Value marketing through corporate reputation: An empirical investigation of Thai hospitals

Z PaginationTFUKRJMMRJMM27(3 4)FinalsRJMM A 545676 dvi Journal of Marketing Management Vol 27, Nos 3–4, March 2011, 243–268 Value marketing through corporate reputation An empirical investigat.

Vol 27, Nos 3–4, March 2011, 243–268

Value marketing through corporate reputation: An empirical investigation of Thai hospitals

Nopporn Srivoravilai, Dhurakij Pundit University, Thailand

T.C Melewar, Brunel University, UK

Martin J Liu, University of Nottingham, China

Natalia Yannopoulou, Nottingham University Business School, UK

Abstract This study examines the value proposition through corporate reputation,

as corporate reputation best communicates to consumers and the public thecompany’s value orientation The research setting for the study was the Thaiprivate hospital industry A quantitative methodology was employed, usingConfirmatory Factor Analysis (CFA) to purify our measurement items and thenusing regression analysis to test our hypothesis A model of corporate reputation

is proposed, based primarily on a combination of institutional theory, impressionmanagement theory, and signalling theory The study contrasts with mostprevious studies on this subject, which employ a single approach in researchingcorporate reputation Lastly, it explores the implications of corporate reputationwith regards to value offering for practitioners and policy makers

Keywords value marketing; corporate reputation; health industry; institutionaltheory; impression management; Thailand

Introduction

Current changes and challenges in the marketplace call upon marketing management

to create better product and service offerings, while convincing customers about theircorporate and product credibility According to Davidow and Malone (1992), thisshift is considered to be a move towards value marketing; a process that begins byguaranteeing customer satisfaction and, through continuous monitoring, ensures thatvalue is delivered (Walters & Lancaster, 1999) More specifically, and in the case

of services, service value has been widely described as a cognitive trade-off betweenperceptions of quality and sacrifice (Dodds, Monroe, & Grewal, 1991)

Research to date has mainly concentrated on either the value of buyer–sellerrelationships (Lindgreen & Wynstra, 2005), or identifying the value criteria, drivers,and key success factors in consumer decision making, in an effort to make thecase for the importance of value-based marketing to customers (Grant, 1995;

ISSN 0267-257X print/ISSN 1472-1376 online

© 2011 Westburn Publishers Ltd.

DOI: 10.1080/0267257X.2011.545676

McTaggart, Kontes, & Mankins, 1994; Rappaport, 1986; Reimann, 1987; Webster,1994; Wikstrom, 1996) Nevertheless, it has neglected to address how value iscommunicated to customers Thus, in this study, we attempt to examine valuemarketing through corporate reputation This is because corporate reputation bestcommunicates to consumers and the public the company’s commitment to valueoffering (Fombrun, 1996).

For the first time, therefore, a model of corporate reputation is examinedbased primarily on the combination of the most influential theories with regards

to reputation – namely, institutional theory, impression management theory, andsignalling theory – in contrast to most existing studies, which employ a singleapproach in researching corporate reputation Hence, the proposed conceptualmodel of this study aims to capture the effects of institutional factors on corporatereputation and to examine the potential outcome of the reputation-building processand value creation

The private hospital industry was chosen as the research setting because it has astrong institutional environment (Arndt & Bigelow, 2000) that makes the departurefrom an institutionalised structure unsafe Furthermore, private hospitals are servicecompanies whose products are largely, though not exclusively, intangible, and thushave to rely heavily on corporate reputation to attract customers (Fombrun, 1996) Inaddition, the Thai private hospital industry was selected because its national culture

is considered different from that of Western countries, where most previous studiesabout corporate reputation have taken place As a result, our study gives us theopportunity to test empirically existing models in a non-Western setting, and thushelps us to examine the applicability of theories in other contexts

In the following section, we discuss the key characteristics of health careservices and the distinctiveness of the Thai private hospitals, contrasting them withWestern service designs We then examine the relevant literature with regards tothe antecedent of corporate reputation, organisational legitimacy, and impressionmanagement tactics, and explore its outcome Relevant hypotheses are presentedafter the discussion of each component of the framework These are first tested usingCFA, in order to access and purify the measurement items of each component, andthen subjected to regression analysis Lastly, we present the study’s implications fortheoreticians and practitioners

Health care and our selected research context

Even though the value concept has long been considered as fundamental for allmarketing activity (Holbrook, 1994), examining marketing towards value creationhas recently been proposed as the contemporary focus for marketing (Sheth &Uslay, 2007) An emerging paradigm shift towards a Service-Dominant (S-D) Logic

as the foundation for marketing theory and practice also seems to have attractedconsiderable attention within the marketing literature (Vargo & Lusch, 2004) It hasbeen argued that all marketing exchange incorporates an actual service, which results

in intangible outcomes (Lusch, Vargo, & O’Brien, 2007) Furthermore, according toHolbrook (2006), all products are services, since products perform services that willprovide the relevant value-creating experiences

Health care services, however, present certain peculiarities compared to productofferings and other service industries They are mainly viewed as high-involvement

services (Hogg, Laing, & Newholm, 2004), with a high degree of emotional

vulner-ability and risk (Jadad, 1998) They are also characterised as asymmetrical in

information and power distribution, taking a domineering approach to alternatives

(Laing, Hogg, & Winkelman, 2005) As a result, they are seen as knowledge-based

services (Mills & Moshavi, 1999) and an archetypal professional service (Wilson,

1994)

Despite recent changes within the health care sector towards a more

consumer-oriented health system, as well as the emergence of post-modernist consumer

cultures (Laing, Lewis, Fowall, & Hogg, 2002), consumers are still largely considered

as passive (Ham & Alberti, 2002) and dependent on professionals’ ‘rational

scientific rhetoric of reliability’ (Laing, Newholm, & Hogg, 2005, p 514) Perhaps

more surprisingly, recent findings based on cases in the West also suggest that

consumers remain somewhat reluctant about expressing consumerist expectations

and behaviours in respect to health care services (Laing & Hogg, 2002)

It has therefore been noted that, although Western service designs have long

been established, they continue to face fundamental challenges due to organisational

complexities and conflict, which has been intensified by the various key actors

involved (Smith & Fischbacher, 2002) Government attempts to move in a more

market-driven direction have largely been unsuccessful in the UK (Laing & Cotton,

1996; Laing & Hogg, 2002), whilst recent US attempts to create a more regulated

heath service has encountered considerable opposition and obstacles

Hence, while the West seems currently to be debating the merits of the

consumerisation of health care, our paper examines value marketing within health

care in a context that differs in the following main ways First, Thai private hospitals

present a case of applied commercially driven marketing concepts Second, their

focus is on competing in the international health care sphere Third, within Thai

private hospitals, producer dominance has been significantly reduced Fourth, the

domineering approach to alternatives is not as high as in other cases

A final interesting characteristic of Thai private hospitals is that they position

themselves within the medical tourism sector and aspire to gain a strong international

presence within this market sector Their main target audience comprises Thai

nationals and foreigners in search of affordable medical treatment Currently, Thai

nationals account for 60% of their customer base and foreigners from 136 countries,

mainly from the Middle East and Southern and Eastern African countries, for

40% Moreover, it has been projected that foreign patients’ contribution to the

revenue of Thai private hospitals will reach 60% over the next five years (IMTJ,

2009)

In sum, our research examines value marketing within health care by studying

private hospitals, in a collectivistic society, that operate with applied marketing

concepts and whose main target audience is patients from other counties

The relationship between organisational legitimacy and corporate

reputation

Corporate reputation has been defined in a number of different ways (Bick, Jacobson,

& Abratt, 2003) based on the different research areas that theoreticians use for its

examination, as shown in Table 1

Nevertheless, studies on corporate reputation have emphasised two traditional

assumptions about rationality and objectivity They have paid less attention,

Table 1 Definitions of corporate reputation.

products’ quality

Shapiro (1982,1983)

attributes or achievement based on what therelevant public knows about the actor

management

Public’s cumulative judgements of firms overtime

Roberts andDowling (2002)Strategic

management

Stakeholder’s knowledge and emotional reactions(e.g affect, esteem) towards a firm

Hall (1992);Fombrun (1996)

attribute of an entity

Herbig andMilewicz (1995)

(1999)

and evolves over time as a result of consistentperformance, reinforced by effective

communication

Gray and Balmer(1998)

(Adapted from Bennett and Kottasz, 2000, p 224).

however, to relevant social assumptions (Fombrun & Van Riel, 1997) According

to institutional theory, it can be argued that corporate reputation is an outcome ofinstitutional processes by which a company comes to accept, correspond to, andtransmit the social norms that are taken for granted as defining the way thingsare or should be done Such processes are multiple, but past investigations mainlyaddress the normative process (Rao, 1994; Staw & Epstein, 2000) Besides, corporatereputation, especially in the case of Thai private hospitals that apply commerciallydriven marketing concepts, can be viewed as a resource, based on the comparativeadvantage theory (Hunt & Morgan, 1995) In this view, hospitals’ market orientationthrough corporate reputation could represent an intangible entity that enables them

to produce efficiently market offerings that offer value to the consumers of themedical tourism sector Moreover, a company is more likely to attract supportand resources from its constituents if it appears to be proper, understandable, anddesirable than if it does not seem so (Staw & Epstein, 2000)

Organisational legitimacy has been conceptualised with different levels ofspecificity and emphasis (Pfeffer & Salancik, 2003; Suchman, 1995) Organisationallegitimacy expresses the importance of the congruence between the social valuesassociated with the activities of organisations and the external (societal) norms orexpectations (Ashforth & Gibbs, 1990; DiMaggio & Powell, 1983; Meyer & Rowan,1977; Suchman, 1995) Furthermore, legitimacy can also be viewed as a type ofresource that is essential for the acquisition of other resources such as high-qualityemployees, financial resources, and governmental support (Ashforth & Gibbs, 1990;Zimmerman & Zeitz, 2002)

Hence, the relationship between corporate reputation and organisationallegitimacy is not a simple one On the one hand, corporate reputation and

organisational legitimacy share several characteristics Both of them, for example,

are socially constructed and subject to evaluations by internal as well as external

constituents (Ashforth & Gibbs, 1990; Deephouse & Carter, 2005; Fombrun

& Shanley, 1990) Moreover, both connect to similar antecedents (e.g financial

performance) and consequences (e.g social support) (Deephouse, 1997; Deephouse

& Carter, 2005; Fombrun & Shanley, 1990) On the other hand, they are different

from each other to the extent that a deterioration in corporate reputation has less

dreadful consequences than that of a deterioration in legitimacy

In linking organisational legitimacy with corporate reputation, this research

concentrates on the evaluative aspect (Suchman, 1995) of organisational legitimacy,

consisting of sociopolitical and pragmatic legitimacy Sociopolitical legitimacy is

composed of regulative and normative legitimacy, both of which are generally

referred to as the result of the legitimation process whereby a company conforms to

the rules, standards, and norms set by governments, professional bodies, and society

as a whole (Aldrich & Fiol, 1994) Pragmatic legitimacy is usually referred to as the

outcome of the legitimation process by which a company responds to the expectation

of its constituents at both individual and wider society levels (Suchman, 1995)

Acquiring both types of legitimacy can help a company to influence stakeholders’

assessments of its corporate reputation positively On the one hand, the company

that obtains regulative legitimacy signals to its stakeholders that it possesses sufficient

transparency, defined as the timely disclosure of adequate information regarding the

company’s operating and financial condition and its corporate governance practices

(Fombrun & Van Riel, 2004) By being transparent, the company signals to the

public that it is comprehensible, relevant, and ready to be further evaluated by its

constituents, referring us to the value of relationships

In addition to the effects of sociopolitical legitimacy, achieving pragmatic

legitimacy can lead stakeholders to perceive the company as authentic, especially

with regards to its claim for value creation for its customers (Tzokas & Saren,

1999) A company is considered to possess common characteristics of both pragmatic

legitimacy and authenticity when it shows honesty, trustworthiness, and wisdom

(Fombrun & Van Riel, 2004), all of which are similar to the characteristics of a

person In other words, the personification of companies arising from institutional

pressure suggests that a stakeholder will consistently look for the appropriate

qualifications of a company as if it were a person and will evaluate its activities

accordingly Alternatively, it can be stated that the more honest and trustworthy a

person believes a company to be, the more favourable reputation that person will

ascribe to the company

H1: Overall organisational legitimacy is positively associated with corporate reputation.

H2: Sociopolitical legitimacy is positively associated with corporate reputation.

H3: Pragmatic legitimacy is positively associated with corporate reputation.

Corporate reputation and impression-management tactics

Impression-management tactics refer to the methods employed by organisations

or people to influence or manipulate the impression formed of them by their

stakeholders Corporate reputation is potentially a consequence of both verbal and

non-verbal impression-management activities Verbal impression management refers

to the tactics that are directly involved with verbal communication (e.g excuses,explanation), whereas non-verbal impression management refers to the tactics thatare mainly involved with other non-verbal actions (e.g providing a free snack anddrink, offering a free newspaper).

Impression-management techniques can affect corporate reputation in at least twoways First, impression management tactics can enhance a company’s reputation byincreasing the company’s positive visibility and distinctiveness (Fombrun, 1996)

In fact, impression management tactics can influence corporate reputation byimproving the following characteristics of the company: visibility, distinctiveness,transparency, consistency, and authenticity (Fombrun & Van Riel, 2004) However,since impression-management theory is principally concerned with the management

of appearance or impression in the mind of constituents, the visibility anddistinctiveness of a company will be the dimensions that are most effectively anddirectly manipulated

A company can thus enhance its reputation by actively employing tactics such

as self-promotion and self-presentation to boost the aforementioned characteristics.Companies that engage in impression-management activities can also enhance theirreputation through the identification of stakeholders with the companies Staw andEpstein (2000) found that the corporate reputation of US industrial corporations inthe Fortune 500 database is positively and statistically associated with the adoption ofpopular management techniques for value creation, such as total quality management(TQM) and empowerment The latest refers us to S-D Logic, which points outthat value is perceived, defined, and co-created by consumers (Vargo & Lusch,

2004, 2008) Lamertz, Heugens, and Calmet (2005) have shown, in addition, thatCanadian brewing companies can represent themselves positively and distinctively

in their external environment through self-categorisation, a tactic that refers to theassignment of companies to groups of similar characteristics, such as age and size(Chatman & Spataro, 2005) Taking into account these two reasons mentionedabove, it is suggested that:

H4: Overall impression-management tactics are positively associated with corporate reputation.

H5: Verbal impression-management tactics are positively associated with corporate reputation.

H6: Non-verbal impression-management tactics are positively associated with corporate reputation.

Other than the direct effect of impression-management tactics on corporatereputation, an indirect effect may also exist In particular, the relationship betweenimpression management and corporate reputation can be mediated by a thirdvariable: organisational legitimacy Overall organisational legitimacy is assumed to

be a generative mechanism (Baron & Kenny, 1986) through which impressionmanagement tactics influence corporate reputation According to institutionaltheorists, a company can gain organisational legitimacy not only via isomorphicactions (e.g conformity with acceptable norms, rules, standards) but also via thedecoupling – that is the adoption of practices that cohere with institutional demandsbut are intentionally different from how work is actually done Decoupling, as

a form of impression management, can help a company to defend successfully

its social suitability from outside threats Ashforth and Gibbs (1990) argue that

organisations can employ tactics such as rejection and concealment, apology offering,

and explanation offering to make themselves appear consistent with social values and

expectations

H7: The relationship between overall impression-management tactics and corporate

reputation is mediated by organisational legitimacy.

Customer support: A consequence of corporate reputation

Empirical research reveals that corporate reputation is positively correlated with

organisational as well as marketing variables, such as financial performance indicators

(Carmeli & Tishler, 2004; Deephouse, 2000), perceived credibility (Ganesan, 1994),

perceived quality (Wheatley & Chiu, 1977), and purchase intention (Yoon, Guffey,

& Kijewski, 1993) All of these variables appear to be essential for customer

retention, which has been seen as the main objective of value marketing (Bolton

& Drew, 1991) It has been shown that the concept of customer value has a strong

relationship to customer satisfaction (Spiteri & Dion, 2004), which in turn derives

from customer support (P Anderson, 1982) Moreover, it has been proposed that

customer support, along with innovation and effective operational processes, forms

the basis for building core capabilities that enhance value marketing (Doyle, 1995)

Conceptually, a company with a favourable reputation is argued to be able

to attract support from employees, customers, investors, and other stakeholders

(Fombrun, 1996) Nevertheless, empirical research has not investigated the

relationship between corporate reputation and support as a construct

Support from customers is of particular interest because it represents social

endorsement from virtually the most important stakeholder of any company, and

has usually been referred to as an outcome of institutional processes (Suchman,

1995; Zucker, 1987) In line with past research (Albrecht & Adelman, 1987; Cobb,

1976; Handelman & Arnold, 1999), this study defines customer support as actions

by customers that reduce uncertainty for a company that lead the company to believe

that it is cared for and loved and that make the company belong to a network of

communication and mutual obligation The social identification concept and the

signalling theory can be used to explain the proposed relationship between corporate

reputation and customer support

Social identification generally refers to the perception of oneness with or

belongingness to a group of persons (Ashforth & Mael, 1989; Dutton, Dukerich,

& Harquail, 1994) Identification with a group can lead to socially and physically

supportive actions by an individual (Ashforth & Mael, 1989; Bhattacharya &

Sen, 2003) Lichtenstein, Drumwright, and Braig (2004) found that customers’

identification with companies was positively related to their purchasing behaviour

More recently, Cornwell and Coote (2005) found a positive relationship between the

willingness to purchase, sponsoring a firm’s products, and consumers’ identification

with a not-for-profit organisation to which the sponsoring firm provides support

In the context of the customer–corporation relationship, when a potential

customer is presented with a choice of companies from which he can buy some

services, all things being equal, the customer tends to identify him/herself with

a more reputable company in order to enhance his/her self-esteem He/she then

and returns his/her support in various forms (e.g word-of-mouth references, pay

premium prices)

Alternatively, according to signalling theorists (Fombrun & Shanley, 1990; Weigelt

& Camerer, 1988), corporate reputation can be viewed as a cue that reflects acompany’s characteristics, service quality, reliability, and credibility to its customers,especially when there is information asymmetry in the market Consumers use well-known brands in order to reduce risks (Ring, Schriber, & Horton, 1980), as brandingprovides guarantees about quality and security (Aaker, 1991) Hence, a strong brandwith a positive corporate reputation is a safe place for consumers because it enablesthem to visualise and understand better the offer and face up to the uncertainty andperceived risk associated with buying and consuming a product or service In otherwords, strong corporate reputation can be used as a mechanism through which acustomer attempts to minimise his/her risk and bolsters his/hers confidence and trust

(Bearden & Shimp, 1982; Roselius, 1971), which is seen as key for value marketing(Fornell, 1992) Taking together, the explanations based on the social identificationconcept and signalling theory, it is proposed that:

H8: The relationship between corporate reputation and customer support is positive.

Yet the relationship between corporate reputation and customer support is notalways straightforward Corporate reputation itself can be a generative mechanismthrough which organisational legitimacy influences customer support In otherwords, corporate reputation can mediate the relationship between organisationallegitimacy and customer support Based on the reasoning above, this study proposesthat:

H9: The relationship between organisational legitimacy and customer support is mediated by corporate reputation.

H10: The relationship between sociopolitical legitimacy and customer support is mediated by corporate reputation.

H11: The relationship between pragmatic legitimacy and customer support is mediated

by corporate reputation.

The conceptual model (see Figure 1) and its hypothesis are designed to capture theeffects of institutional factors on corporate reputation and to examine a potentialoutcome of the reputation building process

Methods

Main survey

A main questionnaire survey was conducted in Thai private hospitals The targetedparticipants of the main survey were a group of customers and managers of all 346private hospitals, of which a complete directory was obtained from the MedicalRegistration Division, Department of Health Service Support, Ministry of PublicHealth (MPH) Methodologically, all private hospitals were included for two reasons:the requirement for the sample size imposed by the chosen analysis techniques (EFAand CFA) and the historical response rates recorded in past research (Powpaka,

Figure 1 An operational model of the antecedents and consequence of corporate

Impression

Management

Corporate Reputation

Organisational Legitimacy

Customer Support Consequence

1998) Customers answered questions regarding corporate reputation, organisational

legitimacy, impression-management tactics, and customer support, whilst managers

answered questions about demographic characteristics and other control variables of

hospitals in separate sets of questionnaires

Measurements

In this study, the domains of corporate reputation are multidimensional and are

based mainly on the 20-item reputation quotient scale (Fombrun, Gardberg, & Sever,

2000) To measure organisation legitimacy, this research follows Handelman and

Arnold (1999) by identifying two domains of the eight-item organisational legitimacy

scale: sociopolitical and pragmatic legitimacy Based on individual impression

management scales (Bolino & Turnley, 2003; Wayne & Ferris, 1990; Wayne & Liden,

1995), the study classifies the concept into two broad domains: verbal and

non-verbal impression management, which was measured by a 22-item scale adopted from

Bolino and Turnley (2003) Since this study identified one domain for the construct

and focused on general rather than specific forms of support, the customer support

construct was measured by the nine-item scale formulated by Long-Tolbert (2000)

Control variables

Three control variables (marketing capabilities, operating performance, and size)

known to affect corporate reputation were incorporated into this study’s model

Marketing capability (the development of products and services, marketing

innovation, customer relationship management, etc.) is found to be a driver of

corporate reputation (Blazevic & Lievens, 2004; Dowling, 2004) A measurement

scale was adapted from Vorhies and Morgan’s (2003) scale and measured in the

questionnaires to control for the marketing capability effect A number of studies

(Carmeli & Tishler, 2005; Fombrun & Shanley, 1990; Roberts & Dowling, 2002)

have also reported a positive relationship between performance outcomes (both

financial and non-financial) and corporate reputation of a company The existing

literature (Black, Carnes, & Richardson, 2000; Shrum & Wuthnow, 1988) posits a

significant and positive relationship between the size of a company and its reputation

To control for variables such as marketing capabilities and operating performance,

we identified a general dimension for marketing capabilities and measures operatingperformance with few self-reported items (E Anderson, Fornell, & Lehmann, 1994;Homburg & Pflesser, 2000; Noble, Sinha, & Kumar, 2002) Note that these proxiesare in line with several studies in marketing and strategic management whose focalsetting is in the hospital industry (Dranove & Shanley, 1995; King & Zeithaml,2001) However, constraints imposed by many hospitals prohibit us from accessingtheir actual operating performance figures Subjective performance indicators areconsequently employed at the expense of objective indicators (Covin & Slevin, 1989;Dess & Robinson, 1984)

Existing literature (Black et al., 2000; Dunbar & Schwalbach, 2000; Shrum &Wuthnow, 1988) has found a significant and positive relationship between size andreputation of a company In this study, the number of beds (provided by the Ministry

of Public Health) was used as an indicator of size

Data-collection procedure

In line with previous studies (Algesheimer, Dholakia, & Herrmann, 2005; Hartline

& Ferrell, 1996; Hartline, Maxham, & McKee, 2000), this study employed a probability sampling technique and relied on managers to distribute questionnaires

non-to their cusnon-tomers Questionnaire packets, each comprising one manager survey, fivecustomer surveys, postage-paid return envelopes, an instruction page, and a coveringletter (Hartline et al., 2000), were sent to hospitals approximately two weeks afterfirst contact was made Instead of having managers freely choose respondents, theywere instructed to follow a system in which only one customer was surveyed eachday at a fixed cashier point and at a certain time for five consecutive days While thismethod is by no means perfect, it does represent an attempt to reduce potential biasesthat might arise from purposively ignoring some types of respondent The instructedmethod is similar to systematic sampling, in that a system is employed to recruitrespondents (Denscombe, 2002) It differs from systematic sampling, however, inthat the sampling unit cannot be identified in advance and the probability of eachsampling unit (customer) being selected is not known (Baker, 2002)

Data analysis

The analysis of data for this study consists of three major parts: (1) purification ofmeasurement scales using exploratory factor analysis (EFA) and reliability analysis(Cronbach’s α); (2) validation of measurement scales using confirmatory factor

analysis (CFA); and (3) hypotheses testing using multiple regression analysis (MRA).MRA was employed at the expense of structural equation modelling (SEM) becauseour sample size was <100, which did not meet the threshold value Instead, we

calculated summated scores (i.e averaging item scores) to represent dimensions ofall constructs; within-group agreement assumption was also assessed before the dataaggregation (Bliese, 2000; De Jong, de Ruyter, & Lemmink, 2004; James, 1982),which resulted in an aggregated sample size of 90 observations When SEM cannot

be used, MRA is an appropriate and the most widely used method for investigatingthe relationship between a dependent variable and two or more independent variablesdue to its well-developed underlying statistical theory (Blaikie, 2003; Hair, Anderson,Tatham, & Black, 1998; Montgomery, Peck, & Vining, 2001)

Main study

Data profile and screening

We received 90 questionnaires from managers and 416 questionnaires from

customers (see Appendix 4), which resulted in an aggregated response rate of 26%

To assess the factorability of items, we examined anti-image matrices and two

other indicators (Kaiser-Meyer-Olin measure of sampling adequacy and Bartlett’s

test of sphericity) For every EFA, it was found that manifest variables have KMO

measures of sampling adequacy >.50 and p-values for Bartlett’s test of sphericity

<.05, suggesting satisfactory factorability for all items.

Before conducting CFA and hypotheses testing, questionnaires from both sources

were subjected to initial data screening First, they were checked for missing value

patterns Little’s MCAR statistics showed that both data sets were not missing

completely at random (customer set: p-value = 001, χ2= 5,711.792, df = 5,382;

manager set: p-value = 013, χ2= 156.108, df = 119) Extensive tests, however,

revealed that they were instead missing at random (MAR), which therefore suggested

that the expectation maximisation (EM) technique could be used to impute all

missing values Additionally, since none of the cases or variables had more than 30%

missing values, it was considered appropriate to retain them for further analyses

Second, since the response rates were relatively low (26% and 24%), tests for

non-response bias were performed to examine whether the questionnaires received

earlier differed statistically from those received later or those that were not returned

Multivariate analysis of variance (MANOVA) was used to compare the means of all

manifest variables for early and late samples of each group (manager and customer)

The early responses were defined as those questionnaires that were returned to us

within four weeks, whereas the late responses were the questionnaires that were

returned after the fourth week The p-values (.188 for customer and 566 for manager

samples) for overall comparisons showed that there was no significant difference

between variables of early and late responses This indicated that the non-response

bias was not a significant problem in this research

Model fit

To evaluate model fit, five indicators were examined: chi-square statistics (χ2), the

goodness-of-fit index (GFI), the root-mean-square error of approximation (RMSEA),

the normed fit index (NFI), and the comparative fit index (CFI) The CFA results (e.g

path coefficients, model fit indices) are summarised in Table 2

Following CFA and reliability assessment, unidimensionality was evaluated The

remaining items were found to be unidimensional, as each of them loaded on only

one corresponding dimension of well-fitted models (p-values associated with χ2 >

0.05) (Hughes, Price, & Marrs, 1986) Furthermore, purified scales were subjected

to validity tests

First, convergent validity was examined via the inspection of path coefficients and

their statistical significance All coefficients were found to be above a recommended

value (.6) (Bagozzi & Yi, 1988) and statistically significant

Second, discriminant validity was assessed via chi-square difference tests for every

pair of estimated constructs (one pair at a time) This was done by comparing

the chi-square values obtained from an unconstrained and a constrained model, in

Table 2 Final CFA results.

Standardised

Group 1

(0.023)

EmotionalAppeal

PragmaticLegitimacy

Table 2 (Continued).

Standardised

Note: ∗Item numbers are referred to those in the main questionnaire (i.e different from those in EFA).

1) The first variable of each dimension was a reference variable: therefore, t-values could not be

calculated for those variables.

2) CP9 = Personal Selling PI = Growth of Total Revenue P2 = Growth of the Number of Customers,

P3 = Overall Operating Performance

which the correlation between two constructs was set at 0 If the difference was

statistically significant, this would prove the discriminant validity of both constructs

(J Anderson & Gerbing, 1988) In general, all constructs were found to have an

adequate discriminant validity (i.e p-values < 01 for all χ2)

Hypothesis testing

Regression analyses were performed to test hypotheses in Eview 5.0 and SPSS The

estimated regression models generally exhibited were moderate All F-statistics were

statistically significant (p < 001), suggesting that the explanatory powers of overall

models were adequate (Myers, 1990, p 37) Overall organisational legitimacy was

significantly and positively related to corporate reputation (b = 841; p < 001).

Likewise, sociopolitical and pragmatic legitimacy were positively and significantly

associated with corporate reputation (b = 403 and 440 respectively; p < 001).

Hence, the three hypotheses (H1, H2, and H3) are strongly supported when hospital

size, market capability, and performance are controlled

The fourth hypothesis (H4) predicted that overall impression management

would be positively associated with corporate reputation The estimated coefficient

(b = 700; p < 001) confirms this assumption H5 and H6, which posited that

verbal and non-verbal impression management would be positively correlated with

corporate reputation, are also supported, as their estimated coefficients were positive

and statistically significant (b = 384 and 364 respectively; p < 001) Hypothesis 8,

that the relationship between corporate reputation and customer support is positive,

is corroborated by the estimated coefficients of model 5 and 6 According to model 6,

the unstandardised coefficient of corporate reputation is 744 (p < 001) In addition,

the coefficients of emotional appeal and products and services are 443 (p < 001)

and 309 (p < 05) respectively Note that all control variables are not statistically

significant in any models (see Table 3)

To test for mediating effects (H7, H9, H10, and H11), we followed the procedure

devised by Baron and Kenny (1986) First, the mediating variable was regressed on

the independent variable (path A) Second, the dependent variable was regressed