servqual servperf

Measuring Service Quality:SERVQUAL vs. SERVPERF Scales

Sanjay K Jain and Garima Gupta

Quality has come to be recognized as a strategic tool for attaining operational efficiencyand improved business performance. This is true for both the goods and services sectors.However, the problem with management of service quality in service firms is that qualityis not easily identifiable and measurable due to inherent characteristics of services whichmake them different from goods. Various definitions of the term ‘service quality’ havebeen proposed in the past and, based on different definitions, different scales formeasuring service quality have been put forward. SERVQUAL and SERVPERF constitutetwo major service quality measurement scales. The consensus, however, continues toelude till date as to which one is superior.

An ideal service quality scale is one that is not only psychometrically sound but is alsodiagnostically robust enough to provide insights to the managers for corrective actionsin the event of quality shortfalls. Empirical studies evaluating validity, reliability, andmethodological soundness of service quality scales clearly point to the superiority of theSERVPERF scale. The diagnostic ability of the scales, however, has not been explicitlyexplicated and empirically verified in the past.

The present study aims at filling this void in service quality literature. It assesses thediagnostic power of the two service quality scales. Validity and methodologicalsoundness of these scales have also been probed in the Indian context — an aspect whichhas so far remained neglected due to preoccupation of the past studies with serviceindustries in the developed world.

Using data collected through a survey of consumers of fast food restaurants in Delhi,the study finds the SERVPERF scale to be providing a more convergent and discriminant-valid explanation of service quality construct. However, the scale is found deficient inits diagnostic power. It is the SERVQUAL scale which outperforms the SERVPERF scaleby virtue of possessing higher diagnostic power to pinpoint areas for managerialinterventions in the event of service quality shortfalls.

The major managerial implications of the study are:Because of its psychometric soundness and greater instrument parsimoniousness,one should employ the SERVPERF scale for assessing overall service quality of afirm. The SERVPERF scale should also be the preferred research instrument whenone is interested in undertaking service quality comparisons across serviceindustries.On the other hand, when the research objective is to identify areas relating toservice quality shortfalls for possible intervention by the managers, the SERVQUALscale needs to be preferred because of its superior diagnostic power.

However, one serious problem with the SERVQUAL scale is that it entails giganticdata collection task. Employing a lengthy questionnaire, one is required to collect dataabout consumers’ expectations as well as perceptions of a firm’s performance on eachof the 22 service quality scale attributes.

Addition of importance weights can further add to the diagnostic power of theSERVQUAL scale, but the choice needs to be weighed against the additional task of datacollection. Collecting data on importance scores relating to each of the 22 service attributesis indeed a major deterrent. However, alternative, less tedious approaches, discussed to-wards the end of the paper, can be employed by the researchers to obviate the data col-lection task.

KEY WORDS

Service Quality

Measurement of Service Quality

Service Quality Scale

Scale Validity and Reliability

Diagnostic Ability of Scale

Executive Summary

R E S E A R C Hincludes research articles that focus on theanalysis and resolution of managerial andacademic issues based on analytical andempirical or case research

VIKALPA • VOLUME 29 • NO 2 • APRIL - JUNE 2004 25

25

Quality has come to be recognized as a strategictool for attaining operational efficiency andimproved business performance (Anderson and

Zeithaml, 1984; Babakus and Boller, 1992; Garvin, 1983;Phillips, Chang and Buzzell, 1983). This is true for theservices sector too. Several authors have discussed theunique importance of quality to service firms (e.g.,Normann, 1984; Shaw, 1978) and have demonstrated itspositive relationship with profits, increased market share,return on investment, customer satisfaction, and futurepurchase intentions (Anderson, Fornell and Lehmann1994; Boulding et al., 1993; Buzzell and Gale, 1987; Rustand Oliver, 1994). One obvious conclusion of these studiesis that firms with superior quality products outperformthose marketing inferior quality products.

Notwithstanding the recognized importance ofservice quality, there have been methodological issuesand application problems with regard to its operation-alization. Quality in the context of service industrieshas been conceptualized differently and based on dif-ferent conceptualizations, alternative scales have beenproposed for service quality measurement (see, forinstance, Brady, Cronin and Brand, 2002; Cronin andTaylor, 1992, 1994; Dabholkar, Shepherd and Thorpe,2000; Parasu- raman, Zeithaml and Berry, 1985, 1988).Despite considerable work undertaken in the area, thereis no consensus yet as to which one of the measurementscales is robust enough for measuring and comparingservice quality. One major problem with past studieshas been their preoccupation with assessing psycho-metric and methodological soundness of service scalesthat too in the context of service industries in the de-veloped countries. Virtually no empirical efforts havebeen made to evaluate the diagnostic ability of thescales in providing managerial insights for correctiveactions in the event of quality shortfalls. Furthermore,little work has been done to examine the applicabilityof these scales to the service industries in developingcountries.

This paper, therefore, is an attempt to fill this existingvoid in the services quality literature. Based on a surveyof consumers of fast food restaurants in Delhi, this paperassesses the diagnostic usefulness as well as the psycho-metric and methodological soundness of the two widelyadvocated service quality scales, viz., SERVQUAL andSERVPERF.

SERVICE QUALITY: CONCEPTUALIZATIONAND OPERATIONALIZATION

Quality has been defined differently by different au-thors. Some prominent definitions include ‘conformanceto requirements’ (Crosby, 1984), ‘fitness for use’ (Juran,1988) or ‘one that satisfies the customer’ (Eiglier andLangeard, 1987). As per the Japanese production phi-losophy, quality implies ‘zero defects’ in the firm’sofferings.

Though initial efforts in defining and measuringservice quality emanated largely from the goods sector,a solid foundation for research work in the area was laiddown in the mid-eighties by Parasuraman, Zeithaml andBerry (1985). They were amongst the earliest researchersto emphatically point out that the concept of qualityprevalent in the goods sector is not extendable to theservices sector. Being inherently and essentially intan-gible, heterogeneous, perishable, and entailing simulta-neity and inseparability of production and consump-tion, services require a distinct framework for qualityexplication and measurement. As against the goods sectorwhere tangible cues exist to enable consumers to eva-luate product quality, quality in the service context isexplicated in terms of parameters that largely comeunder the domain of ‘experience’ and ‘credence’ prop-erties and are as such difficult to measure and evaluate(Parasuraman, Zeithaml and Berry, 1985; Zeithaml andBitner, 2001).

One major contribution of Parasuraman, Zeithamland Berry (1988) was to provide a terse definition ofservice quality. They defined service quality as ‘a globaljudgment, or attitude, relating to the superiority of theservice’, and explicated it as involving evaluations of theoutcome (i.e., what the customer actually receives fromservice) and process of service act (i.e., the manner in whichservice is delivered). In line with the propositions putforward by Gronroos (1982) and Smith and Houston(1982), Parasuraman, Zeithaml and Berry (1985, 1988)posited and operationalized service quality as a differ-ence between consumer expectations of ‘what they want’and their perceptions of ‘what they get.’ Based on thisconceptualization and operationalization, they proposeda service quality measurement scale called ‘SERVQUAL.’The SERVQUAL scale constitutes an important land-mark in the service quality literature and has beenextensively applied in different service settings.

26 MEASURING SERVICE QUALITY: SERVQUAL vs. SERVPERF SCALES

26

Over time, a few variants of the scale have also beenproposed. The ‘SERVPERF’ scale is one such scale thathas been put forward by Cronin and Taylor (1992) inthe early nineties. Numerous studies have been under-taken to assess the superiority of two scales, but con-sensus continues to elude as to which one is a betterscale. The following two sections provide an overviewof the operationalization and methodological issuesconcerning these two scales.

SERVQUAL Scale

The foundation for the SERVQUAL scale is the gapmodel proposed by Parasuraman, Zeithaml and Berry(1985, 1988). With roots in disconfirmation paradigm,1

the gap model maintains that satisfaction is related tothe size and direction of disconfirmation of a person’sexperience vis-à-vis his/her initial expectations (Church-ill and Surprenant, 1982; Parasuraman, Zeithaml andBerry, 1985; Smith and Houston, 1982). As a gap ordifference between customer ‘expectations’ and ‘percep-tions,’ service quality is viewed as lying along a con-tinuum ranging from ‘ideal quality’ to ‘totally unaccept-able quality,’ with some points along the continuumrepresenting satisfactory quality. Parasuraman, Zeith-aml and Berry (1988) held that when perceived or ex-perienced service is less than expected service, it impliesless than satisfactory service quality. But, when per-ceived service is less than expected service, the obviousinference is that service quality is more than satisfactory.Parasuraman, Zeithaml and Berry (1988) posited thatwhile a negative discrepancy between perceptions andexpectations — a ‘performance-gap’ as they call it —causes dissatisfaction, a positive discrepancy leads toconsumer delight.

Based on their empirical work, they identified a setof 22 variables/items tapping five different dimensionsof service quality construct.2 Since they operationalizedservice quality as being a gap between customer’s ex-pectations and perceptions of performance on thesevariables, their service quality measurement scale iscomprised of a total of 44 items (22 for expectations and22 for perceptions). Customers’ responses to their ex-pectations and perceptions are obtained on a 7-pointLikert scale and are compared to arrive at (P-E) gapscores. The higher (more positive) the perception minusexpectation score, the higher is perceived to be the levelof service quality. In an equation form, their operation-alization of service quality can be expressed as follows:

∑=

−=k

1jijiji )EP(SQ (1)

where: SQi = perceived service quality of indivi-dual ‘i’

k = number of service attributes/items P = perception of individual ‘i’ with res-

pect to performance of a service firmattribute ‘j’

E = service quality expectation for at-tribute ‘j’ that is the relevant norm forindividual ‘i’

The importance of Parasuraman, Zeithaml andBerry’s (1988) scale is evident by its application in anumber of empirical studies across varied service set-tings (Brown and Swartz, 1989; Carman, 1990; Kassimand Bojei, 2002; Lewis, 1987, 1991; Pitt, Gosthuizen andMorris, 1992; Witkowski and Wolfinbarger, 2002; Young,Cunningham and Lee, 1994). Despite its extensive ap-plication, the SERVQUAL scale has been criticized onvarious conceptual and operational grounds. Some majorobjections against the scale relate to use of (P-E) gapscores, length of the questionnaire, predictive power ofthe instrument, and validity of the five-dimension struc-ture (e.g., Babakus and Boller, 1992; Cronin and Taylor,1992; Dabholkar, Shepherd and Thorpe, 2000; Teas, 1993,1994). Since this paper does not purport to examinedimensionality issue, we shall confine ourselves to adiscussion of only the first three problem areas.

Several issues have been raised with regard to useof (P-E) gap scores, i.e., disconfirmation model. Moststudies have found a poor fit between service qualityas measured through Parasuraman, Zeithaml and Ber-ry’s (1988) scale and the overall service quality measureddirectly through a single-item scale (e.g., Babakus andBoller, 1992; Babakus and Mangold, 1989; Carman, 1990;Finn and Lamb, 1991; Spreng and Singh, 1993). Thoughthe use of gap scores is intuitively appealing and con-ceptually sensible, the ability of these scores to provideadditional information beyond that already containedin the perception component of service quality scale isunder doubt (Babakus and Boller, 1992; Iacobucci,Grayson and Ostrom, 1994). Pointing to conceptual,theoretical, and measurement problems associated withthe disconfirmation model, Teas (1993, 1994) observedthat a (P-E) gap of magnitude ‘-1’ can be produced insix ways: P=1, E=2; P=2, E=3; P=3, E=4; P=4, E=5; P=5,E=6 and P=6, E=7 and these tied gaps cannot be con-


27

strued as implying equal perceived service qualityshortfalls. In a similar vein, the empirical study by Peter,Churchill and Brown (1993) found difference scores beingbeset with psychometric problems and, therefore, cau-tioned against the use of (P-E) scores.

Validity of (P-E) measurement framework has alsocome under attack due to problems with the conceptu-alization and measurement of expectation component ofthe SERVQUAL scale. While perception (P) is definableand measurable in a straightforward manner as the con-sumer’s belief about service is experienced, expectation(E) is subject to multiple interpretations and as such hasbeen operationalized differently by different authors/researchers (e.g., Babakus and Inhofe, 1991; Brown andSwartz, 1989; Dabholkar et al., 2000; Gronroos, 1990;Teas, 1993, 1994). Initially, Parasuraman, Zeithaml andBerry (1985, 1988) defined expectation close on the linesof Miller (1977) as ‘desires or wants of consumers,’ i.e.,what they feel a service provider should offer rather thanwould offer. This conceptualization was based on thereasoning that the term ‘expectation’ has been useddifferently in service quality literature than in the cus-tomer satisfaction literature where it is defined as aprediction of future events, i.e., what customers feel aservice provider would offer. Parasuraman, Berry andZeithaml (1990) labelled this ‘should be’ expectation as‘normative expectation,’ and posited it as being similarto ‘ideal expectation’ (Zeithaml and Parasuraman, 1991).Later, realizing the problem with this interpretation,they themselves proposed a revised expectation (E*)measure, i.e., what the customer would expect from‘excellent’ service (Parasuraman, Zeithaml and Berry,1994).

It is because of the vagueness of the expectationconcept that some researchers like Babakus and Boller(1992), Bolton and Drew (1991a), Brown, Churchill andPeter (1993), and Carman (1990) stressed the need fordeveloping a methodologically more precise scale. TheSERVPERF scale — developed by Cronin and Taylor(1992) — is one of the important variants of the SERV-QUAL scale. For, being based on the perception com-ponent alone, it has been conceptually and methodolog-ically posited as a better scale than the SERVQUAL scalewhich has its origin in disconfirmation paradigm.

SERVPERF Scale

Cronin and Taylor (1992) were amongst the researcherswho levelled maximum attack on the SERVQUAL scale.

They questioned the conceptual basis of the SERVQUALscale and found it confusing with service satisfaction.They, therefore, opined that expectation (E) componentof SERVQUAL be discarded and instead performance(P) component alone be used. They proposed what isreferred to as the ‘SERVPERF’ scale. Besides theoreticalarguments, Cronin and Taylor (1992) provided empir-ical evidence across four industries (namely banks, pestcontrol, dry cleaning, and fast food) to corroborate thesuperiority of their ‘performance-only’ instrument overdisconfirmation-based SERVQUAL scale.

Being a variant of the SERVQUAL scale and con-taining perceived performance component alone, ‘per-formance only’ scale is comprised of only 22 items. Ahigher perceived performance implies higher servicequality. In equation form, it can be expressed as:

∑=

=k

1jiji PSQ (2)

where: SQi = perceived service quality of indi-vidual ‘i’

k = number of attributes/items P = perception of individual ‘i’ with

respect to performance of a servicefirm on attribute ‘j’

Methodologically, the SERVPERF scale representsmarked improvement over the SERVQUAL scale. Notonly is the scale more efficient in reducing the numberof items to be measured by 50 per cent, it has also beenempirically found superior to the SERVQUAL scale forbeing able to explain greater variance in the overallservice quality measured through the use of single-itemscale. This explains the considerable support that hasemerged over time in favour of the SERVPERF scale(Babakus and Boller, 1992; Bolton and Drew, 1991b;Boulding et al., 1993; Churchill and Surprenant, 1982;Gotlieb, Grewal and Brown, 1994; Hartline and Ferrell,1996; Mazis, Antola and Klippel, 1975; Woodruff, Ca-dotte and Jenkins, 1983). Though still lagging behind theSERVQUAL scale in application, researchers have in-creasingly started making use of the performance-onlymeasure of service quality (Andaleeb and Basu, 1994;Babakus and Boller, 1992; Boulding et al., 1993; Bradyet al., 2002; Cronin et al., 2000; Cronin and Taylor, 1992,1994). Also when applied in conjunction with the SERV-QUAL scale, the SERVPERF measure has outperformedthe SERVQUAL scale (Babakus and Boller, 1992; Brady,Cronin and Brand, 2002; Cronin and Taylor, 1992;


28

Dabholkar et al., 2000). Seeing its superiority, evenZeithaml (one of the founders of the SERVQUAL scale)in a recent study observed that “…Our results are in-compatible with both the one-dimensional view of ex-pectations and the gap formation for service quality.Instead, we find that perceived quality is directly influ-enced only by perceptions (of performance)” (Bouldinget al., 1993). This admittance cogently lends a testimonyto the superiority of the SERVPERF scale.

Service Quality Measurement:Unweighted and Weighted Paradigms

The significance of various quality attributes used in theservice quality scales can considerably differ acrossdifferent types of services and service customers. Secu-rity, for instance, might be a prime determinant of qualityfor bank customers but may not mean much to customersof a beauty parlour. Since service quality attributes arenot expected to be equally important across serviceindustries, it has been suggested to include importanceweights in the service quality measurement scales (Croninand Taylor, 1992; Parasuraman, Zeithaml and Berry,1995, 1998; Parasuraman, Berry and Zeithaml, 1991;Zeithaml, Parasuraman and Berry, 1990). While theunweighted measures of the SERVQUAL and theSERVPERF scales have been described above vide equa-tions (1) and (2), the weighted versions of the SERV-QUAL and the SERVPERF scales as proposed by Croninand Taylor (1992) are as follows:

∑=

−=k

1jijijiji )EP(ISQ (3)

∑=

=k

1jijiji )P(ISQ (4)

where: Iij is the weighting factor, i.e., importanceof attribute ‘j’ to an individual ‘i.’

Though, on theoretical grounds, addition of weightsmakes sense (Bolton and Drew, 1991a), not much im-provement in the measurement potency of either scalehas been reported after inclusion of importance weights.Between weighted versions of two scales, weightedSERVPERF scale has been theoretically posited to besuperior to weighted SERVQUAL scale (Bolton and Drew,1991a).

As pointed out earlier, one major problem with thepast studies has been their preoccupation with assess-ment of psychometric and methodological soundness of

the two scales. The diagnostic ability of the scales hasnot been explicitly explicated and empirically investi-gated. The psychometric and methodological aspects ofa scale are no doubt important considerations but onecannot overlook the assessment of the diagnostic powerof the scales. From the strategy formulation point ofview, it is rather the diagnostic ability of the scale thatcan help managers in ascertaining where the qualityshortfalls prevail and what possibly can be done to closedown the gaps.

METHODOLOGY

The present study is an attempt to make a comparativeassessment of the SERVQUAL and the SERVPERF scalesin the Indian context in terms of their validity, abilityto explain variance in the overall service quality, powerto distinguish among service objects/firms, parsimonyin data collection, and, more importantly, their diagnos-tic ability to provide insights for managerial interven-tions in case of quality shortfalls. Data for making com-parisons among the unweighted and weighted versionsof the two scales were collected through a survey of theconsumers of the fast food restaurants in Delhi. The fastfood restaurants were chosen due to their growingfamiliarity and popularity with the respondents understudy. Another reason was that the fast food restaurantservices fall mid way on the ‘pure goods - pure service’continuum (Kotler, 2003). Seldom are the extremes foundin most service businesses. For ensuring a greater gen-eralizability of service quality scales, it was considereddesirable to select a service offering that is comprisedof both the good (i.e., food) and service (i.e., preparationand delivery of food) components. Eight fast food res-taurants (Nirulas, Wimpy, Dominos, McDonald, PizzaHut, Haldiram, Bikanerwala, and Rameshwar) rated asmore familiar and patronized restaurants in differentparts of Delhi in the pilot survey were selected.

Using the personal survey method, 300 studentsand lecturers of different colleges and departments ofthe University of Delhi spread all over the city of Delhiwere approached. The field work was done duringDecember 2001-March 2002. After repeated follow-ups,only 200 duly filled-in questionnaires could be collectedconstituting a 67 per cent response rate. The sample wasdeliberately restricted to students and lecturers of DelhiUniversity and was equally divided between these twogroups. The idea underlying the selection of these twocategories of respondents was their easy accessibility.


29

Quota sampling was employed for selecting respon-dents from these two groups. Each respondent was askedto give information about two restaurants — one ‘mostfrequently visited’ and one ‘least frequently visited.’ Atthe analysis stage, collected data were pooled togetherthus constituting a total of 400 responses.

Parasuraman, Zeithaml and Berry’s (1988) 22-itemSERVQUAL instrument was employed for collecting thedata regarding the respondents’ expectations, percep-tions, and importance weights of various service at-tributes. Wherever required, slight modifications in thewording of scale items were made to make the question-naire understandable to the surveyed respondents. Someof the items were negatively worded to avoid the prob-lem of routine ticking of items by the respondents. Inaddition to the above mentioned 66 scale items (22 eachfor expectations, perceptions, and importance rating),the questionnaire included items relating to overallquality, overall satisfaction, and behavioural intentionsof the consumers. These items were included to assessthe validity of the multi-item service quality scales usedat our end. The single-item direct measures of overallservice quality, namely, ‘overall quality of these restau-rants is excellent’ and overall satisfaction, namely, ‘over-all I feel satisfied with the services provided’ were used.Cronin and Taylor (1992) have used similar measuresfor assessing validity of multi-item service quality scales.Behavioural intentions were measured with the help ofa 3-item scale as suggested by Zeithaml and Parasur-aman (1996) and later used by Brady and Robertson(2001) and Brady,Cronin and Brand (2002).3

Excepting importance weights and behaviouralitems, responses to all the scale items were obtained ona 5-point Likert scale ranging from ‘5’ for ‘strongly agree’to ‘1’ for ‘strongly disagree.’ A 4-point Likert scaleanchored on ‘4’ for ‘very important’ and ‘1’ for ‘notimportant’ was used for measuring importance weights

of each item. Responses to behavioural intention itemswere obtained using a 5-point Likert scale ranging from‘1’ for ‘very low’ to ‘5’ for ‘very high.’

FINDINGS AND DISCUSSION

Validity of Alternative Measurement Scales

As suggested by Churchill (1979), convergent and dis-criminant validity of four measurement scales wasassessed by computing correlations coefficients for dif-ferent pairs of scales. The results are summarized inTable 1. The presence of a high correlation betweenalternate measures of service quality is a pointer to theconvergent validity of all the four scales. The SERVPERFscale is, however, found having a stronger correlationwith other similar measures, viz., SERVQUAL andimportance weighted service quality measures.

A higher correlation found between two differentmeasures of the same variable than that found betweenthe measure of a variable and other variable implies thepresence of discriminant validity (Churchill, 1979) inrespect of all the four multi-item service quality scales.Once again, it is the SERVPERF scale which is foundpossessing the highest discriminant validity.

SERVPERF is, thus, found providing a more con-vergent as well as discriminant valid explanation ofservice quality.

Explanatory Power of AlternativeMeasurement Scales

The ability of a scale to explain the variation in theoverall service quality (measured directly through asingle-item scale) was assessed by regressing respond-ents’ perceptions of overall service quality on its corre-sponding multi-item service quality scale. Adjusted R2

values reported in Table 2 clearly point to the superiorityof SERVPERF scale for being able to explain greater

Table 1: Alternate Service Quality Scales and Other Measures — Correlation Coefficients

SERVQUAL SERVPERF Weighted Weighted Overall Overall Behavioural(P-E) (P) SERVQUAL SERVPERF Service Satisfaction Intentions

I (P-E) I (P) Quality

SERVQUAL (P-E) 1.000

SERVPERF (P) .735 -

Weighted SERVQUAL I(P-E) .995 .767 -

Weighted SERVPERF I(P) .759 .993 .772 -

Overall service quality .416 .544 .399 .531 -

Overall satisfaction .420 .557 .425 .554 .724 -

Behavioural intentions .293 .440 .308 .459 .570 .528 1.000


30

proportion of variance (0.294) in the overall service qualitythan is the case with other scales. Addition of importanceweights is not able to enhance the explanatory powerof the SERVPERF and the SERVQUAL scales. The resultsof the present study are quite in conformity with thoseof Cronin and Taylor (1992) who also found addition ofimportance weight not improving the predictive abilityof either scale.

Discriminatory Power of AlternativeMeasurement Scales

One basic use of a service quality scale is to gain insightas to where a particular service firm stands vis-à-visothers in the market. The scale that can best differentiateamong service firms obviously represents a better choice.Mean quality scores for each restaurant were computedand compared with the help of ANOVA technique todelve into the discriminatory power of alternative meas-urement scales. The results presented in Table 3 showsignificant differences (p < .000) existing among meanservice quality scores for each of the alternate scales. Theresults are quite in line with those obtained by usingsingle-item measures of service quality. The results thusestablish the ability of all the four scales to be able todiscriminate among the objects (i.e., restaurants), and assuch imply that any one of the scales can be used formaking quality comparisons across service firms.

Parsimony in Data Collection

Often, ease of data collection is a major considerationgoverning the choice of measurement scales for studiesin the business context. When examined from this per-spective, the unweighted performance-only scale turnsout to be the best choice as it requires much less infor-mational input than required by the other scales. Whilethe SERVQUAL and weighted service quality scales(both SERVQUAL and the SERVPERF) require data oncustomer perceptions as well as customer expectationsand/or importance perceptions also, the performance-only measure requires data on customers’ perceptionsalone, thus considerably obviating the data collectiontask. While the number of items for which data arerequired is only 22 for the SERVPERF scale, it is 44 and66 for the SERVQUAL and the weighted SERVQUALscales respectively (Table 4). Besides making the ques-tionnaire lengthy and compounding data editing andcoding tasks, requirement of additional data can haveits toll on the response rate too. This study is a case inpoint. Seeing a lengthy questionnaire, many respon-dents hesitated to fill it up and returned it on the spot.

Diagnostic Ability of Scales in ProvidingInsights for Managerial Interventionand Strategy Formulation

A major reason underlying the use of a multi-item scalevis-à-vis its single-item counterpart is its ability to pro-vide information about the attributes where a given firmis deficient in providing service quality and thus needsto evolve strategies to remove such quality shortfallswith a view to enhance customer satisfaction in future.When judged from this perspective, all the four servicequality scales, being multi-item scales, appear capableof performing the task. But, unfortunately, the scales

Table 2: Explanatory Power of Alternative Service Scales— Regression Results

Measurement Scale R2 Adjusted R2

(Independent Variable)

SERVQUAL (P-E) .173 .171SERVPERF (P) .296 .294Weighted SERVQUAL I(P-E) .159 .156Weighted SERVPERF I(P) .282 .280

Note: Dependent variable = Overall service quality.

Table 3: Discriminatory Power of Alternate Scales — ANOVA Results

Restaurant SERVPERF Weighted SERVQUAL Weighted Overall(P) SERVPERF I (P) (P-E) SERVQUAL I (P-E) Service Quality

Nirulas 3.63 3.67 -0.28 -0.31 4.04Wimpy’s 3.41 3.44 -0.64 -0.58 3.46Dominos 3.40 3.50 -0.45 -0.41 3.52McDonalds 3.72 3.78 -0.21 -0.20 4.23Pizza Hut 3.64 3.72 -0.24 -0.29 4.00Haldiram 3.55 3.51 -0.40 -0.50 3.72Bikanerwala 3.38 3.41 -0.57 -0.62 3.65Rameshwar 3.19 3.30 -0.58 -0.58 3.19Overall mean 3.55 3.61 -0.37 -0.38 3.86F-value (significance level) 6.60 5.31 4.25 3.40 6.77

(p <.000) (p <.000) (p <.000) (p <.002) (p<.000)


31

differ considerably in terms of the areas identified forimprovement as well as the order in which the identifiedareas need to be taken up for quality improvement. Thisasymmetrical power of the four scales can be probed intoby taking up four typical service attributes, namely, useof up-to-date equipment and technology, prompt res-ponse, accuracy of records, and convenience of operat-ing hours as being tapped in the study vide scale items1, 11, 13, and 22 respectively. The performance of arestaurant (name disguised) on these four scale itemsis reported in Table 5.

An analysis of Table 5 reveals the following find-ings. When measured with the help of ‘performance-only’ (i.e., SERVPERF) scale, scores in column 3 showthat the restaurant is providing quality in respect ofservice items 1, 13, and 22. The mean scores in the rangeof 3.31 to 3.97 for these items are a pointer to this in-ference. The consumers appear indifferent to the pro-vision of service quality in respect of item 11. However,when compared with maximum possible attainable valueof 5 on a 5-point scale, the restaurant under consider-ation seems deficient in respect of all the four serviceareas (column 5) implying managerial intervention inall these areas. In the event of time and resource con-straints, however, the management needs to prioritizequality deficient areas. This can be done in two ways:either on the basis of magnitude of performance scores

(scores lower in magnitude pointing to higher priorityfor intervention) or on the basis of magnitude of theimplied gap scores between perceived performance (P)and maximally attainable score of 5 (with higher gapsimplying immediate interventions). Judged anyway, theservice areas in the descending order of interventionurgency are 11, 22, 13, and 1 (see columns 3 and 5). Themanagement can pick up one or a few areas for man-agerial intervention depending upon the availability oftime and financial resources at its disposal. If importancescores are also taken into account as is the case with theweighted SERVPERF scale, the order of priority getschanged to 11, 13, 22, and 1.

In the case of the SERVQUAL scale requiring com-parison of customers’ perceptions of service perform-ance (P) with their expectations (E), the areas with zeroor positive gaps imply either customer satisfaction ordelight with the service provision and as such do notcall for any managerial intervention. But, in the areaswhere gaps are negative, the management needs to dosomething urgently for improving the quality. Whenviewed from this perspective, only three service areas,namely, 13, 11, and 1 having negative gaps, call formanagerial intervention and in that order as determinedby the magnitude of gap scores shown in column 9 ofTable 5. Taking into account the importance scores alsoas is the case with the weighted SERVQUAL scale, orderof priority areas gets changed to 11, 13, and 1 (see column10).

We thus find that though all the four multi-itemscales possess diagnostic power to suggest areas formanagerial actions, the four scales differ considerablyin terms of areas suggested as well as the order in whichthe actions in the identified areas are called for. The moot

Table 4: Number of Items Contained in Service QualityMeasurement Scales

Scale Number of Items

SERVQUAL (P-E) 44SERVPERF (P) 22Weighted SERVQUAL I(P-E) 66Weighted SERVPERF I(P) 44

Table 5: Areas Suggested for Quality Improvement by Alternate Service Quality Scales

Scale Item Performance Maximum Gap Importance I(P) or Expectation Gap (P-E) I(P-E) orItem Description (P) or Score (P-M) Score (I) Weighted Score or Weighted

SERVPERF SERVPERF (E) SERVQUAL SERVQUALScore Score

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

1. Use of up-to-dateequipment and technology 3.97 5.00 -1.03 4.28 16.99 4.37 -0.40 -1.71

11. Prompt response 3.08 5.00 -1.92 4.09 12.60 3.57 -0.49 -6.0113. Accuracy of records 3.51 5.00 -1.49 3.67 12.88 4.05 -0.54 -1.9822. Operating hours

convenient to all 3.31 5.00 -1.69 4.05 13.37 3.04 -0.27 -1.09

Action areas inorder of priority 11, 22, 13, 1 11, 22, 13, 1 11, 13, 22, 1 13, 11, 1 11, 13, 1

Note: Customer expectations, perceptions, and importance for each service quality item were measured on a 5-point Likert scale rangingfrom 5 for ‘strongly agree’ to 1 for ‘strongly disagree.’


32

point, therefore, is to determine which scale providesa more pragmatic and managerially useful diagnosis.From a closer perusal of the data provided in Table 4,it may be observed that the problem of different areasand different ordering suggested by the four scales iscoming up basically due to different reference pointsused explicitly or implicitly for computing the servicequality shortfalls. While it is the maximally attainablescore of 5 on a 5-point scale that presumably is servingas a reference point in the case of the SERVPERF scale,it is customer expectation for each of the service areathat is acting as a yardstick under the SERVQUAL scale.Ideally speaking, the management should strive for at-taining the maximally attainable the performance level(a score of 5 in the case of 5-point scale) in all thoseservice areas where the performance level is less than5. This is exactly what the SERVPERF scale-based anal-ysis purports to do. However, this is tenable only undersituations when there are no time and resource con-straints and it can be assumed that all the areas areequally important to customers and they want maximal-ly possible quality level in respect of each of the serviceattributes. But, in a situation where the managementworks under resource constraints (this usually is thecase) and consumers do not equally importantly wantmaximum possible service quality provision, the man-agement needs to identify areas which are more criticalfrom the consumers’ point of view and call for imme-diate attention. This is exactly what the SERVQUALscale does by pointing to areas where firm’s performanceis below the customers’ expectations.

Between the two scales, therefore, the SERVQUALscale stands to provide a more pragmatic diagnosis ofthe service quality provision than the SERVPERF scale.4

So long as perceived performance equals or exceedscustomer expectations for a service attribute, the SERV-QUAL scale does not point to managerial interventiondespite performance level in respect to that attributefalling short of the maximally attainable service qualityscore. Service area 22 is a case in point. As per theSERVPERF scale, this is also a fitting area for managerialintervention because the perceived performance level inrespect of this attribute is far less than the maximallyattainable value of 5. This, however, is not the case withthe SERVQUAL scale. Since the customer perceptionsof a restaurant’s performance are above their expecta-tion level, there seems to be no ostensible justificationin further trying to improve the performance in this area.

The customers are already getting more than their ex-pectations; any attempt to further improve the perform-ance in this area might drain the restaurant owner ofthe resources needed for improvement in other criticalareas. Any such effort, moreover, is unlikely to add tothe customers’ delight as the customers themselves arenot desirous of having more of this service attribute asrevealed by their mean expectation score which is muchlower than the ideally and maximally attainable scoreof 5.

If importance scores are also taken into consider-ation, the weighted versions of both the scales providemuch more useful insights than those provided by theunweighted counterparts. Be it the SERVQUAL or theSERVPERF scale, the inclusion of weights does representimprovement over the unweighted measures. By incor-porating the customer perceptions of the importance ofdifferent service attributes in the analysis, the weightedservice quality scales are able to more precisely directmanagerial attention to deficient areas which are morecritical from the customers’ viewpoint and as such needto be urgently attended to. It may, furthermore, beobserved that between the weighted versions of theSERVPERF and the SERVQUAL scales, the weightedSERVQUAL scale is much more superior in its diagnos-tic power. This scale takes into account not only themagnitude of customer defined service quality gaps butalso the importance weights that customers assign todifferent service attributes, thus pointing to such servicequality shortfalls as are crucial to a firm’s success in themarket and deserve immediate managerial intervention.

CONCLUSIONS, IMPLICATIONS, ANDDIRECTIONS FOR FUTURE RESEARCH

A highly contentious issue examined in this paper re-lates to the operationalization of service quality con-struct. A review of extant literature points to SERV-QUAL and SERVPERF as being the two most widelyadvocated and applied service quality scales. Notwith-standing a number of researches undertaken in the field,it is not yet clear as to which one of the two scales isa better measure of service quality. Since the focus ofthe past studies has been on an assessment of the psy-chometric and methodological soundness alone of theservice quality scales — and that too in the context ofthe developed world — this study represents a pioneer-ing effort towards evaluating the methodological sound-ness as well as the diagnostic power of the two scales


33

in the context of a developing country — India. A surveyof the consumers of the fast food restaurants in the Delhiwas carried out to gather the necessary information. Theunweighted as well as the weighted versions of theSERVQUAL and the SERVPERF scales were compara-tively assessed in terms of their convergent and discri-minant validity, ability to explain variation in the overallservice quality, ease in data collection, capacity to dis-tinguish restaurants on quality dimension, and diagnos-tic capability of providing directions for managerialinterventions in the event of service quality shortfalls.

So far as the assessment of various scales on the firstthree parameters is concerned, the unweighted perform-ance-only measure (i.e., the SERVPERF scale) emergesas a better choice. It is found capable of providing a moreconvergent and discriminant valid explanation of serv-ice quality construct. It also turns out to be the mostparsimonious measure of service quality and is capableof explaining greater proportion of variance present inthe overall service quality measured through a single-item scale.

The addition of importance weights, however, doesnot result in a higher validity and explanatory powerof the unweighted SERVQUAL and SERVPERF scales.These findings are quite in conformity with those ofearlier studies recommending the use of unweightedperception-only scores (e.g., Bolton and Drew, 1991b;Boulding et al., 1993; Churchill and Surprenant, 1982;Cronin, Brady and Hult, 2000; Cronin and Taylor, 1992).

When examined from the point of view of the powerof various scales to discriminate among the objects (i.e.,restaurants in the present case), all the four scales standat par in performing the job. But in terms of diagnosticability, it is the SERVQUAL scale that emerges as a clearwinner. The SERVPERF scale, notwithstanding its su-periority in other respects, turns out to be a poor choice.For, being based on an implied comparison with themaximally attainable scores, it suggests interventioneven in areas where the firm’s performance level isalready up to customer’s expectations. The incorpora-tion of expectation scores provides richer informationthan that provided by the perception-only scores thusadding to the diagnostic power of the service qualityscale. Even the developers of performance-only scalewere cognizant of this fact and did not suggest that itis unnecessary to measure customer expectations in serv-ice quality research (Cronin and Taylor, 1992).

From a diagnostic perspective, therefore, (P-E) scale

constitutes a better choice. Since it entails a direct com-parison of performance perceptions with customer ex-pectations, it provides a more pragmatic diagnosis ofservice quality shortfalls. Especially in the event of timeand resource constraints, the SERVQUAL scale is ableto direct managerial attention to service areas which arecritically deficient from the customers’ viewpoint andrequire immediate attention. No doubt, the SERVQUALscale entails greater data collection work, but it can beeased out by employing direct rather than computedexpectation disconfirmation measures. This can be doneby asking customers to directly report about the extentthey feel a given firm has performed in comparison totheir expectations in respect of each service attributerather than asking them to report their perception andexpectation scores separately as is required under theSERVQUAL scale (for a further discussion on this aspect,see Dabholkar, Shepherd and Thorpe, 2000).

The addition of importance weights further adds tothe diagnostic power of the SERVQUAL scale. Thoughthe inclusion of weights improves the diagnostic abilityof even the SERVPERF scale, the scale continues to sufferfrom its generic weakness of directing managerial atten-tion to such service areas which are not at all deficientin the customer’s perception.

In overall terms, we thus find that while theSERVPERF scale is a more convergent and discriminantvalid explanation of the service construct, possessesgreater power to explain variations in the overall servicequality scores, and is also a more parsimonious datacollection instrument, it is the SERVQUAL scale whichentails superior diagnostic power to pinpoint areas formanagerial intervention. The obvious managerial impli-cation emanating from the study findings is that whenone is interested simply in assessing the overall servicequality of a firm or making quality comparisons acrossservice industries, one can employ the SERVPERF scalebecause of its psychometric soundness and instrumentparsimoniousness. However, when one is interested inidentifying the areas of a firm’s service quality shortfallsfor managerial interventions, one should prefer theSERVQUAL scale because of its superior diagnosticpower.

No doubt, the use of the weighted SERVQUAL scaleis the most appropriate alternative from the point ofview of the diagnostic ability of various scales, yet a finaldecision in this respect needs to be weighed against thegigantic task of information collection. Following Cro-


34

nin and Taylor’s (1992) approach, one requires collectinginformation on importance weights for all the 22 scaleitems thus considerably increasing the length of thesurvey instrument. However, alternative approaches doexist that can be employed to overcome this problem.

One possible alternative is to collect informationabout the importance weights at the service dimensionrather than the individual service level. This can beaccomplished by first doing a pilot survey of the re-spondents using 44 SERVQUAL scale items and thenperforming a factor analysis on the collected data foridentifying service dimensions. Once the service dimen-sions are identified, a final survey of all the samplerespondents can be done for seeking information inrespect of the 44 scale items as well as for the importanceweights for each of the service quality dimensions iden-tified during the pilot survey stage. Addition of onemore question seeking importance information will onlyslightly increase the questionnaire size. The importanceinformation so gathered can then be used for prioritizingthe quality deficient service areas for managerial inter-vention. Alternatively, one can employ the approachadopted by Parasuraman, Zeithaml and Berry (1988).Instead of directly collecting information from the re-spondents, they derived importance weights by regress-ing overall quality perception scores on the SERVQUALscores for each of the dimensions identified through theuse of factor analysis on the data collected vide 44 scaleitems. Irrespective of the approach used, the data col-lection task will be much simpler than required as perthe approach employed by Cronin and Taylor (1992) forgathering data in connection with the weighted SERV-QUAL scale.

Though the study brings to the fore interestingfindings, it will not be out of place to mention here someof its limitations. A single service setting with a fewrestaurants under investigation and a small database ofonly 400 observations preclude much of the generali-zability of the study findings. Studies of similar kind

with larger sample sizes need to be replicated in differ-ent service industries in different countries — especiallyin the developing ones — to ascertain applicability andsuperiority of the alternate service quality scales.

Dimensionality, though an important considerationfrom the point of view of both the validity and reliabilityassessment, has not been investigated in this paper dueto space limitations. It is nonetheless an important issuein itself and needs to be thoroughly examined beforecoming to a final judgment about the superiority of theservice quality scales. It is quite possible that the con-clusions of the present study might change if the dimen-sionality angle is incorporated into the analysis. Studiesin future may delve into this aspect.

One final caveat relates to the limited power of boththe unweighted and the weighted versions of the SERV-QUAL and the SERVPERF scales to explain variationspresent in the overall service quality scores assessedthrough the use of a single-item scale. This casts doubtson the applicability of multi-item service quality scalesas propounded and tested in the developed countriesto the service industries in a developing country like India.Though regressing overall service quality scores onservice quality dimensions might somewhat improve theexplanatory power of these scales, we do not expect anyappreciable improvement in the results. The poor explan-atory power of the scales in the present study might havearisen either due to methodological considerations suchas the use of a smaller sample or a 5-point rather thana 7-point Likert scale employed by the developers ofservice quality scales in their studies or else — as is morelikely to be the case — the problem has arisen due tothe inappropriateness of items contained in the servicequality scales under investigation in the context of thedeveloping countries. Both these aspects need to bethoroughly examined in future researches so as to be ableto arrive at a psychometrically as well as manageriallymore useful service quality scale for use in the serviceindustries of the developing countries.

ENDNOTES1. Customer satisfaction with services or perception of

service quality can be viewed as confirmation or dis-confirmation of customer expectations of a service offer.The proponents of the gap model have based theirresearches on disconfirmation paradigm which main-tains that satisfaction is related to the size and directionof the disconfirmation experience where disconfirma-tion is related to the person’s initial expectations. For

further discussion, see Churchill and Surprenant, 1982and Parasuraman, Zeithaml and Berry, 1985.

2. A factor analysis of 22 scale items led Parasuraman,Zeithaml and Berry (1988) to conclude that consumersuse five dimensions for evaluating service quality. Thefive dimensions identified by them included tangibility,reliability, responsiveness, assurance, and empathy.

3. The scale items used in this connection were: “The


35

probability that I will use their facilities again,” “Thelikelihood that I would recommend the restaurants toa friend,” and “If I had to eat in a fast food restaurantagain, the chance that I would make the same choice.”

4. Even though a high correlation (r=0.747) existed be-

tween (P-M) and (P-E) gap scores, the former cannotbe used as a substitute for the latter as on a case by casebasis, it can point to initiating actions even in such areaswhich do not need any managerial intervention basedon (P-E) scores.

REFERENCESAndaleeb, S S and Basu, A K (1994). “Technical Complexity

and Consumer Knowledge as Moderators of ServiceQuality Evaluation in the Automobile Service Indus-try,” Journal of Retailing, 70(4), 367-81.

Anderson, E W, Fornell, C and Lehmann, D R (1994).“Customer Satisfaction, Market Share and Profitability:Findings from Sweden,” Journal of Marketing, 58(3), 53-66.

Anderson, C and Zeithaml, C P (1984). “Stage of the Pro-duct Life Cycle, Business Strategy, and Business Per-formance,” Academy of Management Journal, 27 (March),5-24.

Babakus, E and Boller, G W (1992). “An Empirical Assess-ment of the Servqual Scale,” Journal of Business Research,24(3), 253-68.

Babakus, E and Mangold, W G (1989). “Adapting the Serv-qual Scale to Hospital Services: An Empirical Investi-gation,” Health Service Research, 26(6), 767-80.

Babakus, E and Inhofe, M (1991). “The Role of Expectationsand Attribute Importance in the Measurement of Serv-ice Quality” in Gilly M C (ed.), Proceedings of the SummerEducator’s Conference, Chicago, IL: American MarketingAssociation, 142-44.

Bolton, R N and Drew, J H (1991a). “A Multistage Modelof Customer’s Assessment of Service Quality and Va-lue,” Journal of Consumer Research, 17(March), 375-85.

Bolton, R N and Drew, J H (1991b). “A Longitudinal Ana-lysis of the Impact of Service Changes on CustomerAttitudes,” Journal of Marketing, 55(January), 1-9.

Boulding, W; Kalra, A, Staelin, R and Zeithaml, V A (1993).“A Dynamic Process Model of Service Quality: FromExpectations to Behavioral Intentions,” Journal of Mar-keting Research, 30(February), 7-27.

Brady, M K and Robertson, C J (2001). “Searching for aConsensus on the Antecedent Role of Service Qualityand Satisfaction: An Exploratory Cross-National Study,”Journal of Business Research, 51(1) 53-60.

Brady, M K, Cronin, J and Brand, R R (2002). “Perfor-mance–Only Measurement of Service Quality: A Rep-lication and Extension,” Journal of Business Research,55(1), 17-31.

Brown, T J, Churchill, G A and Peter, J P (1993). “Improvingthe Measurement of Service Quality,” Journal of Retail-ing, 69(1), 127-39.

Brown, S W and Swartz, T A (1989). “A Gap Analysis ofProfessional Service Quality,” Journal of Marketing, 53(April), 92-98.

Buzzell, R D and Gale, B T (1987). The PIMS Principles, NewYork: The Free Press.

Carman, J M (1990). “Consumer Perceptions of ServiceQuality: An Assessment of the SERVQUAL Dimensions,”Journal of Retailing, 66(1), 33-35.

Churchill, G A (1979). “A Paradigm for Developing BetterMeasures of Marketing Constructs,” Journal of Market-ing Research, 16 (February), 64-73.

Churchill, G A and Surprenant, C (1982). “An Investigationinto the Determinants of Customer Satisfaction,” Journalof Marketing Research, 19(November), 491-504.

Cronin, J and Taylor, S A (1992). “Measuring Service Quality:A Reexamination and Extension,” Journal of Marketing,56(July), 55-67.

Cronin, J and Taylor, S A (1994). “SERVPERF versus SERV-QUAL: Reconciling Performance-based and Perceptions–Minus–Expectations Measurement of Service Quality,”Journal of Marketing, 58(January), 125-31.

Cronin, J, Brady, M K and Hult, T M (2000). “Assessingthe Effects of Quality, Value and Customer Satisfactionon Consumer Behavioral Intentions in Service Environ-ments,” Journal of Retailing, 76(2), 193-218.

Crosby, P B (1984). Paper presented to the “Bureau deCommerce,” Montreal, Canada (Unpublished), Novem-ber.

Dabholkar, P A, Shepherd, D C and Thorpe, D I (2000).“A Comprehensive Framework for Service Quality: AnInvestigation of Critical, Conceptual and MeasurementIssues through a Longitudinal Study,” Journal of Retail-ing, 76(2), 139-73.

Eiglier, P and Langeard, E (1987). Servuction, Le Marketingdes Services, Paris: McGraw-Hill.

Finn, D W and Lamb, C W (1991). “An Evaluation of theSERVQUAL Scale in a Retailing Setting” in Holman, Rand Solomon, M R (eds.), Advances in Consumer Research,Provo, UT: Association for Consumer Research,480-93.

Garvin, D A (1983). “Quality on the Line,” Harvard BusinessReview, 61(September-October), 65-73.

Gotlieb, J B, Grewal, D and Brown, S W (1994). “ConsumerSatisfaction and Perceived Quality: Complementary orDivergent Constructs,” Journal of Applied Psychology,79(6), 875-85.

Gronroos, C (1982). Strategic Management and Marketing inthe Service Sector. Finland: Swedish School of Economicsand Business Administration.

Gronroos, C (1990). Service Management and Marketing:Managing the Moments of Truth in Service Competition.Mass: Lexington Books.

Hartline, M D and Ferrell, O C (1996). “The Managementof Customer Contact Service Employees: An EmpiricalInvestigation,” Journal of Marketing, 69 (October), 52-70.

Iacobucci, D, Grayson, K A and Ostrom, A L (1994). “TheCalculus of Service Quality and Customer Satisfaction:Theoretical and Empirical Differentiation and Integra-tion,” in Swartz, T A; Bowen, D H and Brown, S W (eds.),Advances in Services Marketing and Management, Green-wich, CT: JAI Press,1-67.

Juran, J M (1988). Juran on Planning for Quality. New York:The Free Press.

Kassim, N M and Bojei, J (2002). “Service Quality: Gapsin the Telemarketing Industry,” Journal of Business Re-search, 55(11), 845-52.

Kotler, P (2003). Marketing Management, New Delhi: Pren-


36

tice Hall of India.Lewis, R C (1987). “The Measurement of Gaps in the Quality

of Hotel Service,” International Journal of HospitalityManagement, 6(2), 83-88.

Lewis, B (1991). “Service Quality: An International Com-parison of Bank Customer’s Expectations and Percep-tions,” Journal of Marketing Management, 7(1), 47-62.

Mazis, M B, Antola, O T and Klippel, R E (1975). “AComparison of Four Multi-Attribute Models in thePrediction of Consumer Attitudes,” Journal of ConsumerResearch, 2(June), 38-52.

Miller, J A (1977). “Exploring Satisfaction, ModifyingModes, Eliciting Expectations, Posing Problems, andMaking Meaningful Measurements,” in Hunt, K (ed.),Conceptualization and Measurement of Consumer Satisfac-tion and Dissatisfaction, Cambridge, MA: MarketingScience Institute, 72-91.

Normann, R (1984). Service Management. New York: Wiley.Parasuraman, A, Berry, L L and Zeithaml, V A (1990).

“Guidelines for Conducting Service Quality Decrease,”Marketing Research, 2(4), 34-44.

Parasuraman, A, Berry, L L and Zeithaml, V A (1991).“Refinement and Reassessment of the SERVQUALScale,” Journal of Retailing, 67(4), 420-50.

Parasuraman, A, Zeithaml, V A and Berry, L L (1985). “AConceptual Model of Service Quality and Its Implica-tions for Future Research,” Journal of Marketing, 49 (Fall),41-50.

Parasuraman, A, Zeithaml, V A and Berry, L L (1988).“SERVQUAL: A Multiple Item Scale for MeasuringConsumer Perceptions of Service Quality,” Journal ofRetailing, 64(1), 12-40.

Parasuraman, A, Zeithaml, V A and Berry, L L (1994).“Reassessment of Expectations as a Comparison Stand-ard in Measuring Service Quality: Implications forFurther Research,” Journal of Marketing, 58(January),111-24.

Peter, J P, Churchill, G A and Brown, T J (1993). “Cautionin the Use of Difference Scores in Consumer Research,”Journal of Consumer Research, 19(March), 655-62.

Phillips, L W, Chang, D R and Buzzell, R D (1983). “ProductQuality, Cost Position and Business Performance: ATest of Some Key Hypothesis,” Journal of Marketing, 47(Spring), 26-43.

Pitt, L F, Oosthuizen P and Morris, M H (1992). ServiceQuality in a High Tech Industrial Market: An Applicationof SERVQUAL, Chicago: American Management Asso-ciation.

Rust, R T and Oliver, R L (1994). Service Quality — NewDirections in Theory and Practice, New York: Sage Pub-lications.

Shaw, J (1978). The Quality - Productivity Connection, NewYork: Van Nostrand.

Smith, R A and Houston, M J (1982). “Script-based Eva-luations of Satisfaction with Services,” in Berry, L,Shostack, G and Upah, G (eds.), Emerging Perspectiveson Services Marketing, Chicago: American MarketingAssociation, 59-62.

Spreng, R A and Singh, A K (1993). “An Empirical Assess-ment of the SERVQUAL Scale and the Relationshipbetween Service Quality and Satisfaction,” in Peter, DW, Cravens, R and Dickson (eds.), Enhancing KnowledgeDevelopment in Marketing, Chicago, IL: American Mar-keting Association, 1-6.

Teas, K R (1993). “Expectations, Performance Evaluation,and Consumer’s Perceptions of Quality,” Journal ofMarketing, 57(October), 18-34.

Teas, K R (1994). “Expectations as a Comparison Standardin Measuring Service Quality: An Assessment of Reas-sessment,” Journal of Marketing, 58(January), 132-39.

Witkowski, T H and Wolfinbarger, M F (2002). “Compar-ative Service Quality: German and American Ratingsacross Service Settings,” Journal of Business Research, 55(11), 875-81.

Woodruff, R B, Cadotte, E R and Jenkins, R L (1983).“Modelling Consumer Satisfaction Processes UsingExperience-based Norms,” Journal of Marketing Research,20(August), 296-304.

Young, C, Cunningham, L and Lee, M (1994). “AssessingService Quality as an Effective Management Tool: TheCase of the Airline Industry,” Journal of Marketing Theoryand Practice, 2(Spring), 76-96.

Zeithaml, V A, Parasuraman, A and Berry, L L (1990).Delivering Service Quality: Balancing Customer Percep-tions and Expectations, New York: The Free Press.

Zeithaml, V A and Parasuraman, A (1991). “The Natureand Determinants of Customer Expectation of Service,”Marketing Science Institute Working Paper No. 91-113,Cambridge, MA: Marketing Science Institute.

Zeithaml, V A and Parasuraman, A (1996). “The BehavioralConsequences of Service Quality,” Journal of Marketing,60(April), 31-46.

Zeithaml, V A and Bitner, M J (2001). Services Marketing:Integrating Customer Focus Across the Firms, 2nd Edition,Boston: Tata-McGraw Hill.

Sanjay K Jain is Professor of Marketing and International Businessin the Department of Commerce, Delhi School of Economics,University of Delhi, Delhi. His areas of teaching and researchinclude marketing, services marketing, international marketing,and marketing research. He is the author of the book titled ExportMarketing Strategies and Performance: A Study of Indian Textilespublished in two volumes. He has published more than 70 researchpapers in reputed journals including Journal of Global Marketing,Malaysian Journal of Small and Medium Enterprises, Vikalpa,

Business Analyst, etc. and also presented papers at various nationaland international conferences.e-mail: [email protected]

Garima Gupta is a Lecturer of Commerce in Kamla Nehru Col-lege, University of Delhi, Delhi. She is currently pursuing herdoctoral study in the Department of Commerce, Delhi School ofEconomics, University of Delhi, Delhi.e-mail: [email protected]


37

servqual servperf

Documents

Transcript of servqual servperf