical conduct influenced the perceptions of

firms’ CSR strategies. The results of the study showed that when consumers saw the

positive financial performance, they provided a better evaluation of CSR activities.

Similarly, when consumers saw high-quality ethics statements, they provided a better

evaluation of CSR activities. The results of their study also showed that when consumers

had a positive CSR perception, they rated corporate reputation and consumer trust and

loyalty higher while rating perceived risk lower. The authors concluded that commitment

to positive financial performance and ethical standards positively influenced perceptions

of CSR activities, which directly related to positive corporate reputation and consumer

loyalty and trust.

Dilling (2011) showed that the CSR strategies of companies significantly

influenced stakeholder perceptions. In Dilling’s study, the age of corporations and their

publishing of CSR reports significantly increased stakeholders’ perceptions of

corporations. Surprisingly, Dilling also demonstrated that CSR efforts in the areas of

cultural diversity and community development decreased stakeholders’ perceptions of

organizations.

Corporate reputation informs how stakeholders view the ethical conduct of

companies. Cian and Cervai (2014) published a study to clarify the definitions of and

relationships between the terms, “corporate image,” “projected image,” “construed

image,” “reputation,” “organizational identity”, and “organizational culture.” Cian and

Cervai provided several conclusions. One conclusion was that the perceptions of

stakeholders informed reputation. Another conclusion was that reputation was the answer

to “who we are”. Similarly, reputation was corporate culture. Reputation was also what

internal stakeholders believe external stakeholders thought about their company. Finally,

reputation was what management communicates to external audiences.

Vos, Shoemaker, and Luoma-aho (2014) demonstrated that corporate

communication was the strategic interface between both internal and external

stakeholders. The discipline of corporate communication aims to develop good will and

mutual relations while acknowledging possible conflicting interests. Vos et al.

recognized that stakeholders do not all have a shared stake in an organization.

Stakeholders will have interests’ specific to their issues and possibly even have opposing

points-of-view. Their viewpoints and expectations inform corporate reputation.

The above-cited studies showed that the perceptions stakeholders had about

CSRrelated economic and legal initiatives, ethical conduct, and publishing of CSR

reports all positively influenced corporate reputation and legitimacy. A study by

Campbell et al. (2012) confirmed that the main effect of CSR on corporate reputation

was both statistically and practically significant. Despite this finding, Campbell et al.

found that when local branches of multinational corporations were far away from the

corporate headquarters, they were less likely to engage in CSR. Local branches that were

proximate to the corporate headquarters were more likely to engage in CSR. Campbell et

al. concluded that despite the positive benefits of CSR, foreign firms at greater distances

were less willing to invest in host-country CSR.

CSR strategies have the potential to help companies obtain the legitimacy they

need to gain cooperation and necessary resources from stakeholders and conduct business

within society (Menck & Oliveira, 2014). Without legitimacy, there is no shareholder

value. Assuming that the purpose of business is to maximize the value to shareholders

within ethical and lawful means, as Carroll (1979) suggested, there is no longer a

question of whether leaders should integrate CSR into their business strategies, but how.

Wood et al. (2006) suggested that the global business leaders could integrate social

responsibility into their business practices by becoming global business citizens.

Numerous studies indicate that positive CSR increases competitive performance. CSR

increases competitive performance because it adds value to stakeholders, especially

consumers, and employees, influences the cost of doing business, in particular, the cost

of capital, and improves financial performance.

Competitive Performance. The value CSR activities add to stakeholders and the

costs associated with CSR affects the competitive performance of companies (Menck &

Oliveira, 2014). The value CSR activities add to stakeholders, especially consumers, and

employees, affect competitive performance. The following focuses on literature

informing how CSR activities add value to stakeholders, affect the cost of capital, and

affect the financial performance.

CSR Value to Stakeholders. Stakeholders significantly influence CSR activities

and business citizenship behavior (Lahouel, Peretti, & Autissier, 2014; Park & Ghauri,

2015). Corporate leaders that build the reputation of being a good business citizen

improve legitimacy in the eyes of their stakeholders (Wolf, 2014). When stakeholders

view a company as being legitimate, they provide access to essential resources (Wolf,

2014). Increasingly, stakeholders are demanding higher levels of CSR performance.

Stakeholders are demanding that corporate leaders conduct their business as though they

are socially responsible citizens of the society (Park & Ghauri, 2015; Öberseder et al.,

2011; Shum & Yam, 2011). Companies must rely on stakeholders to maximize the value

of the corporation (Moura-Leite et al., 2014). Responsible leaders must consider the

consequences of corporate actions on all stakeholders (Voegtlin et al., 2012). MouraLeite

et al. (2014) found that providing what stakeholders demanded had a significant positive

effect on corporate financial performance.

Farmer (2014) identified a shift from corporate leaders being accountable to

shareholders above all to being accountable to stakeholders, including employees,

consumers, and local communities. Farmer noted that European business leaders focus on

investing in the communities in which they do business. As Wood et al. (2006)

contended, Farmer commented that the model is more sustainable because acting in this

socially responsible way should enhance competitiveness and maximize wealth creation

to the overall society.

Carlon and Downs (2014) developed a comprehensive method to account for the

financial value stakeholders bring to a firm. In the model, the first phase was to

determine whether the firm was a stakeholder firm, meaning they were accountable to

stakeholders first, as opposed to a shareholder firm that was responsible to shareholders

first. The second phase was to assess the financial value various stakeholders bring to the

firm. In the financial valuation, Carlon and Downs included the importance of business

citizenship activities. The third phase was to account for and report on that value.

The results from the above studies indicated that stakeholders influence the CSR

behavior of companies. However, a study about stakeholders’ power to influence CSR in

China indicated a different story. Lu and Abeysekera (2014) acknowledged that in the

past few years, the Chinese government had made significant strides toward continuous

economic growth, but CSR disclosure is a relatively new practice in China. The results of

Lu and Abeysekera’s study indicated that while CSR disclosure had a positive

association with profitability, the power of stakeholders to influence CSR disclosure was

weak.

Contrarily, Park, Chidlow, and Choi (2014) found that among South Korean firms, both

primary and secondary stakeholders were able to influence CSR practices positively.

Goyal, Rahman, and Kazmi (2013) determined that between 1992 and 2011, most of the

CSR research analyzed the relationship between CSR and financial performance in

developed countries. Studying stakeholders’ power to influence CSR in developing or

government controlled countries seems to be an area warranting further research.

Consumers are arguably the most important stakeholder group for a company

(Menck & Oliveira, 2014). A company cannot stay in business without consumers

purchasing their products. Research indicated that consumer stakeholders were concerned

about price and the value products provided. Consumers are also concerned about the

meaning, or the social identity related to the product, company, or industry (Menck &

Oliveira, 2014). Surveys of by Cone Communication (2013) of 10,287 actual consumers

showed that fifty-five percent of global consumers said they had refused to buy products

in the last year because of negative social responsibility. While some research supports

direct relationships between CSR and consumers’ intentions to purchase products,

attitudes do not always predict behavior (Papaoikonomou, Ryan, & Giniesis, 2011).

Likewise, purchasing intentions do not necessarily translate into actual purchases

(Papaoikonomou et al., 2011).

In a multi-method qualitative study, Papaoikonomou et al. (2011) examined why

consumers might not make purchases aligned with their CSR concerns. Papaoikonomou

et al. found several reasons for the intention—behavior gap. One reason was that

consumers encountered a lack of comparable products produced by companies aligned

with their ethical concerns. Consumers also complained that ethical alternatives did not

meet their expectations in areas such as functionality or style. Consumers indicated that

they found it difficult to obtain information about how companies made products and felt

that CSR reporting was incomplete or of questionable quality and credibility. Another

reason for the intention—behavior gap was that consumers were on limited budgets and

companies priced ethically responsible products too high. There were also factors of

everyday life such as the “pester power” of children demanding products, the influence

friends and peers, and opting for the easy choice.

Mandhachitara’s (2011) quantitative study of bank customers showed that CSR

had a strong, significant positive relationship with consumers’ loyalty, suggesting repeat

patronage intentions. There was a significant and positive association between repeat

patronage intentions and service quality. Results also showed that perceived service

quality contributed to consumers’ loyalty. While CSR initiatives did not significantly

relate to repeating patronage intentions, CSR did show positive and significant

relationships with loyalty and perceived service quality. Supporting Mandhachitara’s

research, the 2013 Cone survey of 10,287 consumers showed that when companies

engaged in CSR, 96% of respondents had a more positive image. Ninety-four percent

reported they were more likely to trust those companies, and 93% indicated that they

would be more loyal to those companies. Mandhachitara concluded that the positive

associations between CSR and product loyalty and service quality had direct

consequences for banking services.

Öberseder et al., (2011) found that when consumers did not have information about

CSR, it did not play a role in their purchasing decisions. When consumers did have

relevant information, they began a complicated hierarchical process of core, central, and

peripheral factors to assess their purchasing decisions. The researchers concluded that the

complexity of the decision process hindered consumers’ purchasing decisions. Research

supports that incorporating CSR into business practices has the benefit of influencing

consumers’ purchasing intentions, however, their intentions may not translate to actual

purchasing behaviors due to a variety of complicated hierarchical factors.

Shen, Wang, Lo and Shum (2012) found that consumers indicated that they were

willing to pay a higher price for fashions produced in manners consistent with human and

environmental well-being. Their study also reported that when consumers had limited

knowledge about global sweatshop practices of fashion retailers, that lack of knowledge

prevented consumers from translating their concerns into purchases. Likewise, Marquina

and Morales (2012) found that consumers in Peru and Spain indicated that they were

willing to pay more for products from companies with good CSR reputations, including

sound labor and environmental practices.

Employees are a primary stakeholder group of all corporations (Farooq, Farooq,

& Jasimuddin, 2014). Results from a study by Evans and Davis (2014) indicated that the

more employees perceived their employers to exhibit business citizenship behavior, the

more likely they were to identify with the organization. The employees also engage in

organizational citizenship behavior and avoided the potential deviant behavior in the

workplace. CSR activities help organizations retain employees who are loyal and have

positive attitudes (Hansen, Dunford, Boss, Boss, & Angermerier, 2011; Wong & Gao,

2014). The results of a quantitative study by Hansen et al. (2011) showed that employees

who had positive CSR perceptions of their company had decreased intentions to leave.

The employees also demonstrated increased organizational citizenship behavior.

Likewise, Wong & Gao (2014) found that stakeholder related CSR activities resulted in a

significant employee commitment to the organization.

Evans and Davis (2014) conducted a study to predict whether employees’

perception of the business citizenship of their company would influence their

organizational citizenship behavior. In this study, Evans and Davis defined corporate

citizenship as how the corporate leaders were “fulfilling economic, legal, ethical, and

discretionary responsibilities imposed on the firm by its stakeholders” (p 129). Evans and

Davis measured organizational identification, organizational citizenship behaviors,

employee deviance, and socially desirable responding. The results of their study

indicated that when employees perceived their employer as exhibiting business

citizenship, they were more likely to identify with the organization, engage in

organizational citizenship behaviors, and avoid the deviant behavior.

Deviant behavior is costly and harmful to organizations (Aleassa & Zurigat,

2014). Aleassa and Zurigat conducted a study to determine how employees who

identified with their corporation and the company’s ethical values responded when faced

with unethical behaviors committed by their peers. Aleassa and Zurigat found that

employees who identified with their company were more likely to report their peers’

misconduct. The question this raises is how companies can influence how employees

identify with the company.

Farooq et al. (2014) conducted a study to determine if evaluating CSR initiatives

could predict employee behavior. The results of their study demonstrated that CSR

initiatives directed toward the community, employees, and consumers positively

influenced how the employees identified with the organization and how they shared

information. Therefore, employees will identify with companies that engage in CSR

initiatives.

Chen and Hung-Baesecke (2014) found that leaders’ participation and modeling

of CSR directly affected employees’ attitudes about CSR. In particular, Chen and

HungBaesecke found three management behaviors that directly and indirectly affected

employee CSR participation. The three management practices were role modeling or

leading by example, advocating for CSR and facilitating employee CSR participation.

Bohdanowica, Zientara, and Novotna, (2011) demonstrated a significant link

between CSR and employee engagement. Some companies are using this link to promote

their CSR activities as a way to attract employees who align with the company culture

(Bohdanowica et al., 2011). Bohdanowica et al. analyzed Hilton’s We Care! Program,

Hilton’s We Care! Program involved over 16,000 of the hotel’s employees to develop

CSR initiatives in the form of “greening” to reduce energy use and CO2 emissions.

Bohdanowica et al. showed that the program motivated 95% of the Hilton employee

participants. Ninety percent believed the program helped improve Hilton’s profitability,

and 89% replied that for them, the program improved the hotel’s image. Hilton promotes

the We Care! Program to prospective employees.

Socially responsible activities help organizations retain employees who are loyal

and have positive attitudes (Hansen, Dunford, Boss, Boss, & Angermerier, 2011; Wong

& Gao, 2014). When employees perceived their employer as exhibiting corporate

citizenship, they were more likely to identify with the organization, engage in

organizational citizenship behaviors, and avoid deviant behavior (Evans & Davis, 2014).

Likewise, employees who identified with their company were more likely to report their

peers’ deviant behavior (Aleassa & Zurigat, 2014). Companies can influence how their

employees identify with them by engaging in CSR initiatives directed toward

community, employees, and consumers (Farooq et al., 2014). When leaders participate

and model

CSR behaviors, they directly affect employees’ attitudes about CSR (Chen &

HungBaesecke, 2014). Because of the significant link between CSR and employee

engagement, some companies are promoting their CSR activities as a way to attract

employees aligned with the company culture (Bohdanowica, Zientara, & Novotna, 2011).

CSR and the cost of capital. Numerous studies indicate that CSR activities

influence investors and the cost of capital. Research suggested that positive CSR

strategies reduce the cost of capital and improve access to capital by attracting investors.

Credit ratings are an integral factor in the rate companies pay for capital (Attig et al.,

2013). A study by Attig et al. (2013) provided evidence that credit rating companies

awarded relatively higher ratings to firms with good social performance, especially CSR

that extended beyond compliance. By sampling the KLD database, Jiraporn et al. (2014)

found that firms’ credit ratings increased as much as 4.5% when they increased their

KLD CSR score by one standard deviation. Oikonomou et al. (2014) reviewed more than

3,000 bond issues by 742 firms operating in 17 industries. Oikonomou et al. found that

bond yield spreads were higher for companies that demonstrated good social performance

and companies with negative social performance paid more for capital. The authors

concluded that it appeared that good social performance led investors to perceive the

companies as having lower credit risk and better credit quality.

Strugatch (2011) determined that 82% of investors evaluate social responsibility

when making their investment decisions. Cohen, Holder-Webb, Nath, and Wood (2011)

conducted a study to determine what nonfinancial information investors used and desired

to use more of in the future. Cohen et al. found that investors placed the most importance

on economic performance followed by corporate governance and corporate social

responsibility. The respondents also indicated interest in using more nonfinancial

information in the future. Furthermore, the respondents clearly indicated their preference

for corporate social responsibility information provided by a third party. Aspara and

Tikkanen (2011) found that 85% of investors were willing to invest in a company with

lower financial returns if they identified positively with the corporate identity. Likewise,

Dhaliwal, Li, Tsang, and Yang (2011) and Xu, Liu, and Huang (2015) found that

institutional investors invested in companies with superior CSR performance. Dhaliwal et

al. (2011) found that companies that voluntarily disclosed superior CSR performance

paid less for equity capital. Additionally, these firms raised significantly more capital

than companies that did not disclose their CSR activities. Similarly, Girerd-Potin,

Jimenez-Garcès, and Louvet (2014) found that the cost of equity was lower for

companies that demonstrated social responsibility toward business stakeholders, societal

stakeholders, and financial stakeholders.

Cho, Lee, and Pfeiffer (2013) found a significant positive association between the

“bid—ask” spread and the CSR activities of the companies in which they were investing.

The authors concluded that this indicated that investors with knowledge about the CSR

performance of corporations exploited that information when valuing stock. A study by

Cheng, Ioannou, and Serafeim (2014) also indicated that companies with better CSR

performance were able to raise more financing. Dhaliwal et al. also found that companies

with superior CSR performance attracted more analyst coverage, and the analysts

produced reports with fewer forecast errors and dispersion.

Elliott, Jackson, Peecher, and White (2013) found that investors valued corporate

stock based on firm CSR performance and their assessment of the CSR performance.

Furthermore, the researchers found that investors who did not explicitly analyze CSR

performance were more willing to invest in a firm with positive perceived CSR

performance. The results of a study by Sun and Cui (2014) indicated that companies with

positive CSR did reduce the risk of default.

Investors in Bangladesh do not seem to demonstrate the level of rewarding

companies with good CSR by reducing the cost of equity as the above studies indicate.

Ahmed, Islam, Mahta, and Hasan (2014) sampled 152 companies listed on the Dhaka

Stock Exchange. Ahmed et al. found that good CSR performance had a positive, but

insignificant, relationship to the amount of money that institutional investors supplied to

those companies in Bangladesh. In summary, research indicates that CSR activities

attract investors and reduce the cost of capital, which directly affects financial

performance.

CSR and financial performance. Scherer and Palazzo (2011) compiled over 100

empirical studies that all demonstrated a positive relationship between positive CSR and

improved financial performance. For this study, the Walden Library was the resource that

enabled a comprehensive search for articles published in 2013 and 2014 that reported

results examining the relationship between CSR and financial performance. Nine articles

reported a positive relationship; one reported a peripheral positive relationship; three

articles reported no relationship, and one study of Chinese companies reported a negative

relationship.

In the following nine articles, researchers found a positive relationship between

CSR and financial performance. By studying the CSR-related proposals that shareholders

passed by a narrow margin, Flammer (2015) found the CSR activities increased

shareholder value by 1.77%. Wu and Shen (2013) found that CSR positively affected

financial performance in the banking industry. In particular, Wu and Sen found that CSR

positively affected return on assets, return on equity, net interest income, and non-interest

income. Michelon et al. (2013) analyzed the best business citizens as defined by Kinder,

Lyndenberg, Domini Analystics, Inc. (KLD). Wu and Shen found a positive relationship

between CSR and financial performance when those companies strategically aligned their

CSR initiatives to stakeholder interests. In another study using KLD data of companies in

the U.S. telecommunications industry, Wang et al., (2014) found a significant positive

relationship between KDL social rating indexes and corporate performance. Wang et al.

found that companies that implemented CSR had higher corporate efficiency. In a study

of small and medium companies, Torugsa et al. (2013) found that firms that

synergistically integrated the economic, social, and environmental dimensions of CSR

initiatives showed enhanced financial performance.

Boesso, Kumar, and Michelon (2013) studied the relationships between corporate

performance and descriptive, instrumental, and strategic approaches to CSR. Boesso et

al. found an association between all three CSR approaches and corporate performance,

but in different ways. Helpful CSR had a positive association with short-term measures

of corporate performance. Strategic CSR influenced short and medium-term corporate

performance. Descriptive CSR had no definite association with corporate performance.

Lu, Wang, and Lee (2013) demonstrated that CSR had a positive effect on the

financial performance of US semiconductor companies. Mallin, Farag, and Ow-Yong

(2014) studied 90 banks across 13 Islamic countries. Lu et al. discovered a positive

association between CSR disclosure and financial performance and those banks with

better financial performance were more inclined to disclose their CSR performance.

Gallardo-Vázqueza and Sanchez-Hernandez (2014) defined a scale to measure the social,

economic, and environmental dimensions of CSR and competitive success. The results of

sampling 67 regional divisions of medium and large firms concluded positive CSR had a

significant and positive effect on those firms’ overall competitive success, indicating

positive financial performance.

In the comprehensive search of articles published in 2013 and 2014, one studied

indicated a peripheral positive relationship between CSR and financial performance.

Chun, Shin, Choi, and Kim (2013) studied the relationship between corporate ethics and

financial performance. The results did not indicate a direct link between corporate values

and firm performance. However, the researchers did find that the corporations with

higher internal ethical standards enjoyed increased levels of collective organizational

commitment by the employees, which provided meaningful improved financial

performance.

In the comprehensive search of the literature published in 2013 and 2014, three

studies indicated no relationship between CSR and financial performance. Belu and

Manescu (2013) used a Data Envelopment Analysis model to evaluate the effects of CSR

on profitability. Their analysis did not confirm the positive relationship that the

researchers were expecting, but the result also did not find evidence to oppose a positive,

or negative, relationship. Ducassy (2013) studied whether CSR improved financial

performance during times of economic crisis. He found that at the beginning of the 2007

financial crisis, there was a significant positive effect, however after the first six months;

there was no longer a significant connection between CSR and financial performance.

Erhemjamts, Li, and Venkateswaran (2013) found that firms with better financial health,

performance, and R&D were more likely engaging in CSR activities than firms with

poorer financial health, performance, and R&D intensity.

In a review of the literature about CSR and corporate financial performance,

Goyal et al. (2013) determined that between 1992 and 2011, most of the research

analyzed the relationship in developed countries. However, Julian and Ofori-Dankwa

(2013) studied the effects of spending money on CSR initiatives to financial resource

availability in the emerging economy of sub-Saharan Africa. Contrary to developed

nations, Julian and Ofori-Dankwa found a significant negative relationship between CSR

expenditures and return on sales, return on equity, and firm net profitability. Their study

indicated that firms in developing nations were profit motivated to spend less of their

financial resources on CSR activities than were firms in developed countries.

Overview of Exploratory Factor Analysis and its Application

Factor Analysis (FA) is a statistical data reduction method. There are two types of

Factor Analyses: Confirmatory Factor Analysis (CFA) and Exploratory Factor Analysis

(EFA). Below is a summary of both CFA and EFA literature, with emphasis on EFA and

how researchers apply EFA since this was the procedure utilized in this research study.

The key term in FA is “factor”. By definition, a factor is an abstract concept or

latent variable such as intelligence. Researchers cannot directly measure abstract

concepts, such as intelligence. To measure such abstract concepts, researchers must

administer different types of questions or even tests to assess them in a holistic or

composite fashion.

The fundamental idea of both CFA and EFA is that multiple observed Likert

variables have similar patterns of responses because of their association with an

underlying latent variable (the factor) which researchers cannot easily measure. For

example, people may respond similarly to Likert questions about income, education, and

occupation, which are all associated with the latent variable “socioeconomic status.”

Factor Analysis. Factor analysis is a group of statistical methods used to understand

and simplify patterns of relationships underlying measured variables (Beavers et al.,

2013; Fabrigar, Wegener, MacCallum, & Strahan, 1999; Schmitt, 2011). For over sixty

years, social scientists have used factor analysis. Researchers use factor analysis to

develop theories, develop instruments, analyze longitudinal data, compare group means,

and evaluate the validity of measures (Beavers et al., 2013; Fabrigar et al., 1999;

Schmitt, 2011). Factor analysis is a concept that includes both exploratory factor analysis

(EFA) and confirmatory factor analysis (CFA) (Jennrich & Bentler, 2011).

CFA tests whether a known factor model can predict a set of observed data

(DeCoster, 1998). Researchers use CFA to verify or confirm hypotheses or theory

(Ruscio & Roche, 2012; Schmitt, 2011). Researchers also use CFA to establish the

validity of the factor model (Ruscio & Roche, 2012; Schmitt, 2011). CFA is used to

compare two models using the same data (Ruscio & Roche, 2012; Schmitt, 2011). CFA

is a good way to test the significance of factor loading (Ruscio & Roche, 2012; Schmitt,

2011). Researchers use CFA to test relationships between factor loadings and tests for

correlation or lack of correlation of factors (Ruscio & Roche, 2012; Schmitt, 2011).

Finally, CSA assesses the convergent and discriminate validity of measures (DeCoster,

1998).

EFA tests the number of common factors that influence measures and tests the

strength and relationship between each common factor to the corresponding measure

(DeCoster, 1998). Researchers use EFA for several reasons. First is to identify the nature

of constructs that underlie responses (DeCoster, 1998). Second is to determine sets of

items that interconnect (DeCoster, 1998). Third is to demonstrate the depth and breadth

of measurement scales (DeCoster, 1998). The fourth is to classify the most important

features of a group of items (DeCoster, 1998). Fifth is to generate factor scores that

represent the underlying constructs (DeCoster, 1998). Researchers commonly use EFA to

develop and validate self-reporting assessment instruments, especially when there is little

or no a priori knowledge of the structural model (Williams et al., 2012). Researchers also

use EFA to evaluate the construct validity of a measurement scale (Williams et al., 2012).

This statistical analysis is necessary to determine the critical factors that should comprise

a standard composite index. Therefore, EFA was the technique most appropriate for this

study. The remainder of this section provides an overview of the EFA technique and then

provides examples of published EFA studies to demonstrate how researchers apply the

technique.

To summarize, EFA allows researchers to discover constructs or concepts that are

not directly measurable by collapsing a large number of Likert variables into a few

interpretable underlying factors representing such constructs. Researchers use CFA more

often than they use its exploratory counterpart. Researchers use EFA to reduce or identify

the minimal number v of Likert variables to form a set of k factors or latent variables.

Researchers use CFA to determine the extent to which the given set of v Likert questions

measures k predefined factors. EFA is a data reduction method or an inductive theory

procedure. CFA is a procedure for testing hypotheses deduced from that theory.

Overview of the EFA technique. When using EFA appropriately, researchers

must make a series of methodological decisions and subjective judgments (Conway &

Huffcuit, 2003; Izquierdo, Olea, & Abad, 2014). Decisions and subjective judgments

directly affect results, interpretations and reporting (Conway & Huffcuit, 2003; Izquierdo

et al., 2014). In all steps of EFA, there are numerous statistical methods available

(Izquierdo et al., 2014). Using methods that are not optimal for the data can yield

misleading results (Conway & Huffcutt, 2003; Fabrigar et al., 1999; Izquierdo et al.,

2014). Fabrigar et al. demonstrated that EFA was a strong methodology that yielded

significant results, however, questionable decisions about the statistical procedures used

led to misleading results.

Exploratory factor analysis involves a series of statistical analysis steps. Beavers

et al. (2013), Schmitt (2011), and Williams et al. (2012) expanded on the five-step

process. The first is to analyze the factor analysis descriptive statistics. In this stage, the

researcher determines if the data is suitable for EFA. Assumptions about the suitability of

the data include examining correlational values and linear relationships (Beavers et al.,

2013). In EFA, there are no dependent or independent variables; therefore, normality is

not required for EFA (Beavers et al., 2013). The second step is to extract factors. The

linear combinations resulting from this first extraction are the factors (Beavers et al.,

2013). The third step is to determine which factors to retain. The researcher must decide

which factors best represent the data and the relationships, and determine which are not

statistically or theoretically relevant. Once identified, the researcher will retain only those

factors that best represent the data (Beavers et al., 2013; Schmitt, 2011; Williams et al.,

2012). Identifying the optimal number of factors to retain is crucial. Retaining too few or

too many factors affects the stability of factor patterns and interpretation (Hayton, Allen,

& Scarpello, 2004; Preacher, Zhang, Kim, & Mels, 2013). The fourth step is factor

rotation. Because there are an infinite number of solutions, the factors are rotated to

achieve a simpler structure and produce a solution that is easier to interpret (Beavers et

al., 2013; DeCoster, 1998). The fifth step is to interpret the factor structure. EFA is an

iterative process that requires the researcher to interpret both the items and the factors

(Beavers et al., 2013). Each measure linearly relates to each factor (DeCoster, 1998). The

factor loadings revealed by the factor rotation indicate the strength of the relationships

(DeCoster, 1998). DeCoster (1998) included the sixth step of constructing factor scores

for use in developing a composite index. The final seventh step is to apply the final index

to rank companies as to the maturity level of global business citizenship they have

achieved.

EFA Applications. Published EFA studies demonstrate the practical

implementation of the EFA technique. Social science investigators use EFA to research

various issue. EFA is good for testing relationships. Researchers use EFA to determine

whether an instrument developed for one population generalizes to other populations.

EFA helps researchers analyze whether existing instruments are applicable for different

purposes. EFA is good for testing the effectiveness of existing instruments. Most

importantly for this study, researchers use EFA to develop new measurement indexes.

The following EFA studies illustrate these applications of EFA.

Test relationships. Social scientists use EFA to test relationships. EFA reveals

relevant factors and reduces the number of items to those that accurately describe

relationships. Three recent studies illustrate how researchers use EFA in this context.

Rostamnezhad, Zarei, and Jalali (2014) used an EFA approach to testing the

impact of technological entrepreneurship on economic development. The authors defined

technological entrepreneurship as exploiting technological advances to create and

commercialize innovative products. The authors contended that successful technological

entrepreneurship translated to a competitive advantage both in the home country and

abroad. The authors surveyed teachers and experts familiar with technological

entrepreneurship and economic development. Rostamnezhad et al. applied EFA to

identify casual relationships and the latent structure to develop a conceptual model. The

results suggested that technological entrepreneurship had a significant positive effect on

economic development. With this study, the authors tested the relationship that

technological entrepreneurship had on the economic development.

Lizote, Verdinelli, and Silveira (2013) used EFA to determine whether

organizational factors positively related to the entrepreneurial competencies of

employees. In performing a literature review, Lizote et al. found that organizations that

promoted and developed entrepreneurial competencies had improved organizational

performance and quickly adapted to changing business environments. Lizote et al. used

Moriano et al.’s model of five dimensions of organizational factors together with Lenzi’s

model of eight entrepreneurial competencies to develop a 5-point Likert scale of 59

questions. The results of Lizote et al.’s study confirmed that organizational factors

positively related to entrepreneurial competencies. From the two models, Lizote et al.

determined that the entrepreneurial competency of the search for opportunities and

initiatives correlated with the organizational factor of support from top management.

Support from upper management also correlated significantly with demand for quality

and efficiency. Results also indicated that managers should clearly establish goals and

plans. Interestingly, a factor with no influence in any area was the use of rewards. Lizote

et al.’s research were an important initial part of a broader study to understand

relationships between organizational factors and entrepreneurial competencies.

To understand what service dimensions affect Chinese air passengers’

satisfaction, Wang, Qiu, Wang, and Wu (2014) conducted a study to determine the needs

and expectations of Chinese air passengers. After conducting a literature review, the

authors identified 16 key service areas. Wang et al. designed a questionnaire to capture

the perceptions passengers had of the service areas. The researchers administered the

survey to participants after they deplaned at Nanjing Lukou International Airport and

received 335 completed responses. Through EFA analysis, the authors identified

fivedimensional factors: in-flight comfort; flight core benefit; travel service and

flexibility price; equipment and relational benefit; and price. With this study, the authors

tested the relationship between air travel satisfaction and 16 key service areas.

Dhurup, Mafini, and Masitenyane (2014) used an EFA approach to study factors

that influenced customer satisfaction in the precision concrete products industry. Based

on an analysis of 260 responses, the researchers reduced the question to an 18-item scale

with the five factors of responsiveness, problem-solving, physical aspects, service

personnel and physical appearance. The researchers acknowledge that the result did not

generalize because the respondents were all customers of one company. However, the

study does provide insight into the relationship between the extracted factors and

customer satisfaction within the industry.

Generalizing to alternate populations. Researchers also use EFA to determine

whether an instrument developed for one population generalizes to other populations.

Zourbanos, Dimitrious, Goudas, and Theodorakis (2014) used EFA to determine whether

Lawrance’s 1989 Smoking Efficacy Scale (SES) generalized to the Greek population.

After translating the SES instrument into Greek, Zourbanos et al. (2014) administered the

survey to three samples of high school students between the ages of 16 and 17. The

researchers included 536 participants. EFA showed that the results are loaded on three

factors including opportunity, friends, and emotion. The results were consistent with

Lawrance’s original English language instrument designed for adults. Zourbanos et al.

concluded that their Greek version of the SES was a valid and reliable self-reporting

instrument to assess smoking self-efficacy in adolescents. Zourbanos et al. concluded that

SES was a promising tool to understand how to influence Greek adolescents’ ability to

resist smoking or to reduce or quit smoking once they had started.

Dardas and Ahmad (2014) validated whether the often-used World Health

Organization’s Quality of Life Questionnaire-BREF (WHOQOL-REF) was effective in

evaluating the quality of life parents of autistic children. The researchers administered the

WHOQOL-REF to 184 participants with autistic children. After applying EFA, the

researchers determined that the 4-domain model of the questionnaire be useful when they

redistributed the domains. The researchers tested the effectiveness of the existing

instrument and found it valid for the population of parents of autistic children.

Using existing scales. Researchers use existing scale instruments to conduct new

research. Researchers may use a single instrument or combine multiple instruments into

one survey. Three studies illustrate the use of EFA in this context.

Chaudhary (2014) used EFA to examine psychometric properties using the

existing Occupational Self-Efficacy Scale (OSES). The purpose was to identify an

alternative multidimensional occupational scale. Chaudhary’s EFA revealed three of the

six factors from the original scale to be consistent with self-efficacy. Chaudhary

concluded that the three-factor structure was superior to other self-efficacy measurement

models.

Kursunluoglu (2014) studied how customer service affected customer satisfaction

and loyalty. Kursunluoglu created a 51-item questionnaire by combining two existing

instruments, the ACSI scale, and the Customer Loyalty Scale. After applying EFA to the

data, Kursunluoglu removed nine items and identified eight primary factors; however,

only four of them affected customer satisfaction and loyalty. The factors that affected

satisfaction and loyalty were incentives, payment options, atmosphere, and employee

encounter.

Burchell and Tumawu (2014) combined four existing survey instruments to assess

employee motivation and work ethic in Ghana. The researchers combined the British

Household Panel Survey, the European Community Household Panel Survey, the British

Social Attitudes Survey, and the Workplace Employment Resources Survey into one

instrument. The survey assessed teachers and banking professionals’ attitudes toward

working hard to help one’s organization and their attitudes toward the importance of

work ethic. The results of the EFA suggested that employees in the private sector had a

higher level of commitment to their employers and a higher work ethic than those

employed in the public sector.

Test the effectiveness of a measurement instrument. Researchers use EFA to test

the effectiveness of an existing measurement instrument. In the state of Virginia, policy

makers based funding decisions on the Phonological Awareness Literacy Screening for

Grades 1 through 3 (PALS 1-3) (Huang, 2014). Huang set out to test three models that

could represent the factor of PALS 1-3, a one-factor model, a two-correlated factor

model, and a bifactor model. The result of the EFA indicated that the bifactor model was

the best fit and provided for the best generalizability and stability. Huang tested the

effectiveness of the instrument and found the bifactor model was the best fit.

Develop/refine a new measurement instrument. Since the focus of this research is

to use EFA to develop a new survey instrument and accompanying composite index, it is

appropriate to include studies in which researchers used EFA in this context. The

following five studies published in 2014 illustrate how researchers use EFA to develop

new instruments.

Shaw, Kristman, Williams-Whitt, Soklaridis, Huang, Côté, and Loisel (2014)

used EFA to develop a new Job Accommodation Scale. Their scale assessed temporary

job modifications for people returning to work after a medical leave for lower back pain.

Through their EFA, the authors identified five underlying factors including modification

of physical workload, modification of work, the environment, change of work schedule,

alternative work, and arranging for assistance. The authors concluded that the result

supported the applicability, reliability, and validity of their Job Accommodation Scale.

Liou and Kuo (2014) developed an instrument to measure senior high school

students’ motivation and self-regulation toward learning science via technology-based

methods versus in-person methods. Liou and Kuo used EFA to determine the validity and

reliability of the Motivation and Self-Regulation Toward Technology Learning (MSRTL)

instrument. The researchers used 909 completed surveys for their EFA analysis. The

results of the EFA confirmed seven scales for technology learning, including selfefficacy,

value, active learning strategies, environmental stimulation, goal-orientation, self-

regulation-triggering, and self-regulation-implementation. The results also indicated that

male and female participants did not tend toward the same preferences for all of the

scales.

To explore what achievements, skills, and personal attributes made college

graduates most employable, Pool, Qualter, and Sewell (2014) developed a new

CareerEDGE Employability Development Profile (EDP) instrument. With 807 student

participants, Pool et al. used EFA to determine that there were five factors. The five

factors included: emotional intelligence and self-management; academic performance

and study skills; career development learning, problem-solving skills; and work/life

experience. The authors contended that their self-assessment instrument would help

students at higher learning institutions determine the factors that would help them secure

jobs when they graduate.

Researchers have linked excessive Internet use to sleep disorders, personal injury,

depression, and poor social and academic adjustment. Jelenchick, Eickhoff, Christakis,

Brown, Zhang, Bensen, and Moreno (2014) developed the Problematic and Risky

Internet Use Screening Scale (PRIUSS). The PRIUSS applied to adolescents and young

adults and provided practitioners with a tool to help prevent such disorders. After

developing a survey and collecting data, the authors used EFA to explore the factor

structure and reduce the number of items. The final scale was an 18-item instrument.

Through rigorous EFA and CFA, the authors felt the new instrument was a reliable

representation of the theoretical framework and was a strong fit for the empirical data.

Practitioners can use the scale to screen for excessive Internet use, which can lead to

preventative care.

Fullwood, Nicholls, and Makichi (2014) sought to expand on the research about

what motivates people to blog. The researchers developed the Blogging Motivations

Questionnaire (BMQ). For the study, Full et al. also used the International Personality

Item Pool (IPIP). The IPIP provided the researchers with a reliable measure of five

personality traits including extraversion, agreeableness, conscientiousness, emotional

stability, and openness. The researchers applied EFA to the results of administering the

BMQ to 160 blogging participants. The EFA technique reduced the number of factors

and determined factor loadings. The results indicated people blog for six primary reasons

including personal revelation, an emotional outlet, creative outlet, selective disclosure,

social networking, and advertising. Correlating the BMQ results with the IPIP results, the

researchers discovered that conscientiousness predicted social networking; agreeableness

predicted selective disclosure, and openness predicted creative outlet. Fullwood et al.

developed a new instrument and combined it with an existing scale to produce new

results.

EFA has several practical uses for social scientists. Social science researchers use

EFA to test relationships, determine whether an instrument developed for one population

generalizes to other populations, use existing instruments for different purposes or test

the effectiveness of existing instruments. Most importantly for this study, researchers use

EFA to develop new measurement indexes. Using EFA was an appropriate approach to

developing a survey instrument and composite index to understand the maturity level

corporate leaders have achieved toward implementing the GBC framework.

Transition

Section 1 of this doctoral study established the background, problem, purpose,

and nature of the study. Section 1 presented the research questions, hypotheses,

theoretical framework, and significance of this study. Additionally, section 1 provided a

synthesis of the literature. The literature review included studies that supported the

development of the GBC theory, established the case for GBC, described CSR

measurement techniques, and provided an overview of EFA and published applications

of the technique. As explained in section 1, there was a clear need to establish a

measurement index for GBC. Section 2 describes the research design and quantitative

methodology used for this study.

Section 2: The Project

Research shows that responsible and ethical business policies and actions have the

potential to provide companies with legitimacy from society and increase their

competitive performance (Menck & Oliveira, 2014). Implementing the GBC framework

orients corporations in a way that enhances legitimacy by providing a common

foundation for socially acceptable behavior at the individual, organizational, and

systemic levels (Wood et al., 2006). While there have been valuable contributions toward

measuring different aspects of CSR, at the start of this study there was no

selfadministered rating system available for business leaders to report to stakeholders the

steps they have achieved toward becoming an ethically responsible business citizen

(Milne & Gray, 2013).

Purpose Statement

The purpose of this cross-sectional quantitative study was to develop Likert

survey questions (independent variables) and apply EFA to reveal factors (dependent

variables) and assign weights to questions to develop a self-administered rating system to

measure the GBC theory, which assesses the maturity level a company has attained

toward becoming a global business citizen. Development of this rating system required

four methodological steps. First, creating a survey consisting of 1 qualifying question and

22 Likert questions that operationalized the GBC principles of VALUE, IMPLE,

ANALY, and LEARN. Second, administering the survey to members of professional

associations who were senior executives of U.S. corporations with an understanding of

GBC. Third, applying the EFA statistical method to the data. EFA revealed the

relationship between the Likert survey questions and the factors that emerged, reduced

the questions, and assigned weights to the remaining questions. Fourth, use the EFA

assigned weights to develop a composite index. The result of this study provided a rating

system to measure a company’s GBC maturity level. This study contributes to social

change by providing practitioners, academics, and stakeholders with a rating system to

evaluate the maturity level that corporate leaders have attained toward becoming a global

business citizen.

Role of the Researcher

Such things as researchers’ personal bias, experiences, beliefs, and even their

approach can influence research (Hunt, 2011). In this respect, to be as transparent as

possible, I disclose no prior experience in the area of study and my interest in this field

was purely academic. My role as the researcher was to develop, administer, and collect

data from a Likert-type survey instrument and then analyze and report the results

(Croasmun, & Ostrom, 2011). Within my professional role, I had access to business

executives via Executive Suite. As part of the data collection process, participants

received an informed consent form (Appendix C) providing information, ascertaining

comprehension, and ensuring they were participating voluntarily (U.S. Department of

Health & Human Services, 1979). The protocol complied with the ethical principles for

the protection of human subjects of research, in compliance with the Belmont Report

(U.S. Department of Health & Human Services, 1979). Additionally, as the researcher, I

handled the validation and reliability of the instrument and interpretation of the data

collected.

Participants

The process of implementing the four-step GBC framework assumes that business

leaders develop corporate values, implement those values, engage in problem analysis

and experimentation, and learn from the previous steps to institutionalize best practices.

As such, business leaders with knowledge of these constructs were appropriate

participants for this research. The scope of the study focused on medium to large,

multinational business organizations headquartered in the United States.

Research shows that soliciting participants associated with professional societies

improves survey response rates significantly (Melnyk, Page, Wu, & Burns, 2012;

VanGeest & Johnson, 2011). Informed by this research, participants forming the sample

frame were business executives who were members of the Executive Suite professional

society. Executive Suite is an invitation-only, online forum for executive level business

leaders. At the time of survey distribution, Executive Suite had 298,841 members.

Members of Executive Suite should have had the knowledge of and been

competent to identify items that may demonstrate the constructs. These business leaders

should have had direct experience developing corporate values, implementing those

values, analyzing issues related to the values and implementation, and learning,

systemizing, and institutionalizing best practices.

Personalized delivery of surveys significantly improves response rates (Melnyk

et. al., 2012; Sahlqvist, Song, Bull, Adams, Preston, & Ogilvie, 2011; Sinclair, O’Toole,

Malawaraarachchi, & Leder, 2012; VanGeest & Johnson, 2011). Informed by this

research, I distributed the survey via online communication to the entire membership of

Executive Suite. All members had the opportunity to take a survey. Only those interested

in participating in the study clicked through to the online survey.

The working relationship of this study was to ensure the anonymity and

confidentially of participants. The survey did not ask for any personal information that

could identify the individual, such as name or email address. The research question was

to determine how many and what factors were needed to characterize the Likert survey

questions to assess a company’s GBC maturity level. The business leaders in this sample

had knowledge of the four constructs were able to answer the Likert survey questions.

Because the participants represented companies from a variety of industries, the results of

the study generalize to the global business community.

Research Method and Design

This quantitative research study relied on a statistical method called EFA and a

cross-sectional survey design. The design of the study was a Likert-type survey

consisting of 1 qualifying yes/no question and 22 Likert-type questions assumed to

capture the four GBC constructs. The GBC constructs were equivalent to the four-step

GBC framework. The four-step GBC framework provided the assumed constructs. The

constructs were the following. First, developing overarching corporate values (VALUE).

Second, implementing the values (IMPLE). Third, analyzing problems and

experimenting to revise the values or local implementations (ANALY). Fourth, learning

from the previous steps and institutionalizing best practices (LEARN) (Wood et al.,

2006).

In the absence of any other known quantitative studies of this kind, applying the

EFA method to the data obtained from the survey reduced the questions to a minimum

number. EFA also indicated how to group the remaining questions into subsets, called

factors. The EFA-generated eigenvalues weights on each question generated the overall

weighted formula, or index, to measure GBC implementation.

Research Method

The adopted method to develop a rating system index to measure GBC maturity

level was an extensively used statistical method called EFA (Basto & Pereira, 2012;

Izquierdo et al., 2014). Researchers commonly use EFA to develop and validate

selfreporting assessment instruments, especially when there is little or no a priori

knowledge of the structural model (Ruscio & Roche, 2012). Other quantitative

techniques could have produced a subset of questions to understand the maturity level

companies had achieved in implementing GBC. These include using Confirmatory Factor

Analysis (CFA) instead of Exploratory Factor Analysis (EFA) or the Delphi technique.

The following is an explanation of why EFA was superior to these techniques for this

study.

CFA tests whether a known factor model can predict a set of observed data

(DeCoster, 1998). CFA requires that researchers specify a particular factor structure in

advance and then designate which items load on which factor. CFA is a model in which a

specific item (question) maps to a specific construct. In contrast, EFA allows all items to

load on all factors. When there is little or no a priori knowledge of the structural model,

EFA is the preferred method to reveal the item-to-factor structure (Ruscio & Roche,

2012). As such, researchers commonly use EFA to develop and validate self-reporting

assessment instruments. Researchers use CFA to verify or confirm hypotheses or theory

(Ruscio & Roche, 2012; Schmitt, 2011). CFA is good for establishing the validity of the

factor model (DeCoster, 1998). CFA is a method well suited to comparing two models

using the same data (DeCoster, 1998). Researchers use CFA to test the significance of

factor loading, to test relationships between factor loadings, and to test for correlation or

lack of correlation of factors(DeCoster, 1998). CFA is also used to assess the convergent

and discriminate validity of measures (DeCoster, 1998). A recommendation for further

study was to conduct a CFA study to validate the results of this EFA study.

SMEs could have validated that the survey questions adequately represented the

process of implementing GBC. As an extension of this method, a group of SME may

have collected and aggregated information systematically via a Delphi technique (Hasson

& Keeney, 2011). These methods were not appropriate for several of reasons. The first

was defining the meaning of SME (Hasson & Keeney, 2011). Defining who is an expert

in GBC could be a research study itself. Secondly, recruiting a panel of supposed experts

to participate would have been challenging (Hasson & Keeney, 2011). Third, the level of

influence some SMEs may have had on other members of the panel may have skewed

results (Hasson & Keeney, 2011). Fourth, establishing anonymity of the member SMEs

would have been difficult (Hasson & Keeney, 2011). Fifth, it would have been difficult

to define what constitutes a consensus (Hasson & Keeney, 2011). Sixth, determining

what criteria to include would have been uncertain (Hasson & Keeney, 2011). Finally,

the influence of personal bias amongst the SMEs would have been a limitation (Hasson

& Keeney, 2011).

Because EFA is a multivariate statistical approach, it provided an unbiased

method for reducing the number of factors, examining relationships between factors, and

evaluating the construct validity of measurement scales (Williams et al., 2012). The EFA

method should ensure reliability. Measurements should repeatedly produce the same

results, the measurement should be stable over time, and the measurements should be

similar within a given period (Hasson & Keeney, 2011).

The quantitative EFA method was appropriate in its rigor because it involved a

series of statistical analysis steps (Basto & Pereir, 2012; Beavers et al. 2013; Schmitt,

2011; Williams et al., 2012). The following is a brief overview of the EFA method. The

Data Analysis section of this paper provides details of each step. The first step is the

planning step. During this step, the investigator determines if the data obtained by

administering the survey to a participant group is suitable for EFA (Beavers et al. 2013;

Schmitt, 2011; Williams et al., 2012). Once determined that the data be suitable, the

second step is to extract factors (Basto & Pereir, 2012; Beavers et al. 2013; Schmitt,

2011; Williams et al., 2012). These factors are the key questions that best describe each

of the four steps of implementing GBC or each of the constructs. The third step is to

identify the number of factors to retain (Basto & Pereir, 2012; Beavers et al. 2013;

Schmitt, 2011; Williams et al., 2012). The retained factors are those that best represent

the data and the relationships (Basto & Pereir, 2012; Beavers et al., 2013; Schmitt, 2011;

Williams et al., 2012). Retaining the optimal number of factors is crucial because

retaining too few or too many factors affects the stability of factor patterns and

interpretation (Hayton et al., 2004; Preacher et al., 2013). The fourth step is to rotate the

factors (Basto & Pereir, 2012; Beavers et al., 2013; Schmitt, 2011; Williams et al., 2012).

Because there are an infinite number of solutions, the EFA technique includes rotating

the factors to achieve a simpler structure and produce a solution that is more readily

interpretable (Beavers et al., 2013; DeCoster, 1998). The fifth step is to interpret the

factor structure (Beavers et al. 2013; Schmitt, 2011; Williams et al., 2012). The factor

loadings revealed by the factor rotation indicate the strength of the relationships

(DeCoster, 1998). Researchers interpret the data to find the delicate balance between

statistical significance and conceptual relevance (DeCoster, 1998). When developing a

composite index, the sixth step is to construct factor scores (DeCoster, 1998). The

EFAgenerated eigenvalues weights are multiplied with the corresponding Likert response

and then summed to obtain an overall index (DeCoster, 1998). In summary, the EFA

method provided an unbiased method for reducing the number of factors, examining

relationships between factors, evaluating the construct validity of a measurement scale,

and developing a composite index (Williams et al., 2012). A qualitative or less rigorous

quantitative method was not conducive to achieving these outcomes.

Research Design

The research question was how many and what factors (dependent variables) were

needed to characterize the Likert survey questions (independent variables) to assess a

company’s GBC maturity level? The study relied on a cross-sectional survey design

appropriate for EFA data collection to answer the research question. The survey

attempted to capture the assumed constructs of the GBC theory. A cross-section sample

was representatives of the business community familiar with the concept of GBC. The

sample also had to be large enough to apply the EFA technique.

In this research study, a cross-sectional survey design for data collection was

appropriate for the EFA method. Researchers commonly use Likert-type survey

instruments to collect data for quantitative EFA research (Harrison & Reilly, 2011;

MacKenzie et al., 2011). Researchers use cross-sectional surveys to measure constructs,

or the abstract and latent ideas or themes (MacKenzie et al., 2011). In survey research,

survey questions attempt to articulate the common characteristics of the constructs

(MacKenzie et al., 2011).

For this study, the actual survey consists of 1 qualifying yes/no question and 22

Likert-type questions. The questions were designed to attempt to capture the assumed

four constructs of the GBC implementation framework. The constructs were corporate

values (VALUE), implementation (IMPLE), problem analysis and experimentation

(ANALY) and learning and institutionalizing best practices (LEARN). Donna J. Wood,

the lead researcher who developed the theory of GBC, agreed that these were the

necessary and sufficient constructs of the GBC theory (D. J. Wood, personal

communication, August 14, 2014; Appendix E). The survey questions attempted to

articulate the common attributes/characteristics of each step, or construct, and may be

necessary to evaluate whether a company has attained implementation of each step.

Unlike constructs being necessary and sufficient, only one question may have been

necessary and sufficient to describe a construct. Alternatively, questions may have been

neither necessary nor sufficient. Unnecessary questions received low EFA loadings.

However EFA was not able to identify missing questions, and future researchers might

need to add such questions. The Data Collection Instruments section details the

development of the survey and the Likert-type questions. In summary, applying EFA to

the resulting data determined which questions were required to evaluate the maturity

level that a company has attained in implementing the four-step GBC framework.

Archival data to support this original research was not available. Even if archival

data related to business citizenship were available, use of archival data might

compromise the study. Use of archival data might have compromised the study if such

data did not match the constructs, the range of measures, scope, or breadth of this study

(Rabinovich & Cheon, 2011).

The study was cross-sectional, meaning the sample represented a cross-section of

the population for which the measure was designed (MacKenzie et al., 2011). In this

case, I developed the instrument for executive leaders of multinational business

organizations, not limited to any particular industry. The survey was deliberately short to

improve response rates (Meade & Craig, 2012; Sahlqvist et al., 2011). Since the sample

was crosssectional, it should have represented the population so that the results should

generalize to the broader population (MacKenzie et al., 2011).

Population and Sampling

It was necessary to use a representative sample to generate results that apply to

businesses. That is, the results should fulfill the requirement for a cross-sectional

representation of various business views on GBC. Given this rationale, the following

sections describe the population from which the sample came. The discussion also

calculates the required sample size required by EFA.

Population

The sample units consist of corporate leaders interested in business citizenship.

Ideally, one should sample from the worldwide population of corporate leaders, but this

was impractical due to its geographic scope and the time and expense required reaching

participants (Baltar & Brunet, 2012). Instead, sample participants were comprised of

business executives who were members of Executive Suite. Membership of this

professional society included business executives, senior professionals, and emerging

leaders, all leaders who may have an interest or at least an understanding of business

citizenship. Also, the sample size satisfied the needs of EFA.

Population aligns with the overarching research question

The overarching RQ was how many and what factors (dependent variables) were

needed to characterize the Likert survey questions (independent variables) to assess a

company’s GBC maturity level? The sample had to understand GBC to provide data that

was meaningful to answer the RQ. Of the top 100 U.S. companies, 97% claim to engage

in business citizenship activities (Fifka, 2013). Major companies such as Boeing, Dow,

IBM, and Microsoft claim that they are business citizens (Crittenden et al., 2011). Senior

executives of these and other leading U.S. corporations comprise the membership of the

Executive Suite. Therefore, the sample units had sufficient knowledge of the dimensions

or factors and their strength that comprise GBC theory to provide data to answer the

overarching RQ.

Sampling Method

Given that the intent was to use this study to generate results that apply to global

businesses, a cross-sectional representation of various business views on GBC was ideal.

A probabilistic sample was required to achieve cross-sectional representation (Trotter,

2012). In probabilistic sampling most often used with quantitative methods, one selects

cases that are together representative of the total population, even though many cases in

isolation will have a low information value (Trotter, 2012). These cases are sample units

or individual corporate leaders. Sample frames identify all of the sample units, or

members, of the target population (McLeod, Klabunde, Willis, & Stark, 2013). In this

study, the sample frame was members of Executive Suite. As previously explained, the

geographic scope and the time and expense required to reach the worldwide population

of corporate leaders interested in business citizenship was prohibitive. Therefore, the

sample units were members of Executive Suite. At the time of survey distribution,

Executive

Suite had 298,841 members. For this study, the sample frame equaled the population.

By contrast, qualitative methods often use a narrow or purposeful sample to

addresses specific purposes related to the research questions. In qualitative research, each

case is selected to address a particular set of questions. With this tight parameter, each

case has a high information content/value (Palinkas, Horwitz, Green, Wisdom, Duan, &

Hongwood, 2013; Trotter, 2012). However, such a narrow sample is usually not

representative, limiting the research findings to that particular subset of participants and

the limitation of not achieving cross-sectional representation (Palinkas et al., 2013). A

probabilistic sampling method was required to achieve the cross-sectional goal of this

study.

There are multiple probabilistic sampling methods available. Probabilistic

methods include simple random sampling, systematic random sampling, stratified

random sampling, cluster sampling, multiphase sampling, and multistage sampling

(Acharya, Prakash, Saxena, & Nigam, 2013). Researchers use a random process of

selecting participants with all of these methods (Acharya et al., 2013). In the simple

random sampling technique, every individual, or sample unit, of the sample frame have

an equal chance of being selected (Acharya et al., 2013).

Of all of the aforementioned probabilistic sampling methods, the simple random

sampling technique was the most feasible to select a sample of business leaders with

knowledge and interest in GBC. Researchers commonly use simple random sampling

when it is complex, costly, or impossible to randomize to the individual level of a

population (Welton, Madan, Caldwell, Peters, & Ades, 2014). Given a large number of

multinational corporations, it was impossible to randomize to the individual level of the

population. Leaders interested in or having an understanding of business citizenship may

have been members of Executive Suite. Selecting this professional business society as the

sample may seem to be biased sampling. However, the simple randomized sampling

method is recognized as an effective method of achieving a probabilistic sample

(Acharya et al., 2013; Baltar & Brunet, 2012; Welton et al., 2014).

Calculating sample size

For the EFA method, researchers determine sample sizes in two ways. One

method is to determine the minimum number of samples needed (N). Another method is

to determine the sample size as a function of the number of variables. Also known as the

subjects-to-variable ratio, (N:p) (Beavers et al., 2013; Guadagnoli & Velicer 1988;

Hogarty, Hines, Kromrey, Ferron, & Mumford, 2005). There are advantages, and

disadvantages, of both methods. The subjects-to-variable ratio was best suited to

determine the sample for this study. The following is a discussion of the advantages,

disadvantages, and rationale for choosing subjects-to-variable ratio.

There are studies about selecting a minimum sample size. Jung and Lee (2011)

analyzed the factor extraction outcomes using sample sizes of less than 50. Jung and Lee

analyzed the outcomes achieved by applying maximum likelihood factor analysis

(MLFA), principle component analysis (PCA) and regularized exploratory factor analysis

(REFA). The researchers found that REFA recovered acceptable factor loadings, had

smaller mean absolute differences and mean square errors, and provided stable factor

loading estimates with samples of 50 or less. Fabrigar et al. (1999) determined that

samples as small as 100 could yield stable solutions. Beavers et al. (2013) recommended

using samples of at least 150 for multivariate tools, such as EFA. Guadagnoli and Velicer

also determined that when researchers selected variables that were representative

indicators of a component, 150 observations yielded accurate solutions. At the high end

of the minimum number of samples, Guadagnoli and Velicer found that they needed 300

when few variables defined factors with moderate to low loadings.

Given this diverse range of recommended sample sizes, using a formula to

determine the appropriate sample size was appropriate. For this study, the method of

determining sample size as a function of the number of variables, (N:p), or the subject-

tovariable method, was suitable. Hogarty et al. (2005) found that a higher number of

samples were necessary when the goal of the study was to understand which factors

underlie which variables. When the study goal was to ensure that sample loadings

correlated highly with population loadings, fewer samples were necessary (Hogarty et al.,

2005). Comrey and Lee (1992) recommend a minimum of 5 observations per variable, or

ideally 20 observations per variable, for EFA. According to this empirical rule, and given

an initial survey with 22 Likert questions (22 variables), at a minimum 22*5 = 110

observations were needed to perform an EFA analysis adequately. Ideally, 22*20 = 440

observations were needed to perform an EFA analysis. Table 3 shows the minimum and

an ideal number of observations as calculated in a subject-to-variable ratio method using

22 variables. Given this wide range, the intent was to solicit the ideal sample size of 440

cases if possible but satisfy the minimum of 110 observations before undertaking the

analysis. Assuming a 10% rate for invalid surveys and an average industry response rate

of 20%, then a minimum of (110*1.10)/0.20 = 605 surveys should have been

administered. Ideally (440*1.10)/0.20=2420 surveys should have been administered.

In summary, the intended population of this study was corporate leaders of

multinational business organizations in the United States. This population was

appropriate because, collectively, the members could answer the overarching research

question. The research question was how many and what factors were needed to

characterize the Likert survey questions to assess a company’s GBC maturity level?

Members of Executive Suite were business executives. This population should have an

interest or understanding of business citizenship. Researchers recognize the simple

random sampling method as an effective method of achieving a probabilistic sample, and

other methods were cost-prohibitive or possibly even impossible. For the EFA technique,

the method of determining sample size as a function of the number of variables, (N:p), or

the subject-to-variable method, was suitable. A minimum of 110 observations was

required, with an ideal of 440 observations.

Table 3

Subject-to-Variable Ratio to Determine Sample Size

Minimum # of Ideal # of

Observations

Observations Observations

Observations per Variable 5 20

# of Observations needed ( N:p) 110 440

Number of variables (p)=22.

Ethical Research

All data collected for this study adhered to the standards set by the U.S.

Department of Health & Human Services’ Belmont Report (U.S. Department of Health

& Human Services, 1979). Additionally, all data collected complied with the standards

set by Walden University’s Internal Review Board (IRB). All participants were voluntary

and anonymous. I did not collect any data before formal IRB approval.

All participants had ample opportunity to review the informed consent form

before starting the survey. The informed consent form appears in Appendix C. The

informed consent form included information about the research procedure, the purpose,

risks, and anticipated benefits, and a statement offering participants the opportunity to

ask questions and to withdraw from the study at any time (U.S. Department of Health &

Human Services, 1979). The informed consent form contained a statement to ascertain

that participants comprehended the information (U.S. Department of Health & Human

Services, 1979). Finally, the informed consent form included a clause stating that

participants agreed to participate on a volunteer basis, free of coercion and undue

influence (U.S. Department of Health & Human Services, 1979).

Participants could withdraw from the study at any time by contacting the

researcher (U.S. Department of Health & Human Services, 1979). The informed consent

form provided the relevant contact information. However, because of the anonymous

nature of the survey, participants’ individual surveys cannot be identified. Participants

received no compensation or incentives for participating in the study. For five years from

the time of data collection, I will maintain all data in a safe, and secure location in an

anonymous format with no personal information indicated. After five years, I will

destroy the electronic data by using Secure Erase software in compliance with Walden

University’s IRB guidelines.

Data Collection Instruments

In this quantitative study, data collection begins with the instrumentation. This

section explains the construction of the 23-question instrument, Cronbach’s alpha method

to assure instrument reliability, and the methods employed to determine instrument

validity. The next step was the data collection technique. This section explains the

rationale for collecting data by personally delivering surveys to members of a

professional society. The last step was data organization techniques. This section

describes data security, retention, query approval, and destruction techniques.

Instrumentation

The first step of this quantitative research study was the construction of a

23question, 5-point Likert-type survey instrument. I constructed the instrument for use in

this study. Appendix A and Appendix B display two versions of the instrument. The

version in Appendix A shows the four constructs and the questions related to each

construct. Appendix B displays the version of the instrument for distribution to

participants.

Instrument Construction. When constructing a survey instrument, the first

consideration is the research objective and variables required to test the research

questions (Lederer, Comber, & Oswalt, 2014; Sinkowitz-Cochran, 2013; Synodinos,

2003). The ultimate objective of this quantitative study was to develop a composite index

to assess the level of global business citizenship for a given company. The literature

informed the constructs (Burton & Mazerolle, 2011; Sinkowitz-Cochran, 2013).

Specifically, the research included seminal studies and book published by the authors of

the theory of GBC. The four-step GBC framework provided the assumed constructs. The

constructs were the following. First, developing overarching corporate values (VALUE).

Second, implementing the values (IMPLE). Third, analyzing problems and

experimenting to revise the values for local implementations (ANALY). Fourth, learning

from the previous steps and institutionalizing best practices (LEARN) (Wood et al.,

2006). Donna J. Wood, the lead researcher who developed the theory of GBC, agreed

that these terms captured the process of implementing the four-step GBC framework (D.

J. Wood, personal communication, August 14, 2014; Appendix E).

When constructing a survey instrument, the second consideration was the

administration method. Administration methods include personal interviews or

selfadministered questionnaires (Sinkowitz-Cochran, 2013; Synodinos, 2003; VanGeest

& Johnson, 2011; Weigold, Weigold, & Russell, 2013). The survey construction should

fit the method of administration (Sinkowitz-Cochran, 2013; Synodinos, 2003; VanGeest

& Johnson, 2011; Weigold, Weigold, & Russell, 2013). Because the instrument for this

study was a Likert-type survey, it fell into the self-administered category. Personalized

delivery of surveys significantly improves response rates (Melnyk et al., 2012; Sahlqvist

et al., 2011; Sinclair et al., 2012; VanGeest & Johnson, 2011). The survey was an online

survey administered through Survey Monkey (Lederer et al., 2014; Synodinos, 2003). I

sent a personal online communication to all members of Executive Suite informing them

about the nature of this study and requesting their participation.

The third consideration when constructing a survey instrument is forming the

questions themselves. The factors that influence questionnaire construction include the

wording of questions, choice of response options, the sequence of questions, and the

intended audience (Synodinos, 2003). Participants tend to understand words differently

(Porter, 2011). To help participants understand the questions, the wording of the

questions should be concise and simple in structure (Lederer et al., 2014; Money, Lines,

Fernando, & Elliman, 2011; Synodinos, 2003). Additionally, each question should relate

to a single issue (Lederer et al., 2014; Money et al., 2011; Synodinos, 2003).

Response choices can include open-ended or closed-ended questions (Bartkus,

Mills, & Olsen, 2014; Lederer et al., 2014; Synodinos, 2003). Self-administered surveys

are well suited for closed-ended questions. Closed-ended questions are easier for

participants to answer, have a tendency to produce fewer missing data, and are easier to

code and analyze than open-ended questions (Sinkowitz-Cochran, 2013; Synodinos,

2003). The instrument for this study was a closed-ended Likert-type survey.

Oddnumbered Likert-type scales allow respondents to answer neutrally, which may

reduce response bias (Croasmun, & Ostrom, 2011; Sinkowitz-Cochran, 2013). The

interval Likert-type scale allowed participants the following response options: strongly

disagree, disagree, neither disagree nor agree, agree or strongly agree (Croasmun &

Ostrom, 2011). Appendix A presents the survey with questions grouped according to

each construct and their corresponding Likert-type response options.

Presenting the questions in different sequences can influence respondents’

answers (Synodinos, 2003). Some research recommends ordering questions logically,

grouping questions together within a topic, and ordering questions from general to

specific (Lederer et al., 2014; Synodinos, 2003). Other studies indicate that a presenting

the questions in a random way reduces response bias (Krumpal, 2013). In an attempt to

reduce response bias, the questions were randomized. Appendix B presents the survey in

the randomized format that participants received.

The final factor that influences questionnaire construction is the intended

audience (Porter, 2011; Sinkowitz-Cochran, 2013; Synodinos, 2003). Synodinos (2003)

noted that the researcher must be extremely well versed in the topic and understand the

capabilities of the participants. The researcher must select appropriate participants, tailor

the survey for the audience, and be able to explain why participants may not have

answered specific questions (Porter, 2011; Synodinos, 2003). Participants may not

answer questions because they did not feel the question was applicable or because they

did not understand how to respond to the question (Porter, 2011; Synodinos, 2003).

Business leaders who are members of Executive Suite represent corporate leaders

interested in business citizenship. The instrument was a short 23-question survey to

accommodate these busy executives. The first question was a qualifying question asking

participants whether they understood the concept of business citizenship, corporate

citizenship, corporate social responsibility, or the ethical responsibilities of corporations.

The yes or no answer to this question revealed the capability of participants to answer the

subsequent questions. The survey directed the participants answering “no” to this first

question to the end of the survey without answering any of the 22 Likert questions. This

information provided insight into whether unanswered questions were a result of the

participants feeling they were not applicable or their inability to answer the questions

(Synodinos, 2003).

Informed by these guidelines, the survey questions related to the four constructs

with the first question measuring each participant’s understanding of global business

citizenship. The remaining 22 Likert-type questions relate to the four constructs. The

GBC seminal studies, the book published by the theory’s authors, and an extensive

literature review informed the development of the items that may characterize each of the

constructs. Five questions related to developing a companywide overarching code of

ethical conduct consistent with the definition of GBC (VALUE). Seven questions related

to the IMPLE construct. The IMPLE construct was the implementation of the

overarching code of conduct throughout the organization and adaptation to local customs,

norms, and local ethical standards that seem in conflict with the overarching code of

ethical conduct. Four questions related to analyzing problem areas and experimenting

with creative and practical solutions to remediate conflicts (ANALY). Six questions

related to systemizing learnings from the previous steps and institutionalize the best

policies, practices, and behaviors throughout the organization (LEARN).

Instrument Reliability. In quantitative research, reliability assures that a

researcher could replicate the study by using the same methods and a similar group of

participants (Oluwatayo, 2012; Tavakol & Dennick, 2011). In this study, the burden of

reliability falls to the instrument. There are multiple methods available to test the

reliability of survey instruments. Standard alternatives include test-retest, equivalent

forms, split-half coefficients, and Cronbach’s alpha (Oluwatayo, 2012; Yang & Green,

2011).

The test-retest coefficient method assesses transient errors but requires the

researcher to administer the survey on two separate occasions (Yang & Green, 2011).

The equivalent forms coefficient also assesses transient errors but requires developing,

validating, and administering two equivalent surveys (Yang & Green, 2011). The

splithalf coefficient method uses Spearman-Brown prophecy formula to assess internal

consistency reliability, but not transient errors, from one single administration of a survey

(Yang & Green, 2011). Social scientists widely use and recommend the Cronbach’s

alpha method (Croasmun, & Ostrom, 2011; Gadermann, Guhn, & Zumbo, 2012;

Oluwatayo, 2012; Tavakol & Dennick, 2011). Cronbach’s alpha is a good test for

reliability of instruments designed to gather responses in the continuum (Croasmun, &

Ostrom, 2011; Gadermann et al., 2012; Oluwatayo, 2012; Tavakol & Dennick, 2011).

Responses in continuum include Likert scales (Croasmun, & Ostrom, 2011; Gadermann

et al., 2012; Oluwatayo, 2012; Tavakol & Dennick, 2011). The Likert scale for this study

is, strongly disagree, disagree, neither disagree nor agree, agree or strongly agree.

Because the survey for this study was Likert-type, the Cronbach’s alpha method was

appropriate to ensure the consistency and stability of the result of the scale data. After

entering the data into SPSS software, a reliability analysis of each question produced

descriptive reliability statistics for the items and the scale as well as showing the inter-

item correlations. The results of the item-total statistics indicated if removing any of the

questions would lead to higher or lower Cronbach’s alpha scores. Cronbach’s alpha

scores between 0.67 and 0.90 demonstrate acceptable reliability (Oluwatayo, 2012;

Tavakol & Dennick, 2011).

Instrument Validity. Instrument validity means determining how well an

instrument measures what it was intended to measure (Burton & Mazerolle, 2011;

Oluwatayo, 2012). There are four types of validity that inform whether of an instrument

is suitable for the intended purpose. Validity includes construct, face, content, and

criterion validity (Burton & Mazerolle, 2011; Oluwatayo, 2012). The following address

each of these types of validity.

Construct validity means that an instrument measures what it was intended to

measure (Burton & Mazerolle, 2011; Cronbach & Meehl, 1955; Oluwatayo, 2012). The

goal of the research is for the investigator to confirm or disconfirm that the instrument

measures what the investigator hypothesized it would measure (Cronbach & Meehl,

1955). Cronbach and Meehl (1955) proposed three steps to evaluate construct validity.

The first is to state the theoretical framework and assign meaning to each construct. The

second is to develop methods and empirically measure how adequately the instrument

substantiates the assigned construct meanings. The third is to interpret correlations and

present evidence and reasoning to show the reader why the correlations confirm or

disconfirm the hypothesis.

For this study, the first step in establishing construct validity involved using the

GBC theory as the theoretical framework. The set of four interrelated theoretical

concepts captured by the GBC theory informed the constructs. The four theoretical

concepts were VALUE, IMPLE, ANALY, and LEARN. As already mentioned, Donna

Wood, the

principal researcher responsible for the development of the theory, indicated that she

agreed that these were the constructs (D. J. Wood, personal communication, August 14,

2014; Appendix E). The survey questions were an attempt to assign meaning to each

construct. Five SME’s indicated that they felt the questions captured the meaning of the

constructs and were clear. Cronbach and Meehl’s (1955) second step to establish

construct validity is to develop methods and empirically measure how adequately the

instrument substantiated the assigned construct meanings. There are several methods for

examining construct validity. Validity methods include the multitrait-multimethod matrix

(Campbell & Fiske, 1959), factor analysis, and structural equation modeling (Marsh et

al., 2014). The third step is the heart of this research. The data were interpreted using the

EFA technique. Interpretation of the correlations provided evidence and reasoning to

show why the correlations confirmed or disconfirmed the hypothesis. In the case of

applying EFA, the hypothesis equated to propositions about the number of factors to

retain to capture the relevant constructs. In summary, the assumption of the construct

validity of the instrument was strongly satisfied via the original theory author’s

confirmation, SME confirmation, and EFA application.

Face validity is a subjective assessment of whether the instrument is relevant,

reasonable, unambiguous, and clear (Oluwatayo, 2012; Synodinos, 2003). Having a panel

of subject matter experts validate the questions is a recommended way to address face

validity (Burton & Mazerolle, 2011; Oluwatayo, 2012). Accordingly, five experts in the

field of CSR or with a familiarity of global corporate or business citizenship confirmed

that the 22 Likert-type questions were relevant, reasonable, unambiguous, and clear

(Appendix F).

Content validity demonstrates that the measure covers the range of meanings that

apply to the constructs (Oluwatayo, 2012). A frequently used method to address content

validity is factor analysis (MacKenzie et al., 2011; Oluwatayo, 2012; Podsakoff et al.,

2012). The research method of this study was EFA. Therefore, the data analysis validated

the content and mitigated content validity threats.

Researchers use criterion-related validity to demonstrate that the scores from the

new instrument correlate highly with scores from existing instruments that are already

determined to be valid (Oluwatayo, 2012). No existing instruments measure GBC.

Because there were no existing instruments to correlate with, it was beyond the scope of

this study to confirm criterion-related validity. Confirming criterion-related validity was

a recommended area for further research.

Data Collection Technique

The method of survey delivery can have a significant impact on response rates

(Melnyk et al., 2012). Personally delivering surveys improves response rates significantly

(Melnyk et al., 2012; Sahlqvist et al., 2011; Sinclair et al., 2012; VanGeest & Johnson,

2011). Soliciting participants from professional societies also significantly improves

survey response rates (Melnyk et al., 2012; VanGeest & Johnson, 2011). Informed by

this research, participants forming the sample frame were business executives who were

members of Executive Suite. An email Letter of Cooperation from Anthony Vlahos, the

owner of the Executive Suite society, appears in Appendix D.

I distributed the survey via online communication to the entire Executive Suite

membership. The online survey directed participants answering “yes” to the first

qualifying question to the informed consent. Participants had the opportunity to review

the conditions of the informed consent before proceeding to the Likert questions.

Appendix B displays the version of the survey that was disturbed to participants.

Appendix C presents the accompanying informed consent form.

An alternative to personally distributing surveys via electronic communication at

renowned professional societies could have been to email surveys to a distribution list of

executives. There are numerous advantages of online surveys. With online surveys, data

is instantaneously stored in a database. Delivery costs may be lower. There is the ability

to offer multiple languages. The data collection process may be faster. Questionnaires

may be user-friendly. Participants can complete surveys to suit their schedule. With

online surveys, participants answer questions in the order presented by the researcher

(Baltar & Brunet, 2012). There are also drawbacks to online surveys. One drawback may

be that participants may view email as spam and not open it (Baltar & Brunet, 2012).

Non-response rates could be significant (Baltar & Brunet, 2012). Surveys delivery is

impersonal (Baltar & Brunet, 2012). It is unclear who completes the survey (Baltar &

Brunet, 2012). A final drawback is that there may be privacy-related issues (Baltar &

Brunet, 2012). Given these drawbacks and the proven advantages of personally

delivering the surveys to members of a professional society was the data collection

technique chosen.

Data Analysis

The data analysis technique for this study was exploratory factor analysis. Before

beginning any analysis, it was appropriate to identify missing data. Mitigation required

removing individual responses with missing data. Before beginning the EFA technique, it

was appropriate to examine descriptive statistics on the Likert data. These descriptive

statistics included frequency tables, mean, median, standard deviation, the coefficient of

variation, and Cronbach’s alpha. The first step of the EFA technique was to run factor

analysis descriptive statistics including correlation matrix, Bartlett’s test of sphericity,

Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy, and possibly anti-

correlation matrix. The second step was to run the initial extraction. Principal axis factor

was the most appropriate extraction technique to discover latent variables and

relationships between items to achieve the purpose of this study. The third step was a

combination of Kaiser’s eigenvalues-greater-than-1 and scree plot to determine the

factors to retain. The fourth step was using oblique Promax method of factor rotation.

The fifth step was to interpret the factor structure using EFA factor loadings and

Cronbach’s alpha. The ultimate goal of this study was to develop a composite index to

determine integration of GBC. The final step was to construct the factor scores for the

composite index. The EFA technique reduced the survey questions to the optimal number

and provided eigenvalue scores on the resulting questions. The eigenvalues multiplied

with the corresponding Likert question score and summed generated an overall index.

Below is a detailed description of each of these steps.

Step 0: Likert Data Descriptive Statistics

Before running EFA, I identified and mitigated missing data. Calculation of

descriptive statistics on the Likert data included frequency tables for each item, mean,

median, standard deviation, and coefficient of variation. Cronbach’s alpha tested the

consistency and stability of the results of the ordinal scale data

Missing data. Missing data can complicate analysis (Seaman & White, 2013;

Seaman, White & Copas, & Li, 2012). Understanding what data is missing is required to

know if mitigating techniques are required (Seaman & White, 2013; White et al., 2012).

The simplest mitigating technique is to remove the incomplete data (Seaman & White,

2013). I used the default option in SPSS to remove records with missing data.

Determining missing data in SPSS involved clicking analyze – multiple imputations –

analyze patterns – selecting the variables – selecting the options of a summary of missing

values, patterns of missing values, and variable with the highest frequency of missing

values.

Likert data descriptive statistics. Descriptive statistics recommended for ordinal

Likert-type item responses include frequency tables, mean, and median, (Boone &

Boone, 2013). Academics have long argued about whether measurements for interval

data apply to the ordinal data (Sullivan & Artino, 2013). Sullivan and Artino provide the

example of “what does the average of “never” and “rarely” really mean?” (p. 542).

Similarly, the mean may appear to be the neutral response if responses cluster around the

high and low extremes (Sullivan & Artino, 2013). However, researchers have determined

that parametric tests, such as mean, standard deviation, and coefficient of variation, are

more robust than nonparametric tests (Sullivan & Artino, 2013). Frequencies were

meaningful to determine the number of responses of each type to each question (Boone

& Boone, 2013). The coefficient of variation is a standardized frequency distribution

expressed as a percentage calculated as a ratio of the standard deviation of the mean

(Subramani & Kumarapandiyan, 2013). Mean provided information about the most

frequent responses (Boone & Boone, 2013). Median indicated the average of the range of

numbers (Boone & Boone, 2013).

Cronbach’s alpha. Social scientists widely use and recommend using

Cronbach’s alpha to test the consistency and stability of the results of the ordinal scale

data (Croasmun, & Ostrom, 2011; Gadermann, Guhn, & Zumbo, 2012; Oluwatayo, 2012;

Tavakol & Dennick, 2011). Consistency and stability indicate the reliability of the survey

instrument (Croasmun, & Ostrom, 2011; Gadermann, Guhn, & Zumbo, 2012;

Oluwatayo, 2012; Tavakol & Dennick, 2011). After entering the data into SPSS

software, a reliability analysis of each question produced descriptive reliability statistics

for the items and the scale as well as showing the inter-item correlations. The results of

the item-total statistics indicated if removing any of the questions would lead to higher or

lower Cronbach’s alpha scores. Cronbach’s alpha scores between 0.67 and 0.90

demonstrate acceptable reliability (Oluwatayo, 2012; Tavakol & Dennick, 2011).

I calculated frequency tables, mean, median, standard deviation, the coefficient of

variation, and Cronbach's alpha on the Likert items. Calculating frequencies tables, mean,

and median in SPSS involved clicking analyze – descriptive statistics – frequencies –

statistics – select mean, median, and standard deviation. The coefficient of variance was

standard deviation divided by mean. Calculating Cronbach’s alpha in SPSS involved

clicking analyze – scale – reliability analysis, then transferring the variables to the Items

section, and then selecting Model as “Alpha.”

EFA Step 1: Factor Analysis Descriptive Statistics

In the EFA method, the first step is to analyze the descriptive statics of the data to

determine if the data obtained from administering the survey instrument to a participant

group is suitable for EFA (Beavers et al. 2013; Schmitt, 2011; Williams et al., 2012).

Assumptions about the suitability of the data include correlational values, and linear

relationships (Beavers et al., 2013). In EFA, there are no dependent or independent

variables; therefore, normality is not required for EFA (Beavers et al., 2013). Generating

factor analysis descriptive statistics in SPSS involved selecting analyze – dimension

reduction – factor – selecting the variables – descriptives – selecting the options of

univariate descriptives, initial solution, coefficients, significance levels, determinant,

KMO and Bartlet’s test of sphericity, inverse, reproduced, and anti-image. This

procedure produced results to determine correlational values and linear relationships.

Correlational values. The SPSS correlations matrix from the factor analysis

descriptive statistics produced the correlations matrix to determine correlational values.

Producing and examining a correlation matrix addressed the correlational values

assumption. Correlations indicate linear relationships (Beavers et al., 2012). Correlations

that exceed .30 indicate sufficient commonality to continue the evaluation (Beavers et al.,

2013).

Linear relationship. The SPSS Bartlett’s test of Sphericity from the factor

analysis descriptive statistics produced Bartlett’s test. The assumption testing must show

that linear relationships exist (Beavers et al., 2013). Bartlett’s Test of Sphericity,

available in SPSS software, is a single number that indicates if linear combinations exist.

Data is suitable for EFA if the Bartlett’s Test of Sphericity is significant (p<.05)

(Williams et al., 2012).

Kaiser-Meyer-Olkin Measure (KMO) of Sampling Adequacy and the

AntiCorrelation matrix indicate whether the dataset produced distinct and reliable factors

(Beavers et al., 2013). A KMO value below .5 is unacceptable (Beavers et al., 2013). If

the KMO value is below .5, the anti-correlation matrix may indicate items that are

unsuitable for the EFA. Values in the anti-correlation matrix above .5 indicate the item

does not have a linear relationship and indicates removal of the item (Beavers et. al.,

2013).

EFA Step 2: Initial Factor Extraction

Once determined that the data was suitable for EFA, the second step was the

initial extraction (Beavers et al., 2013; Schmitt, 2011; Williams et al., 2012). Factor

extraction identifies the latent variables and the relationships between the measured

variables. The linear combinations resulting from this first extraction are the factors

(Beavers et al., 2013). With the first extraction, the linear combinations are independent

or uncorrelated, also known as orthogonal (Beavers et al., 2013). For EFA, there are two

primary extraction models, component analysis, and common factor model (Schmitt,

2011). There are some methods available for each of these models (Schmitt, 2011).

Component analysis method, such as principal component analysis (PCA), reduces the

number of variables while retaining as much of the original variance as possible (Conway

& Huffcutt, 2003). Researchers use a common factor method, such as principal axis

factoring (PAF), to understand the latent, or unobserved, variables and the relationships

between the measured items (Conway & Huffcutt, 2003). Principal axis factor (PAF), a

common factor model, was well suited for this purpose (Conway & Huffcutt, 2003;

Schmitt, 2011; Williams & Brown, 2012). Therefore, I used the common factor model

method PAF, which was available in the SPSS software package. Running PAF in SPSS

involved clicking analyze – dimension reduction – factor – selecting the variables –

choosing principal axis factoring – selecting the options Correlations Matrix, Scree Plot

and eigenvalues greater than 1.

EFA Step 3: Factor Retention

Factor extraction yields multiple factors. The third step was multiphase. First was

to determine which of those factors best represent the data and the relationships (Beavers

et al., 2013; Schmitt, 2011; Williams et al., 2012). The second was to determine which

factors were not statistically or theoretically relevant (Beavers et al., 2013; Schmitt,

2011;

Williams et al., 2012). Third was to retain only those factors that best represented the

data (Beavers et al., 2013; Schmitt, 2011; Williams et al., 2012). Retaining the optimal

number of factors was crucial. Retaining too few or too many factors affects the stability

of factor patterns and interpretation (Hayton et al., 2004; Preacher et al., 2013). Again,

there are multiple techniques available to determine the number of factors to retain, such

as the Kaiser’s eigenvalues-greater-than-1, scree plot, and parallel analysis (Ruscio &

Roche, 2012).

Kaiser’s eigenvalues-greater-than-1 (K1), “retains factors with eigenvalues

greater than 1” (Hayton et al., 2004, p. 193). It is the default on statistical software

programs such as SPSS (Ruscio & Roche, 2012). As the name implies, K1 sets the

threshold between large and small eigenvalues at 1 (Ruscio & Roche, 2012). Eigenvalues

greater than 1 are retained and eigenvalues less than one are not retained (Ruscio &

Roche, 2012). Hayton et al. identified three issues with K1. First, it indicates upper and

lower bound factors, but in practice, researchers use it to determine the exact number of

factors to retain. Second, it tends to lead to overestimation of the number of factors.

Third, it is arbitrary (Fabrigar et al., 1999).

The scree test produces a graphical plot of the eigenvalues in descending order

(Ruscio & Roche, 2012). The scree begins at the breakpoint or the point at which there is

an abrupt change from large to small eigenvalues (Ruscio & Roche, 202). Factors that do

not belong to the scree are retained (Hayton et al., 2004). The scree test is subjective,

especially when there is no clear break or multiple breaks; however, Hayton et al. found

the method worked well with strong factors. Likewise, Fabrigar et al. (1999) found that

the scree test worked well when underlying factors were distinct.

Parallel analysis (PA) may be an accurate method for identifying the number of

factors to retain. PA was accurate primarily because it adjusts for sampling error (Hayton

et al., 2004; Ruscio & Roche, 2012). However, social scientists underutilize PA in their

research (Hayton et al., 2004). Hayton et al. speculated that the reason researchers

underutilize PA is because it is not available in the widely used statistical packages.

In practice, many researchers use multiple methods to determine the number of

factors to retain (Hayton et al., 2004). Standard techniques that are available to determine

the number of factors to retain are the Kaiser’s eigenvalues-greater-than-1, scree plot, and

parallel analysis (Ruscio & Roche, 2012). Scree test, which is available in SPSS

software, is one acceptable method of factor retention (Beavers et al., 2013). I used

Kaiser’s eigenvalues-greater-than-1 and confirmed the number of factors to retain with

scree test. The factor extraction method mentioned above produced a scree plot to

provide a visual representation of the data.

EFA Step 4: Factor Rotation

The fourth step was to factor rotation (Beavers et al., 2013; Schmitt, 2011;

Williams et al., 2012). The factors are rotated to achieve a more simple structure and

produce a solution that is more readily interpretable (Beavers et al., 2013; DeCoster,

1998). There are two types of rotational methods: orthogonal and oblique. There are

fundamental differences between these two methods (Browne, 2001). The method used

can significantly affect the correlations between each factor, as well as how items

correlate with multiple factors, or cross-loadings (Browne, 2001).

The orthogonal rotation methods, such as varimax, quartimax, and equimax reveal

uncorrelated factors that are easy to interpret. However, these methods do not identify

correlated factors (Schmitt & Sass, 2011). As such, orthogonal methods may not

represent the underlying data structure (Schmitt & Sass, 2011). Orthogonal tools are

appropriate if the factors are conceptually independent, and the goal is to generate factor

scores (Beavers et al., 2013).

The oblique rotation methods, such as direct oblimin, Promax, Orthoblique, and

Procrustes, account for relationships, or correlations, between factors (Beavers et al.,

2013). Browne (2001) stated that the oblique rotation method is more appropriate in most

“practical situations” (p. 114) because correlated factors more accurately represent reality

and produces a simpler factor pattern.

Browne (2001) determined that without standardization, both oblique rotation and

orthogonal rotation methods could reproduce their model’s simple structure reasonably

well. When factors were uncorrelated, orthogonal and oblique rotation both resulted in

factor correlations of about zero and similar factor loadings (Floyd & Widaman, 1995 as

cited in Conway & Huffcutt, 2003). However, Browne (2001) also found that oblique

rotations yielded perfect cluster solutions with two substantial loadings per factor and

demonstrated that oblique rotation resulted in greater simplicity.

Myers, Ahn, and Jin (2013) determined that the target rotation method performed

well when there was little a priori knowledge of the structural model, and the underlying

structure was complex. In their study of target rotation, when factor loading was not low,

two or three items per factor achieved accurate results. When factor loading was low,

they obtained accurate results with sample sizes of at least 200 and four or more items

per factor.

Oblique methods produce superior results with correlated factors, and oblique and

orthogonal methods lead to the nearly identical factor loading solutions when constructs

are uncorrelated. Most social science studies involve correlated factors (Schmitt 2011).

The oblique target rotation method performed well for new model development. Schmitt

also recommended oblique rotation when developing and testing a new measurement.

Since the purpose of this research was to construct a new measurement instrument to

develop a new model, which involves identifying correlated factors, I used Promax

oblique factor rotation. Additionally, Promax oblique factor rotation it was available in

the SPSS software package. Running factor rotation in SPSS involved continuing from

the factor extraction procedures described above, by clicking on method and choosing

Promax.

EFA Step 5: Interpretation of Factor Structure

EFA is an iterative process that requires the fifth step of interpreting both the

items and the factors (Beavers et al., 2013). Each measure linearly relates to each factor

(DeCoster, 1998). The factor loadings revealed by the SPSS factor rotation output

indicated the strength of the relationships (DeCoster, 1998). Results of EFA studies

should be statistically significant (Beavers et al., 2013). However, researchers must also

use their theoretical knowledge of the data to determine the conceptual relevance of the

results (Beavers et al., 2013). The researcher must find the fine balance between

statistical significance and conceptual relevance.

I used EFA to evaluate patterns, determine constructs, and reduce the number of

questions. The first procedure was to determine the items with the highest factor loadings

from the EFA. The next procedure was to retain the items with the best factor loadings.

Statistically, items with loadings of .70 or higher that simultaneously do not load on

another factor greater than .40 are considered good identifiers of the factor (Garson, 2010

as cited in Beavers et al., 2013). The final procedure was to calculate Cronbach’s alpha

for each factor including the questions retained for each factor. Cronbach’s alpha scores

between 0.67 and 0.90 demonstrate acceptable reliability (Oluwatayo, 2012; Tavakol &

Dennick, 2011). The resulting questions were the optimal number to measure integration

of GBC.

Step 6: Construct Factor Scores for Composite Index

The purpose of this study was to develop an index to evaluate the maturity level

that a company has attained in implementing the four-step GBC framework. This sixth

step involved constructing factor scores that could be summed to create factor weights

(DeCoster, 1998). The above detailed EFA technique removed redundant questions and

assigned eigenvalue scores to each remaining question. The result was the minimal set of

questions that were necessary for the final survey and the weight assigned to each

question. The EFA-generated eigenvalues were weights that were each multiplied by the

corresponding Likert question score and summed to obtain an overall index (DeCoster,

1998). When company leaders complete the final survey, the resulting data will yield

weighted scores, indicating the maturity level that the company has attained in

implementing the four-step GBC framework.

Data Application

Future researchers can use the GBC index as a tool to capture the four-step

framework of implementing GBC. Researchers and practitioners can then use the result

to calculate, quantitatively, the GBC Composite Index score for individual companies.

The GBC Composite Index score indicates the stage individual companies have achieved

in implementing GBC. The mean GBC Composite Index indicates whether individual

companies demonstrate a high, medium or low commitment to the four-step framework.

Section 3 presents a proposed ranking scale.

The four-step GBC framework provided the assumed constructs. The constructs

were the following. First, developing overarching corporate values (VALUE). Second,

implementing the values (IMPLE). Third, analyzing problems and experimenting to

revise the values or local implementations (ANALY). Fourth, learning from the previous

steps and institutionalizing best practices (LEARN) (Wood et al., 2006).

Study Validity

Study validity includes both internal and external validity. For this study, the

validity of the instrument indicated the internal validity of the study. Tests to determine

internal validity include construct, face, content, and statistical validity. In quantitative

research, external validity means that the results generalize to the population. The

following provides additional details of establishing the internal and external validity of

this study.

Internal validity. The internal validity tests include construct, face, content, and

statistical validity. Construct validity is whether the definitions accurately reflect the

theoretical framework (Oluwatayo, 2012). In this study, the theoretical framework was

the GBC theory with the constructs of VALUE, IMPLE, ANALY, and LEARN. The

EFA technique established weighted factors that could load to the constructs, which

validated the theoretical framework, thereby ensuring construct validity. To mitigate time

and expense constraints, five experts with the familiarity of corporate or business

citizenship assessed the face validity of the survey. The experts determined that the

instrument was reasonable, unambiguous, and clear, (Oluwatayo, 2012; Synodinos,

2003). The EFA technique determined content validity. Applying EFA demonstrated that

the measure covered the range of meanings that could have been applied to the constructs

(MacKenzie et al., 2011; Oluwatayo, 2012; Podsakoff et al., 2012). Computing the scree

test to determine the optimal number of factors to retain ensured statistical validity.

Construct validity. Construct validity is whether the definitions accurately reflect

the theoretical framework (Oluwatayo, 2012). A significant threat to construct validity is

not correctly identifying the constructs. Donna J. Wood, the lead researcher who

developed the theory of GBC, agreed that the terms VALUE, IMPLE, ANALY, and

LEARN capture the process of implementing the four-step GBC framework (D. J. Wood,

personal communication, August 14, 2014; Appendix E). The EFA process ensured

construct validity.

Face validity. Face validity is a subjective assessment of whether the instrument

is relevant, reasonable, unambiguous, and clear (Oluwatayo, 2012; Synodinos, 2003). To

minimize time and cost constraints, five SMEs helped to mitigate the threats of

ambiguous or poorly worded questions (Oluwatayo, 2012). The five individuals were

SMEs in the field of CSR, and/or had sufficient familiarity of global corporate or

business citizenship. The SMEs agreed that the 22 Likert-type questions were reasonable,

unambiguous, and clear. Therefore threats to face validity were mitigated (Appendix F).

Content validity. Content validity demonstrates that the measure covers the range

of meanings applicable to the constructs (Oluwatayo, 2012). A frequently used method to

address threats to content validity is factor analysis (MacKenzie et al., 2011; Oluwatayo,

2012; Podsakoff et al., 2012). By using EFA, the data analysis validated the content and

mitigated content validity threats.

Statistical validity. Threats to statistical conclusion validity focus on whether the

interpretation and measuring of the scores derived from the instrument are valid

(Oluwatayo, 2012). For EFA, this relates to selecting the optimum number of factors.

Selecting too few or too many factors can result in significant model errors (Schmitt,

2011). Parallel analysis (PA) and minimum average partial (MAP) methods are the most

accurate methods for validating the number of factors (Schmitt, 2011). When modeling

Likert scale surveys, the distribution may be non-normal (Schmitt, 2011). Because the

PA method randomly generates eigenvalues over multiple iterations, the distribution

becomes inconsequential; therefore, the resulting data are accurate (Schmitt, 2011). To

ensure the statistical validity and mitigate threats, I determined the optimal number of

factors to retain by using SPSS software to compute a scree test.

External validity. Quantitative research conducted on samples should

demonstrate what is happening in the world (Oluwatayo, 2012). The samples should

correspond to the population. Establishing this external validity is, arguably, the most

important component of a study (Oluwatayo, 2012) because if the study is not valid, it

will not generalize to the population. The major threat to establishing external validity is

that the sample is not representative of the population. The population of this study was

business leaders of multinational companies in the United States. The sample frame was

business executives who were members of professional business societies, specifically

the Executive Suite. The assumption was that the members had an understanding of

global business citizenship and were capable of understanding and accurately completing

the questionnaire. In this case, the sample should be representative of the populations,

thereby establishing external validity. Establishing internal validity, statistical

conclusions validity, and external validity of this study should translate to the results

generalizing to the population. Therefore, the results of the study should generalize to the

global business community.

Transition and Summary

Section 2 laid out the research plan I intended to execute. The plan included the

purpose and design, methodology, sampling, and my role as the researcher. No data were

collected at this point in time, and section 2 only provided the research plan. Upon

Walden University IRB approval, I progressed to Section 3. Section 3 involved collecting

and analyzing the data, presenting the findings, indicating how the study applies to

business practice and contributes to social change and recommendations for further

actions and study.

Section 3: Application to Professional Practice and Implications for Change

This section begins with a restatement of the purpose of the study and

presentation of the findings. The study provided applications to professional practice

and implications for social change. This section describes the applications and

implications. The results of the study indicate recommendations for action and

recommendations for further research. This section explains the recommendations. I

provide reflections about my experience of the Doctor of Business Administration

doctoral study process. The study closes with the conclusion.

Introduction

The purpose of this cross-sectional quantitative study was to develop Likert

survey questions emanating from the GBC theory and apply EFA to assign importance

weights to the questions and group them into factors. The EFA process allowed me to

develop a minimal or pointed self-administered rating index to measure the maturity

level a company has attained toward becoming a global business citizen.

Presentation of the Findings

The data analysis technique for this study was exploratory factor analysis.

Participants completed 381 surveys. The first survey question was the qualification

question, “I am familiar with the concept of corporate citizenship, business citizenship,

corporate social responsibility, or the ethical responsibilities of corporations”.

Participants responded “no” to 209 surveys. These surveys were not include in further

analysis. Participants responded “yes” to 172 surveys. These data were included in the

usable results.

I then analyzed the collected survey data using IBM SPSS Statistics 21. The

remaining 22 questions of the 23-question survey were Likert-type questions. All

Likertscale responses were entered as a score of 1 to 5. A score of 1 represented strongly

disagree. A score of 2 represented disagree, 3 was neither disagree nor agree, 4 was

agree. A score of 5 represented strongly agree. This paragraph is an overview of the data

analysis procedure, followed by details of the procedures and results. The initial data

analysis step was to determine and eliminate surveys with missing data. Likert data

descriptive statistics involved analyzing responses to individual questions using

frequency distributions, mean, median, and Cronbach’s alpha. EFA was used to identify

and group interrelated variables to factors. Step 1 of the EFA technique was to examine

the factor analysis descriptive statistics. The correlations matrix, Bartlett’s Test of

Sphericity, and KMO all indicated the dataset would produce distinct and reliable factors.

Step 2 was to perform a principal axis factoring initial extraction. Step 3 was to examine

scree test and Kaiser’s eigenvalues-greater-than-1 to determine factor retention. Step 4

was Promax oblique factor rotation. Step 5 was an interpretation of the factor structure

and Cronbach's alpha to test the reliability of the determined factor structure. An

additional sixth step was to construct factor scores for the GBC composite index. The

final step was to apply the index.

Step 0: Likert Data. Missing data can complicate analysis (Seaman & White, 2013;

Seaman, White & Copas, & Li, 2012). The simplest mitigating technique is to remove

the surveys with incomplete data (Seaman & White, 2013). I used the default option in

SPSS to identify missing data. Of the 172 usable surveys, SPSS identified 153 containing

no missing data (Figure 3). I removed the 19 surveys that contained missing data. The

result was 153 surveys used for in data analyses. Since the minimum EFA sample size as

specified in Table 3 of Section 2 was determined to be 110 observations, then this

condition was fully satisfied with the sample of 153 usable surveys.

Figure 3. Missing Data.

Likert data descriptive statistics. The next step was to run descriptive statistics

on the Likert data to become familiar with the data. Descriptive statistics recommended

for ordinal Likert-type item responses include mean, median, coefficient of variation to

explain frequency, and frequency tables (Boone & Boone, 2013; Subramani &

Kumarapandiyan, 2013). Table 4 displays the descriptive statistics. Mean described the

average for each variable (Boone & Boone, 2013). The median was the measure of

central tendency indicating the value in the middle of the range of items or the most

popular response (Boone & Boone, 2013). A median of four on all items was unexpected.

In retrospect, the median responses may have indicated the positive perspective the

executives had toward their company as being a global business citizen company. The

standard deviation was used to calculate the coefficient of variation. Frequency tables

provided the number of responses to each item for each variable (Boone & Boone, 2013).

Appendix H presents the frequencies tables for each variable Q1 through Q22.

Table 4

Descriptive Statistics: Mean, Median, Standard Deviation, Coefficient of Variation

Variable

n M Mdn SD

Coefficient of

Variation

Q1 153 4.20 4.00 1.002 23.85%

Q2 153 3.47 4.00 1.142 32.89%

Q3 153 4.14 4.00 1.058 25.57%

Q4 153 3.96 4.00 1.032 26.04%

Q5 153 3.95 4.00 0.955 24.16%

Q6 153 3.73 4.00 1.008 27.06%

Q7 153 3.61 4.00 1.028 28.48%

Q8 153 3.86 4.00 1.045 27.06%

Q9 153 3.85 4.00 1.050 27.27%

Q10 153 3.88 4.00 1.000 25.75%

Q11 153 4.04 4.00 1.063 26.32%

Q12 153 3.50 4.00 1.125 32.10%

Q13 153 3.72 4.00 1.067 28.68%

Q14 153 4.16 4.00 0.862 20.70%

Q15 153 3.94 4.00 1.008 25.58%

Q16 153 4.00 4.00 1.130 28.24%

Q17 153 3.65 4.00 1.067 29.25%

Q18 153 3.61 4.00 1.101 30.46%

Q19 153 3.71 4.00 1.043 28.10%

Q20 153 3.75 4.00 1.017 27.14%

Q21 153 3.58 4.00 1.074 30.05%

Q22 153 3.92 4.00 1.118 28.55%

Cronbach’s alpha measures how items relate to each other as a group, or their

intercorrelations (Croasmun, & Ostrom, 2011; Gadermann, Guhn, & Zumbo, 2012;

Oluwatayo, 2012; Sullivan & Artino, 2013; Tavakol & Dennick, 2011). Cronbach’s

alpha will increase as the intercorrelations among test items within groups of questions,

such as those related to a construct, increase. Alpha for a related subset of questions, such

as those in a construct, are regarded as offering a reliable set of questions to measure the

construct when 0.67 < alpha < 0.90 (Oluwatayo, 2012; Tavakol & Dennick, 2011). In this

range, the researcher has increased confidence to pursue EFA analysis to weight each

question within a given construct. The Cronbach's alpha scores of .921 for the VALUE

construct, .916 for the IMPLE construct, .910 for the ANALY construct, and .932 for the

LEARN construct indicated highly acceptable reliability (Table 5).

Table 5

Reliability Statistics: Cronbach’s alpha

Construct Questions n Cronbach’s alpha

VALUE 1, 3, 4, 11, 22 5 .921

IMPLE 6, 9, 12, 13, 14, 17, 21 7 .916

ANALY 2, 5, 7, 18 4 .910

LEARN 8, 10, 5, 16, 19, 20 6 .932

Given the reliability of the data, I continued the EFA analysis. EFA analysis

identified and grouped interrelated variables to factors. EFA was the appropriate

approach to answering the research question.

Research Question. The purpose of applying EFA was to answer the research

question (RQ): how many and what factors characterized the Likert survey questions to

assess a company’s GBC maturity level? The EFA process targeted the above RQ as it

identified how the survey questions, or items, related to the four assumed constructs of

VALUE, IMPLE, ANALY, and LEARN. The following are details of the analysis. EFA

Step 1: Factor Analysis Descriptive Statistics. In the EFA method, the first step was to

calculate and interpret several key factor analysis descriptive statistics. These included

correlations matrix, Bartlett’s Test of Sphericity, and Kaiser-Meyer-Olkin Measure

(KMO) of Sampling Adequacy. These descriptive statistics were a complement to

Cronbach’s alpha in determining the appropriateness of utilizing the results from

administering the survey instrument to the participant group in an EFA analysis (Beavers

et al. 2013; Schmitt, 2011; Williams et al., 2012). The next sections present the various

factor analysis descriptive statistics.

Correlational values. Correlations indicate the common relationship, or

intercorrelation, between any pair of variables (Beavers et al., 2012). Items that are

strongly intercorrelated may represent the same underlying factor. Correlations that

exceed.30 indicate sufficient commonality to continue EFA evaluation (Beavers et al.,

2013). Table 6 shows the correlations matrix from the SPSS factor analysis descriptive

statistics. As seen from Table 6, all correlations equaled or exceeded .30. Correlations

exceeding .30 indicate that they are intercorrelated sufficiently to identify common

factors.

Table 6

Factor Analysis Descriptive Statistics: Correlation Matrix

Correlation Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Q15 Q16 Q17 Q18 Q19 Q20 Q21 Q22

Q1 1.000 .422 .824 .574 .532 .459 .371 .429 .523 .424 .838 .300 .509 .540 .630 .552 .320 .381 .471 .477 .404 .644

Q2 .422 1.000 .529 .519 ,575 .627 .775 .562 .614 .556 .462 .603 .650 .503 .642 .464 .580 .674 .634 .654 .583 .557

Q3 .824 .529 1.000 .650 .560 .492 .455 .499 .570 .526 .802 .351 .565 .552 .649 .622 .346 .424 .549 .547 .387 .755

Q4 .574 .519 .650 1.000 .512 .540 .569 .532 .517 .563 .715 .471 .600 .555 .542 .559 .454 .531 .552 .599 .365 .665

Q5 .532 .575 .560 .512 1.000 .547 .665 .620 .754 .683 .565 .481 .620 .593 .701 .616 .546 .690 .607 .659 .507 .631

Q6 .459 .627 .492 .540 .547 1.000 .670 .588 .564 .549 .532 .575 .705 .552 .586 .566 .588 .675 .588 .573 .536 .540

Q7 .371 .775 .455 .569 .665 .670 1.000 .666 .683 .679 .460 .730 .649 .459 .575 .487 .713 .819 .637 .710 .605 .544

Q8 .429 .562 .499 .532 .620 .588 .666 1.000 .658 .652 .490 .524 .561 .551 .654 .602 .552 .645 .627 .648 .493 .542

Q9 .523 .614 .570 .517 .754 .564 .683 .658 1.000 .579 .571 .611 .591 .602 .713 .605 .681 .724 .603 .642 .608 .611

Q10 .424 .556 .526 .563 .683 .549 .679 .652 .579 1.000 .537 .463 .598 .473 .724 .652 .504 .562 .756 .702 .400 .550

Q11 .838 .462 .802 .715 .565 .532 .460 .490 .571 .537 1.000 .396 .602 .496 .671 .652 .424 .463 .574 .563 .389 .761

Q12 .300 .603 .351 .471 .481 .575 .730 .524 .611 .463 /396 1.000 .601 .369 .456 .409 .840 .742 .545 .545 .734 .463

Q13 .509 .650 .565 .600 .620 .705 .649 .561 .591 .598 .602 .601 1.000 .523 .682 .562 .641 ,647 .660 .577 .584 .614

Q14 .540 .503 .552 .555 .593 .552 .459 .551 .602 .473 .496 .369 .523 1.000 .677 ,547 .399 .531 .448 .558 .452 .493

Q15 .630 .642 .649 .542 .701 .589 .575 .654 .713 .724 .671 .456 .682 .677 1.000 .716 .488 .602 .747 .691 .475 .603

Q16 .552 .464 .622 .559 .616 .566 .487 .602 .605 .652 .652 .409 .562 .547 .716 1.000 .442 .518 .681 .602 .325 .620

Q17 .320 .580 .346 .454 .546 .588 .713 .552 .681 .504 .424 .840 .641 .399 .488 .442 1.000 ,768 .517 .499 .816 .499

Q18 .381 .674 .424 .531 .690 .675 .819 .645 .724 .562 .463 .742 .647 .531 .602 .518 .768 1.000 .607 .623 .645 .593

Q19 .471 .634 .549 .552 .607 .588 .637 .627 .603 .756 .574 .545 .660 .448 .747 .681 .517 .607 1.000 .694 .419 .566

Q20 .477 .654 .547 .599 .659 .573 .710 .648 .642 .702 .563 .545 .577 .558 .691 .602 .499 .623 .694 1.000 .509 .600

Q21 .404 .583 .387 .365 .507 .536 .605 .493 .608 .400 .389 .734 .584 .452 .475 .325 .816 .645 .419 .509 1.000 .435

Q22 .644 .557 .755 .665 .631 .540 .544 .542 .611 .550 .761 .463 .614 .493 .603 .620 .499 .593 .566 .600 .435 1.000

Two variables computing R^2 > 0.90 indicates that one of the two variables must

be eliminated so as to avoid multicollinearity. Haitovsky’s significance test indicates

whether the correlation matrix has the issue of multicollinearity. Multicollinearity does

not exist if the result is not significant. Table 7 shows the Haitovsky’s significance test.

Since no pairs of variables compute R^2 > 0.90, and the Haitovsky’s score is not

significant, I proceeded to the next step of the analysis.

Table 7

Factor Analysis Descriptive Statistics: Haitovsky’s Significance Test

Factor Result

det .000000000108486

K 22

N 484

df 231

H .000000051512820

ɑ .05

p-value 1

H-crit 267.45

sig No

Linear relationship. I calculated two key descriptive statistics to assess the

adequacy of the sample, as shown in Table 8 below. Specifically, the Bartlett’s Test of

Sphericity assessed whether there was redundancy between the variables that could be

summarized with a few number of factors. Formally, the test checks the H0: whether the

variables are orthogonal. Researchers reject the H0 when the p-value < alpha. As shown

Table 8, the chi-square test statistic was significant χ2 (231) = 3301.755, p<.0000. Thus,

the variables were not orthogonal, indicating that I could proceed with EFA.

The Kaiser-Meyer-Olkin Measure (KMO) of Sampling Adequacy has the same

goal as Bartlett’s Test of Sphericity but proceeds differently. The KMO test checks if we

can efficiently factorize the original survey questions. KMO does this by comparing the

values of correlations between variables and those of the partial correlations, that is, it

removed the effect of the remaining variables. When KMO is close to 0, then EFA is not

relevant and should not be applied. If however, KMO is close to 1, then EFA can perform

the factorization efficiently because the variables are highly correlated. As shown in

Table 8, the KMO of 0.939 is close to 1 indicating that the sample data would produce

distinct and reliable factors and was adequate for EFA.

Table 8

Factor Analysis Descriptive Statistics: KMO and Bartlett’s Test of Sphericity

Kaiser-Meyer-Olkin Measure of .939

Sampling Adequacy

Bartlet’s Test of Sphericity Approx. Chi-Square 3301.755 df

231

Sig. .000

EFA Step 2: Initial Extraction. The second step of EFA was the initial extraction

(Beavers et al., 2013; Schmitt, 2011; Williams et al., 2012). I used principal axis

factoring (PAF) to obtain eigenvalues for each item and understand the latent variables

and the relationships between the measured items (Table 9). Table 9 shows the initial

eigenvalues before extraction and the extraction sums of squared loadings after

extraction. The extraction sums of squared loadings that occurred after extraction and

based on the eigenvalues > 1 criterion left three factors. The three factors had eigenvalue

totals over Kaiser’s criterion of 1 or higher. These three eigenvalues combined explained

69.999% of the variance (Table 9). The initial scree plot (Figure 4) showed three factors.

Table 9

EFA Initial Extraction: Initial Eigenvalues and Total Variance Explained

Factor

Initial Eigenvalues Extraction Sums of Squared

Loadings

Total % of

Variance

Cumulative

Total % of

Variance

Cumulative

1 13.104 59.566 59.566 12.806 58.208 58.208

2 2.012 9.146 68.711 1.789 8.132 66.340

3 1.064 4.836 73.547 .805 3.659 69.999

4 .739 3.357 76.905

5 .608 2.765 79.670

6 .555 2.521 82.191

7 .520 2.364 84.555

8 .444 2.018 86.573

9 .386 1.754 88.327

10 .353 1.606 89.933

11 .315 1.434 91.367

12 .298 1.353 92.720

13 .256 1.166 93.886

14 .227 1.033 94.919

15 .207 .942 95.861

16 .194 .882 96.744

17 .180 .818 97.562

18 .149 .678 98.240

19 .126 .573 98.812

20 .095 .432 99.244

21 .088 .400 99.644

22 .078 .356 100.000

Figure 4. Initial scree plot.

EFA Step 3: Factor Retention. The next step involved generating the

communalities to evaluate the appropriateness of the Kaiser criterion. Based on Kaiser’s

rule, I extracted the recommended three factors 13.104, 2.012, and 1.064 (Table 9).

Kaiser’s rule recommends less than 30 variables, a sample size >250, the majority of

communalities >0.7 and average communality >0.6 (Field 2009). This study passably met

Kaiser's criteria with 22 variables, 153 sample size, 13 of the 22 questions with

communalities >0.7 and with the average communality of the 22 questions of 0.73 (Table

10). Communalities ranging between .60 and .80 indicate excellent congruence (Gaskin

& Happell, 2014). Kaiser criterion indicated that retaining the three factors with

eigenvalue values greater than one was appropriate.

Table 10

EFA Initial Extraction: Communalities Table

Item Initial Extraction

Q1 1.000 .846

Q2 1.000 .647

Q3 1.000 .847

Q4 1.000 .620

Q5 1.000 .678

Q6 1.000 .608

Q7 1.000 .810

Q8 1.000 .662

Q9 1.000 .696

Q10 1.000 .784

Q11 1.000 .867

Q12 1.000 .816

Q13 1.000 .668

Q14 1.000 .513

Q15 1.000 .789

Q16 1.000 .697

Q17 1.000 .864

Q18 1.000 .802

Q19 1.000 .744

Q20 1.000 .712

Q21 1.000 .788

Q22 1.000 .724

Figure 4 above shows the scree plot of the eigenvalues in descending order. The

number of factors to retain is the point at which there was an abrupt change from large to

small eigenvalues. The scree plot concurred with Kaiser’s rule in retaining three factors

(Figure 4). I continued with factor rotation using three factors.

EFA Step 4: Factor Rotation. The fourth step was Promax oblique factor rotation using

the k=3 retained factors. I used Promax oblique factor rotation because it performed well

for new model and measurement instrument development (Beavers et al., 2013; Schmitt,

2011; Williams et al., 2012). Table 11 indicated the variables that loaded to each of the

three factors.

Table 11

EFA Factor Rotation: Pattern Matrix

Note: Extraction Method: Principal Component Analysis.

Rotation Method: Promax with Kaiser Normalization.

Rotation converged in 5 iterations.

EFA Step 5: Interpretation of Factor Structure. EFA is an

iterative process that requires the fifth step of interpreting

both the items and the factors (Beavers et al., 2013). From the

Factor Rotation Pattern Matrix (Table 10), I retained ten items

with EFA factor loadings of .70 or higher that simultaneously

did not load on another factor greater than .40. Table 12

shows the retained questions that loaded to each factor. EFA

generated high factor loadings for Factor 1 with questions

Q10, Q15, Q19, and Q20. All of these questions related to the

LEARN construct; thus Factor 1 represented LEARN. EFA

generated high factor loadings for Factor 2 with questions Q12, Q17, and Q21. All of

these questions related to the IMPLE construct; thus Factor 2 represented IMPLEMENT.

EFA generated high factor loadings for Factor 3 with questions Q1, Q3, and Q11. All of

these questions related to the VALUE construct; thus Factor 3 represented VALUE.

Table 12

Component

1 2 3

Q1 -.090 -.018 .980

Q2 .397 .465 .048

Q3 .085 -.046 .887

Q4 .211 .124 .563

Q5 .595 .169 .162

Q6 .325 .425 .158

Q7 .490 .587 -.129

Q8 .691 .185 -.007

Q9 .374 .411 .191

Q10 .946 -.058 -.044

Q11 .053 .004 .895

Q12 -.002 .926 -.048

Q13 .274 .422 .272

Q14 .369 .088 .366

Q15 .704 -.029 .283

Q16 .684 -.152 .325

Q17 -.030 .954 -.015

Q18 .327 .680 -.044

Q19 .825 .030 .030

Q20 .713 .129 .069

Q21 -.213 .947 .129

Q22 .131 .164 .667

EFA Interpretation: Factor Summary

Factor Number Factor Name Questions

1 Learn Q10 Q15 Q19 Q20

2 Implement Q12 Q17 Q21

3 Value Q1 Q3 Q11

Cronbach’s alpha scores calculated for each factor and all ten retained factors was

above .9 indicating excellent reliability. Tables 13 - 15 show Cronbach’s alpha scores for

Factors 1, 2, and 3 of .911, .921, and .932 respectively. Table 16 shows Cronbach’s alpha

of .925 for all ten retained questions.

Table 13

Cronbach’s alpha: Factor 1

Cronbach’s alpha n

.911 4

Table 14

Cronbach’s alpha: Factor 2

Cronbach’s alpha n

.921 3

Table 15

Cronbach’s alpha: Factor 3

Cronbach’s alpha n

.932 3

Table 16

Cronbach’s alpha: 10 Retained Items

Cronbach’s alpha n

.925 10

The overarching RQ was how many and what factors (dependent variables) are

needed to characterize the Likert survey questions (independent variables) to assess a

company’s GBC maturity level? EFA concluded that the three factors of VALUE,

IMPLE, and LEARN are needed to assess a company’s GBC maturity level. EFA

identified ten questions that best represent these three factors. The conclusions to the

subRQs are presented in Table 17. Sub-RQs 1, 2, and 4 were met. EFA indicated

retaining no questions from the ANALY construct because no questions from the

ANALY construct had factor loadings of >0.70. The ANALY questions all involved

analyzing and experimenting with integrating the overarching principles/values with local

customs or norms. Not retaining questions from the ANALY construct aligned with the

factor retention of three factors. Table 17

Sub-RQ Conclusions

Sub-RQ Conclusion

SRQ1:

Does the survey adequately capture the

VALUE construct of the GBC theory?

Yes

SRQ2:

Does the survey adequately capture the

IMPLE construct of the GBC theory?

Yes

SRQ3:

Does the survey adequately capture the

ANALY construct of the GBC theory?

SRQ4:

Does the survey adequately capture the

Yes

LEARN construct of the GBC theory?

Continuing with the analysis of the three emerging factors LEARN, ANALY, and

VALUE, I identified items with the highest factor loadings and examined inter-item

correlations and Cronbach’s alpha to identify the questions that performed well for each

construct. The original survey included four questions intended to capture the ANALY

construct. EFA analysis indicated that the original survey did not adequately capture the

ANALY construct of the GBC theory. The EFA process could also have identified an

alternative explanation that GBC may be adequately measured with the three constructs

of VALUE, ANALY, and LEARN. The result was the selection of 10-questions with the

corresponding factor analysis results as shown in Table 18.

Table 18

Factor Weights for Each Question by Factor

Original

Number

Final

Number

VALUE

(f3)

IMPLE

(f2)

LEARN

(f1)

Q1 Q1 .980

Q11 Q2 .895

Q3 Q3 .887

Q21 Q4 .947

Q17 Q5 .954

Q12 Q6 .926

Q15 Q7 .704

Q19 Q8 .825

Q10 Q9 .946

Q20 Q10 .713

Table 19 displays a simplified GBC Index survey using the ten retained questions

that represented the three emerging factors. The table includes each factor and the

questions that loaded to each factor. The table includes the full questions, the original

question number of each question, and the newly assigned final question number. A

suggestion for further research is to develop different questions within the ANALY

construct. Further research would indicate whether the survey questions developed for

this study were inadequate, or whether the theory of GBC can be adequately captured and

measured with the three constructs of VALUE, IMPLE, and LEARN.

Table 19

Final Survey Questions by Factor

Question Original

Number

Final

Number

VALUE QUESTIONS (f3)

Have a written code of conduct and policies that reflect the

company’s principles/values.

Q1 Q1

Have a written code of conduct and policies that govern their

conduct everywhere they operate around the globe.

Q11 Q2

Have a written code of conduct and policies that reflect a high

degree of ethical standards.

Q3 Q3

IMPLEMENTATION QUESTIONS (f2)

Implement local variations of their principles/values based on

local customs, culture, norms, or national standards.

Q21

Engage local employees and stakeholders in establishing local

variations of company principles/values to meet local customs,

culture, norms, or national standards.

Q17 Q5

Empower local employees to establish local variations of

Q12 Q6

company principles/values to meet local customs, culture, norms,

or national standards.

LEARN QUESTIONS (f1)

Have a formal, systematic process to organize and communicate

organizational performance to facilitate learning within the

organization.

Q15

Have a formally structured knowledge bank, available to

everyone in the company, where employees can enter tacit

knowledge, questions, and lessons learned.

Q19 Q8

Institutionalize lessons learned into policies, practices, and

behaviors.

Q10 Q9

Share important lessons learned and best practices with

stakeholders and other companies outside the company.

Q20 Q10

Step 6: Construct Factor Scores for Composite Index

The purpose of this study was to develop an index to evaluate the maturity level

that a company has attained in implementing the GBC framework. The sixth step

involved constructing factor scores for the three factors VALUE, IMPLEMENT, and

LEARN and summing them to create factor weights (DeCoster, 1998). EFA generated

eigenvalues, or weights, that are each multiplied by the corresponding Likert question

score. The sum of the results from each factor equates to the overall composite index.

Multiplying the overall composite index by 2.279 normalizes the score to a scale of 100

for ease of interpretation by laypeople (Table 18). When a company representative

completes the final survey, the resulting data will yield a single weighted score,

indicating the maturity level that the company has attained in implementing three steps of

the GBC framework of VALUE, IMPLEMENT, and LEARN.

Table 20 shows the GBC Composite Index calculation for one person. An

individual would answer the final ten survey questions (Table 17). The Likert-scale

responses available to the person are as follows: 1 = strongly disagree, 2 = disagree, 3 =

neither disagree nor agree, 4 = agree, 5 = strongly agree. The response to each question

will be recorded in Table 20 and multiplied with the corresponding eigenvalue as shown

in Table 20. The resulting values are subtotaled for each factor. The sum of the subtotals

is then multiplied by 2.279 to normalize to a scale of 100. The final normalized number is

the GBC Composite Index score. The use of a weighted questions formula based on

eigenvalues is appropriate due to one key reason: all questions are Likert questions from

a scale of 1 to 5. Therefore the measurement scale across all the questions is identical and

thus can be added.

Table 20

GBC Composite Index Calculation

Eigenvalues / question /

factor Calculated Scores

Likert

Survey

Question

Responses

f3 f2 f1

VALUE IMPLE LEARN

f3 f2 f1

weighted weighted weighted

score score score

1 5 .980 4.900

2 5 .895 4.475

3 5 .887 4.435

4 5 .947 4.735

5 5 .954 4.770

14.135 15.940

6 5 .926 4.630

7 5 .704 3.52

8 5 .825 4.12

9 5 .946 4.73

10 5 .713 3.56

Subtotal score per factor 13.810

Add Sub f3, Sub f2, Sub f1 43.885

Subtotal x 2.279 = normalized scale of 100 = final GBC Composite Index Score 100.000

Data Application

Future researchers and practitioners can use the result of this study to calculate the

GBC Composite Index score related to the factors of VALUE, IMPLEMENT, and

LEARN for individual companies. The GBC Composite Index score will indicate the

stage individual companies have achieved in implementing these three GBC constructs.

The mean GBC Composite Index will indicate whether individual companies

demonstrate a high, medium or low commitment to becoming a global business citizen.

The maximum composite score is 43.885 achieved by rating every item 5, or highly

agree. Normalized to a scale of 100 this equates to a score of 100.0 (Table 21).

Table 21

GBC Composite Index: Level 5

Likert

Survey

Question #

Responses

Eigenvalues / question /

factor Calculated Scores

f3 f2 f1

VALUE IMPLE LEARN

f3 f2 f1

weighted weighted weighted

score score score

1 5 .980 4.900

2 5 .895 4.475

3 5 .887 4.435

4 5 .947 4.735

5 5 .954 4.770

6 5 .926 4.630

7 5 .704 3.520

8 5 .825 4.125

9 5 .946 4.730

10 5 .713 3.565

Subtotal score per factor 13.810 14.135 15.940

Add Sub f3, Sub f2, Sub f1 43.885

GBC Composite Index Score 100.000

The minimum composite score from scoring 1, or highly disagree, to each

question is 8.777 (Table 22). A minimum score indicates a company has not started

becoming a global business citizen.

Table 22

GBC Composite Index: Level 1

Likert

Survey

Question #

Responses

Eigenvalues /

question /

factor Calculated Scores

f3 f2 f1

VALUE IMPLE LEARN

f3 f2 f1

weighted weighted weighted

score score score

1 1 .980 .980

2 1 .895 .895

3 1 .887 .887

4 1 .947 .947

5 1 .954 .954

6 1 .926 .926

7 1 .704 .704

8 1 .825 .825

9 1 .946 .946

10 1 .713 .713

Subtotal score per factor 2.762 2.827 3.188

Add Sub f3, Sub f2, Sub f1 = Score 8.777

GBC Composite Index Score 20.000

Scoring 1 on every item yields a composite score of 8.777 and a normalized score

of 20.0. Scoring 2 on every item yields a composite score of 17.554 and a normalized

score of 40.0. Scoring 3 on every item yields a composite score of 26.331 and a

normalized score of 60.0. Scoring 4 on every item yields a composite score of 35.108 and

a normalized score of 80.0. Scoring 5 on every item yields a composite score of 43.885

and a normalized score of 100.0.

Table 23 displays a proposed ranking scale. Future researchers should confirm

this ranking scale. A normalized score between the lowest possible score of 20.0 and 39.0

was assigned Level 1 indicating that a company does not qualify as a global business

citizen. A score between 39.0 and 58.0 was assigned Level 2 indicating that a company

may have started implementing few aspects of GBC but is immature. A score between

58.0 and 77.0 was assigned Level 3 indicating that a company has implemented limited

aspects of GBC and had novice experience. A score between 77.0 and 96.0 was assigned

Level 4 indicating that a company has successfully implemented several aspects of GBC

and is at an intermediate level. A score between 96.0 and the highest possible score of

100.0 indicates a company has implemented most or all of the VALUE, IMPLEMENT,

and LEARN components and was assigned as a Level 5 advanced GBC.

Table 23

GBC Maturity Level Ranking Scale

GBC Composite Index Score Range Level GBC Maturity Level

(20.0, 39.0] Level 1 Not applicable

(39.0 , 58.0] Level 2 Immature

(58.0 , 77.0] Level 3 Novice

(77.0 , 96.0] Level 4 Intermediate

(96.0 , 100.0] Level 5 Advanced

Applications to Professional Practice

Business leaders that can demonstrate to their stakeholders that they are good

business citizens may gain the benefits of improved legitimacy (Wolf, 2014), improved

cost of capital (Brooks & Pavelin, 2014), and improved profitability (Flammer, 2015).

Stakeholders are demanding that corporate leaders conduct their business as though they

are socially responsible citizens of society (Park & Ghauri, 2015). The general business

problem was that there was no self-administered rating system available for business

leaders to report to stakeholders the steps they have achieved toward becoming an

ethically responsible business citizen (Milne & Gray, 2013). The goal of this research

was to operationalize the four high-level GBC theory steps elaborated by Wood et al.

(2006), into a useful survey instrument and weighted index.

The results of the study measured three of the four steps of the GBC

implementation process. The retained survey questions, or items, related to the three

factors of VALUE, IMPLEMENT, and LEARN. The third step of the four-step GBC

implementation process involves analyzing problem areas and experimenting with

solutions to remediate conflicts between the overarching principles/values with local

customs or norms, the ANALY construct. EFA indicated that participants did not rate any

ANALY questions with enough significance to be included in the final survey.

The retained survey questions indicated that the first factor is to create a written

code of conduct that reflects a high degree of ethical standards and the principles/values

of the company. The Code governs the conduct of the company’s employees everywhere

they operate around the globe. The second factor is to implement the code at the local

level. Local employees and stakeholders are empowered to establish local variations of

the code to meet local customs, culture, norms, and national standards. The third factor is

to learn from the previous steps. Learning means organizing and communicating how the

company performs on GBC. Learning means the company institutionalizes lessons

learned into policies, practices, and behaviors and maintains a knowledge bank that is

available to everyone within the company. A mature global business citizen shares

important lessons learned and best practices with stakeholders and others outside the

company.

The artifacts resulting from this study will allow academics to conduct further

research toward the development of a quantitatively validated survey that business

leaders can self-administer. The results of this survey provided a suggested rating system

in the form of a composite index indicating the stage of creating values, implementing,

and learning that business leaders have achieved toward becoming a global business

citizen. Within the context of these three constructs, the results of this study provided a

suggested scale of 5 levels of maturity. A GBC Composite Index score between 20.0 and

39.9 equates to Level 1 indicating that a company does not qualify as a global business

citizen. A GBC Composite Index score between 39.0 and 58.0 equates to Level 2

indicating that a company may have started implementing few aspects of GBC but is

immature. A GBC Composite Index score between 58.0 and 77.0 equates to Level 3

indicating that a company has implemented limited aspects of GBC and had novice

experience. A GBC Composite Index score between 77.0 and 96.0 equates to Level 4

indicating that a company has successfully implemented several aspects of GBC

and is at an intermediate level. A GBC Composite Index score above 96.0 indicates that a

company has implemented most or all of the VALUE, IMPLEMENT, and LEARN

components and is a Level 5 advanced GBC.

The intent of this study was to provide business leaders with a tool to show

stakeholders how they rank as a global business citizen. The result was a study that

researchers can use for further research related to the ANALY construct. The question

was whether the initial survey inadequately captured the ANALY construct, or whether

the three constructs of VALUE, IMPLE, and LEARN adequately describe and measure

the theory of GBC. Follow-up research is required before final development of a

quantitatively validated tool to assess the step or steps corporate leaders have achieved in

implementing the four steps of becoming a global business citizen.

Implications for Social Change

The value of this GBC calculated index is that it provides a practical survey-based

assessment to evaluate the steps that business leaders have achieved toward transforming

their company into a global business citizen. The composite index scale provides a

number that ranks the relative adherence to the GBC steps, rather than a qualitative yes or

no answer. This ranking allows corporate leaders, stakeholders, and academics to

evaluate the progress over time of a company by utilizing the same survey and looking

for improvement in specific areas as captured by the various survey questions related to

the constructs. Business leaders will have this quantitative tool to help communicate to

their stakeholders the steps they have achieved toward becoming a global business

citizen. GBC is an indicator of the level to which companies are maximizing shareholder

value and gaining a competitive advantage at the same time that they are incorporating

laws, public policies, political issues, and the interests of stakeholders. It also indicates

that they are acting ethically and responsibly for the benefit of individual managers,

corporations, industries, and society as a whole.

Recommendations for Action

The concept of global business, or corporate, citizenship has rapidly gained

popularity in the corporate, academic, and political arenas (Crittenden et al., 2011).

Business citizenship has emerged as the preeminent term to describe the ethically

responsible roles of corporations within society (Crittenden et al., 2011). Major

companies, such as Boeing, Dow, IBM, and Microsoft all claim that they are business

citizens (Crittenden et al., 2011). Of the top 100 U.S. companies, 97% claim to be

business citizens (Fifka, 2013). Publically publishing a GBC Composite Index score

would demonstrate the maturity level corporate leaders have achieved toward becoming a

global business citizen. Further research is required to validate whether this study

inadequately captured the analysis step of the GBC four-step implementation process, or

the three factors of VALUE, IMPLEMENT, and LEARN adequately describes and

measures the theory of GBC.

I intend to conduct a follow-up study to determine whether the survey questions

related to the ANALY construct were insufficient, or whether the ANALY step of the

GBC implementation process is not required. I intend to publish the results of this study

in a peer-reviewed journal such as Corporate Social Responsibility and Environment

Management, Global Business Review, Journal of Business Ethics, or Journal of

Management. After the follow-up study, I will distribute the findings of this study and the

subsequent study to professional societies for business executives. I have personal

relationships with, and will distribute information to the following professional societies;

Dallas Business Club, Executives Club of Chicago, Executive Suite, Global Business

Development Center, Leadership Think Tank, Los Angeles Area Chamber of Commerce,

Los Angeles World Affairs Council, and Town Hall Los Angeles. Finally, Donna J.

Wood, the lead researcher responsible for the GBC theory, personally asked me to send

her a copy this study upon completion. She is retired, but I will offer to communicate the

results of this study with her contacts.

Recommendations for Further Research

Recommendations for further research include conducting follow-up research

around the ANALY construct, conducting a CFA study, confirming the proposed ranking

scale, replicating the study in countries outside the United States, and confirming

criterion-related validity. The results of this survey indicated that the sub-research

question is asking whether the survey adequately captured the ANALY construct of the

GBC theory was not met. The first suggestion for further research is to develop different

questions within the ANALY construct. Further research would indicate whether the

survey questions developed for this study were inadequate, or whether the theory of GBC

can be adequately captured and measured with the three constructs of VALUE, IMPLE,

and LEARN.

Researchers use CFA to test the significance of factor loading, to test relationships

between factor loadings, and to test for correlation or lack of correlation of factors

(DeCoster, 1998). The second recommendation for further study is to conduct a CFA

study of the simplified GBC Index survey. The simplified survey does not represent the

ANALY construct. A CFA study would validate the results of this EFA study for the

three constructs of VALUE, IMPLE, and LEARN.

Table 23 presents a suggested ranking scale for the GBC maturity level. The

ranking scale may not accurately reflect the maturity levels of not applicable, immature,

novice, intermediate, and advanced. A final recommendation for further research is to

apply the survey to numerous business leaders to confirm the proposed ranking scale.

The sample frame was a subjective sample due to geography, demographics, and

economic conditions. Consequently, the results of this study may not generalize to

businesses headquartered in countries outside the United States. An area for further

research is to replicate this study in other countries to determine whether the results

generalize to other countries.

No existing instruments measure GBC. Because there were no existing

instruments to correlate with, it was beyond the scope of this study to confirm

criterionrelated validity. Researchers use criterion-related validity to demonstrate that the

scores from the new instrument correlate highly with scores from existing instruments

that are already determined to be valid (Oluwatayo, 2012). Confirming criterion-related

validity is a recommended area for further research.

Reflections

Having worked as an executive at multi-national, multi-billion dollar companies

and as an executive and serving on boards of medium to large nonprofit companies

biased my view of the ethical responsibilities of corporate leaders. I was pleased to find

the theory of Global Business Citizenship. The concept of global business citizenship

allows corporate leaders to maximize shareholder value and gain competitive advantage

at the same time that they are integrating responsible and ethical business policies and

actions, incorporating laws, public policies, political issues, and the interests of

stakeholders. It also indicates that they are acting ethically and responsibly for the benefit

of individual managers, corporations, industries, and society as a whole (Wood et al.,

2006). I started the study firmly believing the EFA results would align with the four-step

process of implementing GBC. I was surprised that participants gave low scores to the

questions related to the third step of analyzing problem areas and experimenting with

solutions to remediate conflicts between the overarching principles/values with local

customs or norms, the ANALY construct. Participants did not rate ANALY questions

with enough significance to generate an ANALY factor or load to any of the three latent

factors. Either the survey questions related to the ANALY construct were inadequate, or

participants did not feel that the ANALY step was necessary. If the latter is the case, the

GBC theory may need revising. Further research is warranted.

Conclusion

Business leaders claim that their companies are business citizens to achieve the

possibly of long-term economic, social, and environmental benefits that may come from

being a global business citizen (Campbell, Eden & Miller, 2012; Menck & Oliveira,

2014; Wood et al., 2006). Business leaders are also making the claim because

stakeholders are demanding that they conduct their business as socially responsible

citizens of society (Park & Ghauri, 2015; Öberseder, Schlegelmilch, & Gruber, 2011;

Shum & Yam, 2011). This study provided the foundation for the first quantitative rating

system that business leaders could self-administer to measure global business citizenship.

After follow-up studies related to the analyzing step of the GBC implementation process,

the GBC Composite Index will be a tool that corporate leaders can use to demonstrate to

their stakeholders the stage of maturity they have attained in becoming a global business

citizen.

A Composite Index to Measure Integration of Global Business Citizenship

Students also viewed