Antecedents and consequences of brand loyalty: An empirical study

These influences, in turn, are reflected in the behavioural consequences of loyalty.31,32 According to this view, one may hold a favourable attitude towards a brand but still not purchas

thus reducing the probability of ing brands.6 Finally, brand loyalty hasbeen identified as a major determinant

switch-of brand equity.7

The concept of brand loyalty hasnot, however, been uniquely definedand operationalised in the marketingliterature For example, brand loyaltyhas been defined as a repeat purchase,8

preference9 and commitment,10 and asretention and allegiance.11 Thesediverse definitions of brand loyalty are

in part due to the fact that loyalty is avery complex construct.12 Further-more, there exist various aspects ofbrand loyalty (such as behavioural andattitudinal brand loyalty) If theseaspects were to be integrated, however,then one could come up with a

INTRODUCTION

There is no doubt, among academicsand practitioners alike, that the con-cept of brand loyalty is of strategicimportance for companies in order

to obtain a sustainable competitiveadvantage This is due to a num-ber of reasons First, brand-loyal con-sumers are less expensive, since theyreduce the marketing costs of doingbusiness.1–3 Secondly, brand extensionsare less risky for brands that exhibithigh loyalty.4Thirdly, brand loyalty hasbeen shown to be associated withhigher rates of return on investmentthrough increases in market share.5

Fourthly, brand-loyal consumers havefewer reasons to engage in an extendedinformation search among alternatives,

Spiros Gounaris

Assistant Professor of Marketing,

Department of Marketing and

Communications,

Athens University of Economics

and Business, Patission 76,

the different types of brand loyalty.After discussing the findings and theirmanagerial implications, this paperconcludes with study limitations anddirections for future research.

TYPES OF LOYALTY

The authors’ review of past literaturesuggests that brand loyalty has beenviewed from three different, albeitcomplementary, perspectives, namely:the behavioural, the attitudinal and thereasoned action perspectives

More specifically, the behaviouralperspective has conceptualised brandloyalty in terms of repeated pur-chases (for example, Cunningham19

and Kahn et al.20) In fact, severalmodels have been proposed in theliterature in order to study brandloyalty from the behavioural perspec-tive, the Dirichlet model being one ofthe most prominent.21–23 These ap-proaches model the consumers’ faithfulenactment of a promise to consistentlypurchase only one brand, althoughthey fail to model the reason(s) behindthis behaviour

One possible insight could befound in the attitudinal perspective inconceptualising loyalty According tothis perspective, brand loyalty con-sists of a strong internal dispositiontowards a brand leading to repeatedpurchases.24–26 As such, the attitudinalapproach conceives brand loyalty based

on stated preferences, commitment, orpurchase intentions One would ex-pect attitudinal and behavioural brandloyalty to be positively correlated,although not perfectly, otherwise therewould be little need for differentconcepts.27 Thus an increase in at-titudinal brand loyalty should lead to anincrease in behavioural brand loyalty

more accurate definition and thusoperationalisation of brand loyalty

Hence the first objective of this paper

is to conceive a better definition

of brand loyalty and validate itsoperationalisation

Furthermore, until now there havebeen few studies that have examinedthe antecedents of brand loyalty (forexample, Dick and Basu,13 Ha14 andHog et al.15) Hence the secondobjective of this study is to add to thisstream of research by empiricallyexamining the role of context inshaping the development of brandloyalty Finally, a third objective of thismanuscript is to empirically examinethe effects of brand loyalty on con-sumers’ behaviour For instance there isempirical evidence that demonstratesthat loyalty is not necessarily reflectedupon the systematic purchase of asingle brand.16 In fact, researchers havelong questioned whether the systematicpurchase of a single brand is the result

of increased levels of loyalty to thisbrand or whether it is the outcome ofloyalty to a store which carries alimited number of brands for a givenproduct category.17 Moreover, empiri-cal research has demonstrated thatbrand loyalty does not result only in aspecific purchase pattern For instance,

it can also bring about positive of-mouth communication, which isnot necessarily tied with the purchase

word-of the brand to which the consumerfeels loyal.18

The rest of the paper is organised asfollows First, the different types ofbrand loyalty are discussed Next, theauthors advance a conceptual modeland associated research hypotheses

Then a description is given of anempirical study designed to test thehypotheses and compare the effects of

This then implies that the attitudinalperspective is of limited value in grasp-ing the notion of loyalty If, however,there are changes in the marketplace, as

is often the case (for example, a new orimproved product is introduced, andthere is increased perceived risk), theconsumer is likely to engage in adecision-making process, breaking thecycle of habitual purchases

The cycle of purchases may or maynot break, however, if the consumerholds strong positive sentiments andidentifies with the brand Including theattitudinal perspective in conceptualis-ing loyalty is useful, since it allows thedecision-making process occurring inthe consumer’s mind during the pur-chase to be more realistically described

It is the cognitive activities that onedescribes with this perspective

Similarly, the cycle of purchases may

or may not break because of sures exercised by the consumer’ssocial environment Thus embodyingthe reasoned action approach in theproposed conceptualisation recognisesthe fact that there are some situa-tions where consumers’ behaviour isnot fully under their control, but isinfluenced by the expectations ofrelevant others

pres-Therefore, one could conceivebrand loyalty as comprising threedimensions Each of them determinesthe type of loyalty a consumer willexhibit towards a brand For instance,

a consumer who is unfavourable to thepurchase of a certain brand may stillpurchase the brand This loyaltybehaviour is likely to be converted into

a behaviour of switching the brandwhen the consumer is no longer forced

to keep purchasing the brand Thus the

following four generic types of brand

loyalty can be identified: ‘no loyalty’,

Another possible explanation can,

however, be derived from the theory

of reasoned action According to this

perspective, the consumer’s behaviour

may be influenced by social pressures,

thus explaining how a consumer’s

brand attitude may be unfavourable,

while the consumer repeats the

pur-chases of the particular brand In such

a case, the consumer’s brand loyalty

would be superficial.28Recognising the

above difficulties in defining and

ex-plaining brand loyalty, Ha29 proposed

the theory of reasoned action to

explain brand loyalty According to the

reasoned action paradigm — based on

the theory of reasoned action,

intro-duced by Fishbein30 — brand loyalty is

conceived as a notion that is dependent

on normative influences (such as

in-fluences deriving from social peers)

These influences, in turn, are reflected

in the behavioural consequences of

loyalty.31,32 According to this view, one

may hold a favourable attitude towards

a brand but still not purchase it because

of not being able to afford it, a partner

disliking the brand, or for many other

reasons.33,34 Such an individual,

al-though having never actually

pur-chased the brand, promotes it in public,

recommends it, and compels others to

buy it This situation is similar to the

theoretical discussion by Oliver35of the

loyalty phases, and particularly the

cognitive phase, where loyalty is based

merely on ‘brand belief’ and not on

brand experience

For the purposes of this research, a

conceptualisation of loyalty is adopted

that attempts to combine all three

approaches to brand loyalty in

ex-plaining purchasing behaviour Thus,

incorporating the behaviour paradigm

suggests that repeat purchases are often

the outcome of habitual behaviour

self-perception and personality Theconsumer trusts it and is willing torecommend it to peers, friends

or relatives, although, for reasonsbeyond the consumer’s control, thepurchase itself may never occur Insuch cases, the consumer is stronglydiscouraged to be loyal to a certainbrand by social influences Forinstance, a young, newly appointedlecturer in a business school mightcovet a Mercedes, but not purchase

it because he cannot afford it orbecause he might not wish topublicise his economic status Thelecturer may, however, still recom-mend the brand

— Inertia loyalty: An individual, though purchasing the brand, does

al-so out of habit, convenience or forsome other reason, but not as aconsequence of emotional attach-ment to the brand or a realsocial motive Inertia loyalty ischaracterised by a habitual attach-

‘inertia loyalty’, ‘premium loyalty’ and

‘covetous loyalty’ (see Figure 1)

The four types of brand loyalty arecharacterised as follows:

— No loyalty: No purchase at all, and

a complete lack of attachment tothe brand Also no social influences

to be even cognitively loyal to abrand

— Covetous loyalty: No purchase but,

unlike the case of ‘no loyalty’, the

individual exhibits a very high level

of relative attachment to the brand

as well as a strong positive position towards the brand, which isdeveloped from the social en-vironment This condition arisesfrom perceived human characteris-tics which a consumer identifies in

predis-a specific brpredis-and.36 The individualcomes to like the brand andthus emotional attachment with thebrand increases The brand becomes

an extension of the consumer’s own

High

Low High

Low

High COVETOUS

INERTIA NO

vinced that the selected brand is insome way the best brand to buy.43This conviction arises from bothpersonal and social motives Varia-tions in the price of their favouritebrand may affect the quantity of thebrand they purchase, but not thebrand they choose to buy, sincethese consumers are committed tothe brand.

CONCEPTUAL MODEL AND RESEARCH HYPOTHESES

Figure 2 presents the conceptual model

of brand loyalty that guides thisresearch The model shows potentialdrivers of brand loyalty These driversare classified in three basic categories:consumer drivers, brand drivers andsocial drivers By focusing on potentialdrivers, it may be possible to managebrand loyalty better In addition, themodel used in this study focuses onconsumers’ behavioural responses tobrand loyalty — word-of-mouth com-munication, buying alternative brand,going to different point of sale (store)and buying nothing

Consumer drivers

Both normative and empirical studieshave substantiated the importance ofthe individual’s characteristics in decid-ing to purchase a specific brand.44Twosuch characteristics are examined inthis study: risk aversion and varietyseeking Although many characteristics

of consumers may have an impact onthe decision to purchase a specificbrand, this study chooses to focus onthese two specific attributes, whichboth relate to how consumers handlerisks Loyalty has been described as ameans of handling the risk associated

ment that is to a large

ex-tent unemotional and convenience

driven The consumer may

sys-tematically choose the specific

brand over other brands, but this

choice involves little emotional

involvement, little personal

invest-ment, and no brand commitment.37

Hence this is a very fragile

relationship that may be easily

terminated by a rival product

capable of breaking the

con-sumer’s habitual behavioural

pat-tern Oliver38 terms this type of

loyalty ‘phantom loyalty’, while

Day39 and Dick and Basu40 call it

‘spurious loyalty’.

— Premium loyalty: An individual

exhibits a high degree of relative

attachment to the brand, a high

instance of repeat purchases, and

appears to be highly influenced by

social pressure Premium loyalty is

characterised by the greatest degree

of consumer attachment to the

brand, and in this case the

con-sumer purposefully seeks to

pur-chase the particular brand, while

attempting to overcome obstacles

This is similar to the

descrip-tion by Oliver41 of ‘action loyalty’

— ‘commitment to the action

of re-buying’ Premium loyalty

propels individuals to suffer various

sacrifices in order to acquire their

favoured brand Football fans are a

good, although extreme, example of

people showing this type of loyalty

They may see their team losing one

game after the other, and yet be

willing to travel with the team or

watch its games on television

Consumers who exhibit ‘premium

loyalty’ have been won over by the

brand alternative through the value

it provides to them42 or are

con-of rational market behaviour — inorder to reduce perceived risk Therisk element may be either a functionalrisk or a social acceptance risk In factthe perceived risk can be so intensethat individuals become reluctant toproceed with the action Instead, theyprocrastinate until they have reducedthe perceived complexity or the uncer-tainty associated with the situation.53,54

Hence the consumer’s need to controlthe risk will be a significant positivefactor in the formation of brand loyalty

On these grounds, this paper tigates the following hypothesis:

inves-H 1 : Risk aversion will relate to the type of brand loyalty the individual develops towards a specific brand.

Variety seeking

On the other hand, uncertainty of theoutcome of a purchase and the riskassociated with a certain decisionprovides stimulation to the consumer.55

If the stimulation obtained is withinthe bounds of the optimal stimulation

with the decision to purchase a specificbrand.45

Risk aversion

Individuals are often confronted withsituations that differ in the degree ofuncertainty or complexity they present

to them.46Typically, decisions linked tohighly valued goals47 such as thepurchase of a new car and/or deci-sions on high-involvement productcategories48 encompass greater risk forthe individual buyer Such decisionsmay evoke negative emotions that thebuyer attempts to deal with.49 Emo-tions accentuate the risk associatedwith the purchase of a specific brand,leading to a greater search,50 which, inturn, may lead to lower levels of brandloyalty On the other hand, emo-tions may lead consumers to ex-hibit avoidance behaviour51 and/orgreater dependence on previously heldchoices, which result in higher levels ofbrand loyalty Furthermore, Sheth andParvatiyar52 argued that consumersbecome brand loyal — a manifestation

CONSUMER DRIVERS Risk aversion

Variety seeking

BRAND DRIVERS Brand reputation Availability of substitute brands

SOCIAL DRIVERS Social group influences Peers’ recommendation

Type of brand loyalty

Buy nothing

Buy alternative brand

Word-of-mouth communication

Visit other store

Figure 2 Conceptual model

of the transaction and the cost if hedecides to acquire the product Hence

to understand how product tics contribute to brand loyalty, it isnecessary to comprehend what createsand determines the value individualsderive from a brand The followingtwo brand characteristics are examined

characteris-in this study: brand reputation andavailability of substitute brands

Brand reputation

Although not part of the physicalproduct itself, the reputation of abrand’s name has been described as anextrinsic cue, that is, an attributerelated to the product.61 A reputablebrand name conveys a strong indica-tion of the product’s quality and equitythat is not necessarily related to detailedknowledge of the intrinsic — technical

— specification of the product.62Therefore, the choice between alterna-tive brands within a single product class

is facilitated, since brands are tiated easily by their consumers AsOliver63 suggests, loyalty is not merelyabout product superiority and satisfyingcustomers Loyalty is about havingcustomers who can become deter-mined defenders of the brand Ifthe firm cannot develop, supportand maintain brand uniqueness andperceived brand equity, then it isnot possible to expect loyalty todevelop.64

differen-Thus having a brand with a strongreputation will be a significant positivefactor in the formation of brand loyalty,since the brand’s reputation strengthensits perceived equity.65,66 Moreover, thereputation of the brand strengthens thehabitual behaviour of consumers byrewarding their choice and making thebrand more desirable and alluring.67As

level, then it is desirable and the

consumer actively seeks to attain it

However, if the optimal stimulation

level is exceeded, it becomes too

intensive, leading consumers to try to

reduce the complexities that are

as-sociated with such a condition and, as

pointed out earlier, they attempt to

routinise the decision-making process

and its outcomes In fact, Sheth and

Parvatiyar56 pointed out that

routinisa-tion and variety-seeking behaviour

become cyclical over time, but the

cycles are asymmetrical in favour of the

longer duration of routinised

be-haviour

Routinisation, although initially

helpful, may, however, lead an

individual to feelings of monotony and

boredom, which may lead to

ex-perimentation with new brands.57

Moreover, it appears that the level of

variety-seeking behaviour depends on

the intrinsic need of consumers to seek

variety (personal differences) and

on the product category level of

involvement.58 It is within this

framework that studies report a break

in the link between satisfaction and

loyalty.59 Indeed, as Homburg and

Giering60 have demonstrated, variety

seeking is one of the key consumer

characteristics which moderates the

relationship between perceived quality

and satisfaction with the loyalty to a

specific brand Within this framework,

this paper hypothesises that:

H 2 : Variety-seeking behaviour will relate to

the type of brand loyalty an individual

develops towards a specific brand.

Brand drivers

An individual’s intention to purchase a

product reflects a search for value out

do so in the absence of any attractivealternative — as is the case when

no substitute brands are available73

— the relationship tends to lastonly for as long as there is noalternative.74 Research suggests thatcustomers in such constrained situa-tions attempt to restore their freedom

to choose.75 According to dependence theory,76 consumers mayattempt to break free from constrainedrelationships by identifying acceptablesubstitutes Hence the perception ofsimilar substitutes may be expected toinfluence negatively the creation ofrelational ties to the brand within thespecific category, and it might therefore

resource-be considered as a deterrent to theformation of brand loyalty On thesegrounds the following hypothesis isinvestigated:

H 4 : The availability of substitute products will influence the type of brand loyalty

an individual develops towards a ific brand.

spec-Social drivers

Finally, when studying the antecedents

of loyalty, one should not neglect thesocial norms which may influence con-sumers’ behaviour patterns Consumers

do not take decisions isolated fromsocial influences Rather, they are sub-jected to heavy social control over theattitudes they have and the behaviourthey develop.77

Social group influences

One strong type of such social fluence is that derived from referencegroups — the social groups that have

in-a direct or indirect influence on theperson’s attitude or behaviour.78 In the

a result, reputable brands enjoy higherloyalty due to their higher marketshare.68 This higher market share isattributed to the fact that higher-sharebrands are not only bought by moreconsumers, but they are also boughtmore frequently In other words, high-share brands benefit both from greatermarket penetration and higher pur-chase frequency This is the well-known double-jeopardy phenomenon,

an ‘empirical law’ that researchers haveobserved and modelled for nearly 30years.69 On these grounds, the follow-ing hypothesis is investigated:

H 3 : Brand reputation will relate to the brand loyalty type an individual develops towards a specific brand.

Availability of substitute brands

Brand reputation is subjected to theshopper’s perception of both the range

of competing products and brands

as well as the class of substituteproducts

When a product class comprisesseveral brands which are perceived byconsumers to be similar to eachother, discriminating among them ishard Consequently, individuals have

no reason to show loyalty towards one

or another In fact, the more alike thebrands are perceived to be, the lesslikely loyalty is to emerge.70 Rather,consumers are prone to make theirpurchases from a predetermined set ofalternative products without showing aparticular preference to any specificbrand from this set.71,72 Thus theavailability of substitute products isexpected to affect brand loyalty sig-nificantly

Moreover, when customers stay in arelationship because they are forced to

loyalty, since the desire for the brandmay be affected by group preference.87

On these grounds, this paper tigates the following hypothesis:

inves-H 5 : Social group influences will relate to the type of brand loyalty an individual develops towards a specific brand.

Peers’ recommendation

Another strong source of socialinfluence is the recommendations andsuggestions made by the individual’speers Hite and Hite88 found that aparty’s reputation could lead to posi-tive expectations about the partywhich, in turn, leads other parties todevelop reciprocity and loyalty for thereputable party When it comes tobrand names, their reputation reflectsthe opinion of others that a specificbrand possesses or does not possesscertain characteristics.89 While adver-tising and/or public relations helpbrands to demonstrate their qualities,peers are among the most influentialsources of information used by con-sumers in shaping their opinionconcerning a brand’s qualities.90 Peersexercise both normative (conformist)and identificational influences on con-sumers Informative influences help toguide consumers in product, brandand store searches,91 whereas norma-tive influences direct and controlevaluations, choices and loyalties.92Thus peers’ recommendations areexpected to significantly affect brandloyalty.93

Following the reasoned-actionparadigm, Bearden and Etzel94 suggest,however, that the recommendations ofpeers may not necessarily convert intoactions (ie purchase) Under certaincircumstances, they could merely

context of the present work, two types

of social influence are considered:

so-cial group influences and peers’

recom-mendations

A group becomes a reference one

when an individual identifies with it so

much that he takes on many of the

values, attitudes and/or behaviour of its

members.79The power of the influence

of a reference group is dependent on

the individual’s susceptibility to this

influence, the strength of his

involve-ment with the group and the degree

of product conspicuousness.80 Powerful

reference groups may easily change the

behaviour of their members, or their

aspirant members, and align it more

with the norms and standards that the

group considers to be acceptable.81

Hence the individual’s loyalty

towards a product is also dependent on

the acceptance of his preference for a

certain product by the social group the

individual refers to, particularly when

the conditions under which individuals

feel coerced to give in to the group’s

norms are met By adapting their

attitudes and behaviour, consumers

fulfil their aspirations and at the same

time reduce the perceived risk of

making a decision.82 Besides, recent

empirical studies have attested to the

impact of social stimuli (or normative

information) on loyalty.83,84

For instance, Mascarenhas and

Higby,85 in their study of how

youngsters choose a brand, indicated

that parents’ consistent choice of a

particular brand influences children to

perceive the brand as ‘good’, and thus

become loyal to it Furthermore, Hog

groups led young consumers to form a

more positive image of a brand Hence

group social influences are expected to

have a strong positive impact on brand

On these grounds this paper tigates the following hypothesis:

inves-H 7 : The type of brand loyalty will depend

on the occurrence or not of mouth communication between con- sumers.

word-of-Buy alternative brand

An interesting situation arises when

a consumer is loyal to a specificbrand, but the brand is unavailablewhen required at a particular store.98

How likely is it that the individualwill betray the brand and purchaseanother?

Oliver99 has shown that consumers,when faced with uncertainty abouthow to handle a decision and about itsoutcomes (concerning, for example,specifying relative uncertainties, whatinformation to seek, or how to assessconsequences), tend to delay the actualdecision This is in line with theempirical findings of Greenleaf andLehmann,100 demonstrating that suchprocedural uncertainty causes con-sumers to delay a decision Hence, forinstance, when consumers are deprived

of the brand towards which they havedeveloped a feeling of loyalty, theymay delay their purchase until either

‘their’ brand is available again or theyhave managed to handle the newsituation

On the other hand, some sumers might find delaying the pur-chase too ‘costly’ and thus decide

con-to switch brands Many consumersadapt their brand preferences accord-ing to the time when they prefer

to shop.101 Therefore they wouldrather stay in one store and switchbrands Such behaviour is in line withthe consequences of developing the

influence the consumer’s emotionalattachment to the brand Consider, forinstance, a young teenager whodevelops a high attachment to Sony’sPlaystation II after it wasrecommended by a friend, but still feelsreluctant to purchase it because heperceives that his parents woulddisapprove Nonetheless, following theconceptualisation of loyalty outlined,the teenager in question is (covetously)loyal to the brand On these grounds,this paper investigates the followinghypothesis:

H 6 : Peers’ recommendation will relate to the type of brand loyalty an individual develops towards a specific brand.

Consequences of loyalty

Scholars studying the notion of brandloyalty have discussed a number ofbehavioural consequences In thecontext of the present work, fouralternative consequences of loyalty areexamined, namely: word-of-mouthcommunication, buy alternative brand,

go to different point of sale (store) andbuy nothing

Word-of-mouth communication

Perhaps the single most expected havioural outcome of loyalty is brandrecommendation Consumers becomeloyal as a result of the satisfactionthey experience with their purchase.95

be-Satisfied consumers who share theirexperiences with other individuals arethe best advocators of any company orits products.96 In fact, as Oliver97 sug-gests, in certain cases it is the sharing ofthe experience regarding the brand thatultimately provides the satisfaction andnot the brand itself

available), and in doing this there is lessrisk associated with the decision Theindividual has developed the highestlevel of loyalty — ‘action loyalty’.Betraying the brand will be likebetraying himself.104 Based on theabove discussion, this paper investigatesthe following hypothesis:

H 10 : The type of brand loyalty will relate

to whether the consumer decides to buy nothing if the brand is unavail- able.

METHODOLOGY Data collection and sample

The sample for the study was drawnfrom the area of Athens, Greece.Trained personnel conducted inter-views, based on a questionnaire, inorder to increase the validity andreliability of the responses

The sample consisted of 850 sumers of whisky who were randomlyapproached in the street and shop-ping malls Although the sample wasclearly chosen for convenience, theinterviews were conducted at differentlocations and on different days, as well

con-as at uniform intervals, in order toreduce location-, date- and time-re-lated response bias

The decision to focus the study onwhisky buyers was based on threefactors First, there is a remarkablevariety of alternative whisky brands inthe Greek market, a fact whichgives many options to the shopper.Moreover, whisky is bought quiteoften by the majority of the adultGreek population Consequently, itcould reasonably be expected thatsignals of loyalty, where identified interms of behaviour, emotional attach-

‘spurious’ type of loyalty as

sug-gested by Dick and Basu.102 On these

grounds, this paper investigates the

following hypothesis:

H 8 : The type of brand loyalty will relate

to the purchase or not of alternative

brands.

Go to a different store

In the absence of the desired brand,

loyal consumers may choose to go to

a different point of sale to seek the

brand Once more, this behaviour

depends on the perceived risk

as-sociated with the decision of

purchas-ing an alternative brand, but perhaps

also with the emotional

disappoint-ment of not finding the brand with

which the consumer has an emotional

attachment.103 It can be expected that

the decision of whether to go to a

different store to find the desired brand

will be determined by the type of

brand loyalty Based on the above

discussion, this paper investigates the

following hypothesis:

H 9 : The type of brand loyalty will affect

the decision to go to a different store.

Buy nothing

The decision to buy nothing if the

preferred brand is unavailable is by

definition a strong indicator of

premium loyalty To make this

decision, the individual has to go

through the same cognitive/conative

process to decide on an alternative as

he did originally to choose the

preferred product, and this creates both

cognitive and emotional discomfort

The result is that the decision to buy

is postponed (for when the product is

This process resulted in a battery of 21items worded to capture the differenttypes of loyalty Next, a mini pilotsurvey was conducted among a ran-domly selected sample of 250 stu-dents Exploratory factor analysis wasemployed to refine their answers onloyalty by deleting items with highloadings on multiple factors Thisprocess resulted in a four-factor solu-

tion, namely premium loyalty prised of three items), covetous loyalty (three items), inertia loyalty (four items) and no loyalty (five items) These 15

(com-items were employed in the analysis ofthe results, which is reported in thismanuscript

Dimensionality and psychometric attributes

of the loyalty scale

In the main study the psychometricattributes of the scales employed tomeasure loyalty were assessed usingconfirmatory factor analysis (CFA).This was preferred to exploratoryfactor analysis (EFA), a method aboutwhich various concerns have beenraised For instance, according toMulaik,106 EFA may ‘find optimalknowledge’ (p 265) Mulaik107made itclear that ‘there is no rationallyoptimal way to extract knowledgefrom experience without making cer-tain prior assumptions’ (p 265) Inaddition, the factor structures yielded

by an exploratory factor analysis aredetermined by the mechanics of themethod and are dependent on specifictheories and the mechanics of extrac-tion and rotation procedures This,too, can result in inaccurate results.Mulaik108 also made it clear thatexploratory techniques do not provideany way of indicating when something

is wrong with one’s assumptions,

ment or social stimuli, would beauthentic and not the consequence of

a constrained choice Finally, the largenumber of alternative whisky brandsthat can be found in the Greek marketdiffer markedly in terms of theirpositioning strategy (product, price,promotion and distribution) in Greece

The same is also true for the ing occasion Whisky is bought inGreece for private or public (in-home

purchas-or on-premises) consumption, as well

as for offering as a gift In addition,whisky can be bought legally fromoff-licences, supermarkets and con-venience stores Therefore, the authorsexpected that there would be enoughscope for all types of loyalty todevelop

Measures

Although attempts were made to useexisting measures, they were not avail-able for several constructs or wereotherwise limited in their applicability

to the context of this study It wastherefore necessary to adapt the currentmeasures or develop new ones (as dis-cussed subsequently)

Development of the loyalty scale

The review of the current ture did not reveal any empiricallyvalidated scales with respect to thedifferent types of loyalty Thus newscales were developed to measurethe loyalty construct In doing so,the scale-development instructions sug-gested by Churchill105 were followed

litera-More specifically, once the domain ofthe construct (brand loyalty) was speci-fied, the preliminary set of items wasdeveloped through discussion withconsumers and group brand managers