further consulted and adapted some standardized items provided 8 by Parasuraman

Further consulted and adapted some standardized items

This preview shows page 7 - 10 out of 20 pages.

further consulted and adapted some standardized items provided
Image of page 7
- 8 - by Parasuraman et al.(1988) into the survey. The scale items for assessing behavioral consequences of overall satisfaction were adapted from validated measures emanating from prior studies (Bloemer et al., 1999; Zeithaml et al., 1996). The respondents were requested to indicate the extent to which they agree or disagree, based on recent online transaction experiences, by checking the appropriate response to the questionnaire items regarding the key constructs of the study. For each item, except that of overall satisfaction, five-point Likert scales anchored by 1=strongly disagree, and 5=strongly agree, with 3=netural (neither agree nor disagree) as the midpoint, were utilized. In turn, the overall satisfaction measurement items were gauged by seven-point Likert scales. Five academic experts in relationship marketing were asked to assess the face or content validity of the measurement scales. Subsequently, the questionnaire was forwarded, by e-mail attachment, to 50 online customers selected from two news groups: online financial investment and e-commerce. A total of 14 respondents replied with useful suggestions. Based on their feedback, the questionnaire was further revised and finalized. 2. Data Collection A Web-based survey was employed in the study. A solicitation letter was transmitted by an e-mail to 4,000 subjects randomly selected from an e-mailing list provided by an e-mail broker 1) . The e-mail message described the purpose of the research and invited each receiver to participate in the survey. Sample members who evidenced a willingness to participate were required to click through the URL address provided in the invitation e-mail. A total of 1,101 e-mails were returned as undeliverable. Thus, the actual undeliverable rate was 27.5% (1101 of 4000), which is similar to the experience of Sheehan and Hoy (2000). Their undeliverable rate was 26%. The responses from 257 participants were forwarded to the leading author via e-mail. Of course, 22 were eliminated because they were incomplete or duplicated (the ISP address of each respondent was checked). Thus, the effective sample size was 235 and the final response rate was 8.1%(235 of 2899). Some complaints regarding the sampling effort were forwarded to the researchers' employer via a third party. It was not possible to find and delete their e-mail addresses. Since further e-mailing without deleting those addresses in question could have evoked potential legal issues and the number of collected useable responses was sufficient for further data analysis, no follow-up e-mails were sent. A comparison was not made between early and late responses for checking non-response bias, since appropriately 90% of the responses were gathered within five days after the initial e-mail. 1) e-mail list manager in Tsinghua university among students in Tsinghua university and alumni.
Image of page 8
- 9 - Construct Mean S.D. 1 2 3 4 5 6 7 8 9 10 1. Fulfillment/Reliability 4.27 0.71 1.00 2. Customer Services 3.56 0.74 0.50 1.00 3. Ease of Use 3.63 0.74 0.57 0.53 1.00 4. Service Portfolio 3.76 0.70 0.36 0.38 0.53 1.00 5. Security 3.80 0.71 0.60 0.51 0.57 0.37 1.00 6. Overall Satisfaction 5.65 1.06 0.63 0.59 0.62 0.51 0.56 1.00 7. Recommendations
Image of page 9
Image of page 10

You've reached the end of your free preview.

Want to read all 20 pages?

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture