AWERProcedia Information Technology & Computer Science

AWERProcedia Information Technology & Computer Science 1 (2011) 1569-1574

2nd World Conference on Information Technology (WCIT-2011)

E-service websites quality measurement through a revised E-SQual Vahab.VahdatZada a*, Samira.Abbasib, Fateme.Barazeshc, ReisFarshid.Abdid a

Islamic Azad University, South Tehran Branch, IndustrialEngineering department,Tehran, Iran

Abstract The growth of online business emerged variety of studies about customer’s satisfaction, trust, acceptance, technology readiness and etc. Most of scales do not provide a comprehensive assessment of web site’s service, though different researchers have proposed scales to measure websites’ attributes and efficiencies. This paper outlines a study that investigates and rescales the dimensions of E-S-QUAL model, proposed by Parasuraman, Zeithaml and Malhotra (2005), for websites which are merely offering E-service instead of selling goods. The other objective was to evaluate a university web site and its’ student satisfaction. For this purpose, data were collected through Iran Azad university web site. Brief modifications were done to E-S-QUAL model. Also, validity and reliability tests approved that the scale fitted to e-service. Keywords: E-service Quality, E-S-QUAL, Web Site Quality Management; Selection and peer review under responsibility of Prof. Dr. Hafize Keser. ©2012 Academic World Education & Research Center. All rights reserved.

1. Introduction Internet retains a crucial method to sell variety of goods and services, although no longer is the revolution previously conceived *1+. Companies’ selling behavior through the internet are like Amazon.com which distributes products and services solely through Web channels, while others, like traditional organizations, are investing a huge amount of money and effort in information systems to provide online services. Another category arose like E-government and university web sites which are merely e-service based web sites. They do not sale any services or products and are made to spurt customers’ satisfaction by more available and accessible services .The rest of this article consists of these parts. In the first part, we will overview main models on eservice quality and will discuss about E-S-Qual. Next part begins with rescaling E-S-QUAL in order to fit it to the eservice sites. We will discuss about data collection and explore results and suggestions to web site developers. At

*ADDRESS FOR CORRESPONDENCE: Vahab.VahdatZada, Islamic Azad University, South Tehran Branch, IndustrialEngineering department,Tehran, Iran Tel.: +98-21-88686635 E-mail address: [email protected], [email protected], [email protected]

Vahab.Vahdat Zada et all. / AWERProcedia Information Technology & Computer Science (2012) 1569-1574

last, future research, limitation and implication will be reviewed. 1.1. E-ServiceQuality Background A lot of researches, to date, have been conducted on service quality dimensions and customer satisfactions [2]. Some studies of e-service quality employed traditional instruments such as SERVQUAL survey. For instance, researchers used SERVQUAL to examine the perceived service quality factors of Websites usage, despite the fact that it was not designed to measure service quality in an online environment and its applicability is therefore unlikely to extend to that context [3].One proposed scale to evaluate websites is WebQual describes websites performance evaluation in 12 dimensions that are: Information fit to task, interaction, trust, response time, design, intuitiveness, visual appeal, innovativeness, flow-emotional appeal, integrated communication, business process and substitutability [4]. A weakness referred to this model is that the first purpose of this research was generating information for designers of websites rather than measuring service quality or customer satisfaction. In addition Parasuraman et al believes that, the scale developers miss to include a dimension for customer service [1]. Yoo and Donthu (2001) have introduced a nine item scale, SITEQUAL, assessing the perceived quality of an online shop [5]. This nine-item scale led to four dimensions. SITEQUAL scale excludes dimensions concentrated on evaluation of website service quality. Some researchers believe that, SITEQUAL (similar to WebQual) does not organize a universal and detailed evaluation of a site’s service quality. Lots of studies and models are associated with online shopping web sites, such as Online retailing services by Wolfinbarger and Gilly (2003) named ECOMQ/ETAILQ with four dimensions: Web site design, reliability, Privacy/security, customer service [6]. Santos (2003) made a model for e-service with dimensions as followed: Ease of use, appearance, linkage, structure, content, efficiency, reliability, communication, security, incentive and customer support [7]. In addition, Yang, Zhou, and Zhou (2004) proposed a model for web portals with dimensions below: usability, usefulness, adequacy of information, accessibility and interaction. Yang model sounds to be comprehensive and have been used widely [8]. Recently many of these concerns have been addressed by the original authors of the SERVQUAL instrument and it led to development of a multi-item scale for examining website service quality [1].The proposed scale, named E-S-QUAL, has four-dimensional, 22-item scale to investigate most critical dimensions of e-service quality outlined in the literature. Along with E-S-QUAL, a subscale called E-RecS-QUAL is introduced for handling service problems especially for customers who have nonroutine recovery service with the website. E-RecS-QUAL includes three-dimensional, 11 item scale. The concerning purpose of both dimensions is measurement of website service quality, and have been tested through variety of reliability and validity tests and show well psychometric properties. Seven dimensions of E-SQUAL/E-RecS-QUAL that form core and recover services scale are: Core e-SQ: 1. Efficiency 2.System availability 3.Fulfilment 4.Privacy Online shopping sites Recovery e-SQ: 1. Responsiveness 2.Compensation 3.Contact Kim et al (2006) believe that opposed to other studies which used samples of students, E-S-QUAL dimensions were developed from respondents who are qualified with sufficient online shopping experience [9]. In addition, Sun and Wang (2009) believe that Parasuraman et al, provided the most comprehensive work on e-service quality so far [10]. Boshoff (2007) subject the E-S-QUAL scale to a psychometric assessment. The results of his research revealed that the scale is effective in capturing the essence of electronic service quality [11]. He also suggested that both scholars and practitioners must assess the underlying factor structure of their data before drawing any conclusions from their study. Jun Z et al (2009), E-G-S-QUAL and E-G-RecS-QUAL has been developed on the basis of E-SQUAL and E-RecS-QUAL by taking the service characteristics of e-gov web sites into account [12].Since there are no data collection for proposed scales, nor any reliability and validity tests, there is no evidence of good psychometric properties for E-G-S-QUAL. Nevertheless rescaling E-S-QUAL to evaluate e-gov is a beginning step to measure web sites which are merely service based, without monetary benefits. Despite the power of E-S-QUAL in measuring e-service quality, it is imperfect on measuring service based web site which their purpose is not selling products nor goods, but service [12].

1570


1.2. Methodology We rescaled E-S-QUAL at the first step. Dimensions are used without any changes, while some questions are adjusted to fit service based web sites better. In the next step, reliability tests and correlations are calculated. Upon results of this step, some questions were dropped or changed. And at last, a university web site quality is measured with the designed questionnaire. As mentioned, at the very first step, we are going to discuss about E-S-QUAL dimensions, evaluate them and modify necessary changes. The first dimension of E-S-QUAL is efficiency that defines “the ease and speed of accessing and using the site” *1+. It refers to ease of use, well organized information and simplicity of a web site. Santos (2003) believes it is vital to have a well-organized, well-structured site with concise and understandable content, terms, and conditions *7+. Since these conditions, without considering web site’s purpose, are crucial, efficiency dimension is used without any changes. The fulfillment dimension in E-S-QUAL is explained as “the scope to the site’s promises for delivering of an order and availability of an item” *1+. Other researchers also set fulfillment as a vital factor in measuring online service quality [6]. Since fulfillment dimension in E-S-QUAL is merely about truthful offering, accurate ordering and delivering products, this dimension needs revision. Revised questions are quite the same to E-G-S-QUAL dimension Jun Z et al (2009) because e-gov web sites and e-service based web sites have similar procedures in fulfillment *12+. System Availability’s endeavor is to point out the availability of websites in any conditions. This dimension discusses the software applicability. In fact System Availability search for software flaws of a website such as crash or freeze. Even if a website has rigorous, well organized, well-structured user interface and contents, it won’t satisfy users if it crashes. The last dimension in ES-QUAL model is the privacy which is defined “the degree that the site is safe and protects customer information (Parasuraman et al., 2005, p. 220)” *1+. The concerning about improper use of personal information is still a major anxiety for internet users. In e-service based web sites, it is not meaningful to share web-shopping behavior, but personal information. This dimension is adjusted for more validation. But unlike E-G-S-QUAL because many e-service web sites has a payment section, which sometimes enable a particular service, the question about credit card remain unchanged. In addition, Parasuraman et al. (2005) defined E-RecS-QUAL with three dimensions named: Responsiveness, compensation and contact. The responsiveness dimension introduced as “effective handling of problems and returns through the site” *1+. As declared, responsiveness discuss about meaningful guarantee of a web site and a mechanism for returning goods and products. In e-service based website Responsiveness refers to meaningful explanations, response and handling if a service is not in touch or available in time. Other dimensions of E-RecSQUAL are defined and reviewed, where “Compensation is related to refunding, shipping and handling costs when a problem occurs” and contacts refers to “the availability of assistance through telephone or online representatives” *1+. Although compensation faints in e-service web sites, but contact is an essential factor. In revisions, contact remains with less changes where compensation shrinks.

1.3. Methodology Results In order to assess the proposed scale, at first we developed an initial exploratory reliability analysis. We valued the 32 items of proposed questionnaire according to the Cronbach’s alpha indicator, considering a minimum value of 0.7 [13]. The item-total correlation was used to improve the levels of Cronbach’s alpha, considering a minimum value of 0.3 [14]. In favor of measuring E-SQ, Industrial engineering portal of Azad University, Tehran, Iran was chosen. It consists of various sections, for instance, student’s pages, university news, notifications, Students’ enrollment, accounting, loan and fund, contacts and etc. Among 2500 students of the industrial engineering faculty,93 questionnaires were distributed due to 95% confidence level and 10% margin of error. The questionnaires’ aid is 1571


to test reliability of the scale and measure the quality of the defined university website. SPSS 13.0 is used for all tests and validations analysis. The descriptive information of the selected population was 71% of respondents were Male and 29 percent were Female,11.8 % were less than 20 years and 77.4% were 20-25 years old and 10.8% were between 25 to 35 years old. 81.5% were College students and 18.5% were attending Graduate school. 45.7%visit web site 3 times or less in month and 44.6%visit web site 4 to 12 times per month and 9.8% visit web site 13 times or more per month. Results of item-total correlation and Cronbach’s alpha indicate that two items had to be eliminated. FULL1 “It offers service when needed information and materials are submitted.” Due to low coefficient with other questions in Fulfillment dimension and SYS4 “Pages at this site do not freeze after I enter my order information” considering item-total correlation less than 0.3. Finally, the scale was composed by 30 items, with levels of item-total correlations near or higher than 0.3 The Cronbach’s alpha is calculated for each item separately, representing higher proportion than 70%. The overall Cronbach’s alpha has amount of 0.875 which shown in table1.

Table 1. Reliability Test For Proposed Model Cronbach's Alpha 875

Cronbach's Alpha Based on Standardized Items

N of Items 30

.876

As it Shown in table 2, Pearson Correlation among all elements of a dimension is calculated to robust the model. Upon the results, elements in all dimensions are fully correlated to each other. In next step, correlation among dimensions in a particular questionnaire is calculated. Dimensions related to E-S-QUAL proposed model have significance correlations, where also dimensions in E-RecS-QUAL have meaningful correlation too. Table 2.Pearson Correlation of Revised E-S-QUAL Dimension EFF EFF

SYS

FULL

PRI

SYS

FULL

PRI

Pearson Correlation 1 .482(**) .512(**) .450(**) Sig. (2-tailed) .000 .000 .000 N 93 93 93 93 Pearson Correlation .482(**) 1 .394(**) .915(**) Sig. (2-tailed) .000 .000 .000 N 93 93 93 93 Pearson Correlation .512(**) .394(**) 1 .445(**) Sig. (2-tailed) .000 .000 .000 N 93 93 93 93 Pearson Correlation .450(**) .915(**) .445(**) 1 Sig. (2-tailed) .000 .000 .000 N 93 93 93 93 ** Correlation is significant at the 0.01 level (2-tailed).

Finally, correlation between two questionnaires is measured. Results demonstrate meaningful correlation. In conclusion, Results of statistical tests testify that the model is reliable and valid to use and all sections and qualities related to every e-service web site are tested. As mentioned, the other purpose of this research was measuring web site service quality of given university. To achieve this objective, through all dimensions of E-SQual and E-Rec-S-Qual Hypothesis were defined and tested through one sample T Test. We assume students’ satisfaction level in different ranges. Our tests are based on good satisfaction level meaning three forth of best possible answer. For the sake of analyzing the data Hypothesis are defined:H0: µ ≥3/4best possible answer 1572


(Students’ satisfaction is in appropriate level), H1: µ < 3/4 best possible answer (Students’ satisfaction is in appropriate level). Table 3 express one-sample T Test based on discussed Hypothesis. The results indicate that there is a statistically significant difference between the mean of Efficiency, Fulfillment, System availability and privacy and proper students’ satisfaction level. It is clear that the website needs many accurate modifications on these dimensions. Other dimensions depict a good result and H0 is accepted.

Table 3.One-Sample T Test, satisfaction level of given university website

EFF FULL SYS PRI RES COM CON

95% Confidence Interval of the Difference Lower Upper 5.30108 1.271 ∞ 2.76344 1.271 ∞ 1.02688 1.271 ∞ 17.28495 1.271 ∞ 0.52151 1.271 ∞ -0.02151 1.271 ∞ 0.38172 1.271 ∞ *H0 is rejected due to p-value