E-S-QUAL: A Multiple-Item Scale for Assessing Electronic Service Quality
Summary (5 min read)
TRADITIONAL SERVICE QUALITY VERSUS ELECTRONIC SERVICE QUALITY
- Extensive research on traditional SQ has been conducted during the past 20 years (see Parasuraman and Zeithaml 2002 for a review).
- In contrast, only a limited number of scholarly articles deal directly with how customers assess e-SQ and its antecedents and consequences.
- The authors briefly overview the relevant aspects of traditional SQ and describe the reasons why that research needs to be repeated in the electronic context.
Traditional Service Quality
- By traditional SQ the authors are referring to the quality of all non-Internet-based customer interactions and experiences with companies.
- Using insights from these studies as a starting point, Parasuraman, Zeithaml, and Berry (1988, 1991) conducted empirical studies in several industry sectors to develop and refine SERVQUAL, a multiple-item instrument to quantify customers’ global (as opposed to transaction-specific) assessment of a company’s SQ.
- This scale measures SQ along five dimensions: reliability, responsiveness, assurance, empathy, and tangibles.
- As such, whether the preceding conclusions extend to e-SQ contexts and what the similarities and differences are between the evaluative processes for SQ and e-SQ are open questions.
- The items in the scale were changed to adapt to the electronic context (e.g., tangibles were represented in part by an item about appearance of the Web site), and therefore the scales were not comparable across the research contexts.
Why e-SQ?
- Insights from studies dealing with people-technology interactions imply that customer evaluation of new technologies is a distinct process.
- Another major qualitative study by the same authors (Mick and Fournier 1998), focusing on people’s reactions to technology, suggests that technology may trigger positive and negative feelings simultaneously.
- Earlier studies focusing on specific technologies have also shown that consumers’ beliefs about, and reactions to, the technology in question are distinct and positively correlated with acceptance (Cowles 1989; Cowles and Crosby 1990; Dabholkar 1996; Eastlick 1996).
- Other research shows that perceived usefulness and ease of use are correlated significantly with self-reported (Davis 1989) and actual (Szajna 1996) usage of technology.
- In other words, customer-specific attributes (e.g., technology readiness) might influence, for instance, the attributes that customers desire in an ideal Web site and the performance levels that would signal superior e-SQ.
Research on e-SQ
- Some academic researchers have developed scales to evaluate Web sites.
- This scale’s primary purpose is to generate information for Web site designers rather than to measure service quality as experienced by customers.
- Therefore, although some WebQual dimensions might influence perceived service quality, other dimensions (e.g., innovativeness, business processes, and substitutability) are at best tangential to it.
- Using an online survey, Szymanski and Hise (2000) studied the role that customer perceptions of online convenience, merchandising (product offerings and product information), site design, and financial security play in esatisfaction assessments.
- A number of studies have examined various aspects of these criteria.
Definition and Domain of e-SQ
- The extant literature and extensive focus group research in Zeithaml, Parasuraman, and Malhotra’s (2000) study suggested that customers’assessment of a Web site’s quality includes not only experiences during their interactions with the site but also postinteraction service aspects (i.e., fulfillment, returns).
- To represent the full range of evaluative criteria emerging from their focus groups, the researchers proposed a theoretical framework that is anchored in the means-end-chain approach to understanding consumers’ cognitive structures.
- Fourth, the linkages implied in the theoretical framework (Figure 2) between the e-SQ evaluative process (i.e., perceptual/dimension-level assessments) and its consequences (i.e., higher-order abstractions) constitute a natural “nomological net” for verifying the construct validity of an e-SQ scale consisting of perceptual-attribute level items.
- Zeithaml, Parasuraman, and Malhotra’s (2000) study identified dozens of Web site features at the perceptualattribute level and categorized them into 11 e-SQ dimensions: 1. Reliability: Correct technical functioning of the site and the accuracy of service promises (having items in stock, delivering what is ordered, delivering when promised), billing, and product information.
- Site is simple to use, structured properly, and requires a minimum of information to be input by the customer, also known as Efficiency.
Preliminary Scale
- A set of 121 items representing all facets of the e-SQ domain formed their initial scale.
- The authors incorporated this scale into two questionnaire versions with different scale anchors and formats.
- The authors then evaluated the alternative versions in two focus group interviews with graduate students at a major university in the eastern United States.
- The specific goals of the focus groups were to (a) understand respondents’ reactions to alternative ways of phrasing scale items and anchors (Likert-type scale versus low-high performance anchors); (b) reword items to improve clarity; (c) eliminate redundant items; and (d) obtain feedback on the length, format, and clarity of the instructions and initial questionnaire draft.
- The revised questionnaire had 113 items with 5-point scales ranging from 1 (strongly disagree) to 5 (strongly agree).
Sample Design and Data Collection
- The authors selected two online stores—amazon.com and walmart.com—to verify the psychometric properties of the E-S-QUAL and E-RecS-QUAL Scales.
- Table 3 contains descriptive profiles of the two samples.
- The third part contained multiple-item measures of two constructs—perceived value and loyalty intentions—that were used subsequently in assessing the scales’ validity.
Data Analysis and Scale Reduction
- The survey data, pooled across all sites, were subjected to various scale-reduction/refinement analyses consistent with standard procedures for developing and refining scales.
- The authors then went through a series of iterations, each involving elimination of items with low loadings on all factors or high cross-loadings on two or more factors, followed by factor analysis of the remaining items.
- The authors next analyzed the items they had set aside earlier for constructing a scale for measuring the quality of recovery service provided by Web sites.
- The degree to which the site compensates customers for problems.
- As in the case of E-S-QUAL, the authors conducted CFA analysis to verify the factor structure of the E-RecS-QUAL Scale.
Reliability and Validity Assessment
- These values together with the strong loadings of the scale items on their corresponding factors (in both EFA and CFA) support the convergent validity of each scale’s component dimensions.
- Estimating such a second-order measurement model requires at least two other reflective indicators for the second-order construct, in addition to the formative indicators already in the model (Diamantopoulos and Winklhofer 2001; Jarvis, Mackenzie, and Podsakoff 2003).
- The CFA loadings and fit statistics from this reanalysis were very similar to those reported in Table 1 for The interfactor correlations between pairs of dimensions in the CFA ranged from .67 to .83 for the E-S-QUAL dimensions and .68 to .73 for the E-RecS-QUAL dimensions.
- The correlations of the E-SQUAL dimensions with the overall measures ranged from .47 to .60 for quality and .44 to .54 for value.
Reassessment of Reliability and Validity
- Table 4 presents coefficient alpha values for, and intercorrelations among, the four E-S-QUAL dimensions, the three E-RecS-QUAL dimensions, and the measures of perceived value and loyalty intentions.
- The overall goodness-of-fit statistics and results from the structural-model portion of Table 6 (i.e., the structural coefficients and R2 values) imply that the data from each of the two samples fit the proposed model reasonably well— all fit indices are well above conventional cutoff values (Hu and Bentler 1999), although the RMSEA values are somewhat high.
- A related issue is statistical power, which depends on sample size as well as the degrees of freedom in the structural model and can be either too low (leading to nonrejection of incorrect models) or too high (leading to rejection of correct models) (McQuitty 2004).
- Given the aforementioned results, it is reasonable to interpret each factor-score measure as predominantly representing the corresponding E-S-QUAL dimension.
DISCUSSION
- Informed by insights from the extant literature and using the means-end framework as a theoretical foundation, the authors set out to conceptualize, construct, refine, and test a multiple-item scale (E-S-QUAL) for measuring the service quality delivered by Web sites.
- In the preliminary stage of scale development, a large number of respondents did not provide ratings on a subset of the initial pool of items.
- An examination of these items revealed that they all pertained to nonroutine or recovery service encounters that many respondents apparently had not experienced.
- Therefore, in subsequent stages of scale development and refinement, the authors created a subscale of E-S-QUAL—called E-RecS-QUAL—containing items focusing on handling service problems and inquiries, and being salient only to customers who had had nonroutine encounters with the sites.
- The authors hope that the two scales will stimulate and facilitate additional scholarly research on e-SQ and also assist practitioners in systematically assessing and improving e-SQ.
Directions for Further Research
- Both scales demonstrate good psychometric properties based on findings from a variety of reliability and validity tests.
- Of the four E-S-QUAL dimensions, customers’ assessments of a Web site on these two dimensions have the strongest influence not only on overall quality perceptions but also on perceived value and loyalty intentions.
- Fifth, E-S-QUAL and E-RecS-QUAL are generic and parsimonious scales, intended for obtaining a global (as opposed to transaction-specific) assessment of a Web site’s service quality.
- Companies can also enhance the diagnostic value of the perceptual ratings from the two scales by comparing those ratings with customers’ minimum- and desired-service levels (Parasuraman, Zeithaml, and Berry 1994a).
E-S-QUAL
- Respondents rated the Web site’s performance on each scale item using a 5-point scale (1 = strongly disagree, 5 = strongly agree).
- The items below are grouped by dimension for expositional convenience; they appeared in random order on the survey.
- The symbols preceding the items correspond to the variable names in Tables 1 and 5 in the body of the article.
Efficiency
- This site makes it easy to find what I need.
- Information at this site is well organized.
System Availability
- This site is always available for business.
- Pages at this site do not freeze after I enter my order information.
Fulfillment
- This site makes items available for delivery within a suitable time frame.
- It has in stock the items the company claims to have.
- It makes accurate promises about delivery of products.
Privacy
- It protects information about my Web-shopping behavior.
- It does not share my personal information with other sites.
- This site protects information about my credit card.
E-RecS-QUAL
- Respondents rated the Web site’s performance on each scale item using a 5-point scale (1 = strongly disagree, 5 = strongly agree).
- The items below are grouped by dimension for expositional convenience; they appeared in random order on the survey.
- The symbols preceding the items correspond to the variable names in Table 2 in the body of the article.
Compensation
- This site compensates me for problems it creates.
- It compensates me when what I ordered doesn’t arrive on time.
Contact
- This site provides a telephone number to reach the company.
- This site has customer service representatives \available online.
- The value measure consisted of four items; respondents rated the Web site on each item using a scale of 1 (poor) to 10 .
- The overall value you get from this site for your money and effort.
Did you find this useful? Give us your feedback
Citations
2,514 citations
Cites background from "E-S-QUAL: A Multiple-Item Scale for..."
...Social media’s effects on sales and its interactionswith attitudes and firm-owned touch points have also been examined (e.g., Onishi and Manchanda 2012; Pauwels, Aksehirli, and Lackman 2016)....
[...]
2,337 citations
Cites background from "E-S-QUAL: A Multiple-Item Scale for..."
...…how customers assess he quality of service delivered through websites – “e-service uality” – and developing scales to measure that construct e.g., Parasuraman, Zeithaml, and Malhotra 2005; Wolfinbarger nd Gilly 2003; Zeithaml, Parasuraman, and Malhotra 2002). hough findings from previous…...
[...]
...…xperience that are worth investigating. uture research Insights from extant research suggest that it is best for etailers to offer a blend of employee-based and SST-based serice options (e.g., Parasuraman, Zeithaml, and Malhotra 2005; einders, Dabholkar, and Frambach 2008; Selnes and Hansen 001)....
[...]
...Likewise he type of service – e.g., routine versus recovery service – may nfluence the appropriate blend of employee- and technologyased service (Parasuraman, Zeithaml, and Malhotra 2005)....
[...]
1,811 citations
1,164 citations
1,006 citations
References
76,383 citations
40,975 citations
40,720 citations
37,235 citations
25,611 citations
Related Papers (5)
Frequently Asked Questions (13)
Q2. What future works have the authors mentioned in the paper "A multiple-item scale for assessing electronic service quality" ?
As the discussion in the preceding section illustrates, there is a need for further research to deepen their understanding of the assessment, antecedents, and consequences of e-SQ. The four perceptual attributes that constitute system availability suggest that companies may not have full control over performance on this dimension ; the equipment at the customer ’ s end ( e. g., type of computer and Internet connection ) is also likely to affect performance on this dimension. Companies should be ( a ) sensitive to potential deleterious effects of sophisticated Web site design features on system availability and ( b ) proactive in identifying aspects of system availability that are beyond their control and devising appropriate communication scripts to appease complaining customers. Online companies can best use the scales in tandem ( with the latter being administered only to customers who have had problems or questions ) to track over time—and across competing Web sites—customers ’ overall e-SQ perceptions.
Q3. What incentive was used to get participants to complete the surveys?
As an incentive for participation, respondents completing the surveys were entered into a random drawing to receive one of several cash prizes.
Q4. Why was the pooling of data appropriate at this scale refinement/reduction stage?
The pooling of data was appropriate at this scale refinement/reduction stage because the purpose was to produce a general scale that would be appropriate for assessing service quality of a variety of sites.
Q5. Why did the authors set aside these items for analysis?
Because recovery was an important aspect of service, the authors set aside these items for separate analysis to develop an e-recovery service scale.
Q6. What did they use to develop a scale for measuring site quality?
Yoo and Donthu (2001) developed a nine-item SITEQUAL scale for measuring site quality on four dimensions: ease of use, aesthetic design, processing speed, and security.
Q7. What scale is relevant for a Web site’s entire customer base?
The basic E-S-QUAL Scale (relevant for a Web site’s entire customer base) is a four-dimensional, 22-item scale, whereas E-RecS-QUAL (relevant for the portion of the customer base with recovery service experience) is a three-dimensional, 11-item scale.
Q8. What are the three dimensions of the e-SQ scale?
Parasuraman, and Malhotra’s (2000) study identified dozens of Web site features at the perceptualattribute level and categorized them into 11 e-SQ dimensions:1. Reliability: Correct technical functioning of the site and the accuracy of service promises (having items in stock, delivering what is ordered, delivering when promised), billing, and product information.
Q9. What is the reason for the nonsignificant effects of system availability and privacy?
multicollinearity due to the strong correlations among the summed-score measures of the four E-S-QUAL dimensions seemed to be a plausible explanation for the nonsignificant effects of system availability and privacy.
Q10. What are the five broad sets of criteria that are relevant to e-SQ perceptions?
On the basis of a comprehensive review and synthesis of the extant literature on e-SQ, Zeithaml, Parasuraman, and Malhotra (2002) detailed five broad sets of criteria as relevant to eSQ perceptions: (a) information availability and content, (b) ease of use or usability, (c) privacy/security, (d) graphic style, and (e) reliability/fulfillment.
Q11. What did they use to develop a scale called eTailQ?
Wolfinbarger and Gilly (2003) used online and offline focus groups, a sorting task, and an online-customer-panel survey to develop a 14-item scale called eTailQ.
Q12. What is the strongest predictor of customer satisfaction and quality?
In fact, Wolfinbarger and Gilly (2003) found that reliability/ fulfillment ratings were the strongest predictor of customer satisfaction and quality, and the second strongest predictor of intentions to repurchase at a site.
Q13. How many observations were needed to verify the dimensionality of the refined scales?
To have a sufficient number of observations for verifying the factor structures and dimensionality of the refined scales through confirmatory factor analysis, the authors set a minimum sample size of 200 respondents for each site.