Since the proposed framework was derived from the literature, and the aim of the empirical research was to test this framework, it was important to verify the reliability and validity of the measures used in the research to draw valid inferences from the research leading to theory building. Reliability deals with how consistently similar measures produce similar results (Rosenthal and Rosnow, 1984 recited in Hoang Trong and Chu Nguyen Mong Ngoc (2008), whereas validity of a measurement instrument refers to how well it captures what it is designed to measure (Rosenthal and Rosnow, 1984 recited in Hoang Trong and Chu Nguyen Mong Ngoc (2008).
TABLE3.1: List of indicators to measure e-readiness Code Indicator
Technology
T1 Percentage of computer connected to the Internet T2 Disconnection and speed of the Internet
T3 Business software installed T4 Percentage of data digitalized T5 Types of network using T6 Website update
T7 Advertising
T8 Search for market information Management M1 Management has IT policy
M2 Management planning to implement e-commerce M3 Management awareness on benefit of network M4 Management vision on e-commerce
M5 Management vision on ICT equipment M6 Management support on e-commerce M7 Frequency of Hardware upgrade M8 Frequency of Software upgrade
People
Pe1 Availability of professional labors Pe2 Employees’ ICT literacy
Pe3 Staff Experience with web-based application Pe4 Organizational culture support e-commerce Pe5 Human Resource support e-commerce
Process Pr1 Level of process analysis
Pr2 Level of organization adapt to change Pr3 Flexibility encouragement at work
Pr4 Identification of bottleneck in business process Pr5 Identification of inefficient in business process Pr6 Business process flexibility to accommodate with
TABLE3.2: List of indicators to assess factors affecting e-commerce Code Indicator
Perceived Organizational E-Readiness B1 Believe website promote firm
B2 Believe network help doing business effectively
B3 Believe investment in computer application is effective B4 Believe e-commerce helps cut cost
B5 Believe e-commerce helps expand market share
B6 Believe ICT staff contribute effectively to implementing e-commerce B7 Believe management interest in e-commerce
B8 Believe organization has strong relationship with customers B9 Believe power supply effect on implementing e-commerce B10 Believe Quality of Network effect on implementing e-commerce B11 Believe Internet cost effect on implementing e-commerce
B12 Believe software cost effect on implementing e-commerce B13 Believe Organization understanding e-commerce model
B14 Organization has necessary technical and human resource for e-commerce Perceived Environmental E-Readiness
B15 Believe Customers ready for e-commerce B16 Believe Partners ready for e-commerce
B17 Local technology infrastructure ready for e-commerce B18 Law environment convenient for e-commerce
B19 Government encourage e-commerce
B20 Government support ICT and e-commerce implementation
Reliability has two dimensions referred to as repeatability and internal consistency (Zigmund,1995). Internal consistency refers to the ability of a scale item to correlate with other items in the scale that are intended to measure the same construct. Items measuring the same construct are expected to be positively correlated with each other.
A common measure of the internal consistency of a measurement instrument in so- cial sciences research is Cronbach’s alpha (Zmud and Boynton, 1991 recited in Hoang Trong and Chu Nguyen Mong Ngoc (2008).
Cronbach’s alpha is widely used because it provides many advantages over other reli- ability measures. Besides its easy computation, it does not pose any restriction on the types of variables used, and it removes the memory effect possibility when measuring reliability (Bollen, 1989 recited in Hoang Trong and Chu Nguyen Mong Ngoc (2008).
If the reliability is not acceptably high, the scale can be revised by altering or delet- ing items that have scores lower than a pre-determined cut-off point. If a scale used to measure a construct has an alpha coefficient greater than 0.70, the scale is consid- ered reliable in measuring the construct (Nunnally, 1978; Leedy, 1997 recited in Hoang Trong and Chu Nguyen Mong Ngoc (2008). This indicates a high level of internal con- sistency or homogeneity among the constructs under each dimension (Straub, 1989 recited in Hoang Trong and Chu Nguyen Mong Ngoc (2008). Schuessler (1971 recited in Hoang Trong and Chu Nguyen Mong Ngoc (2008) suggested that a scale is consid- ered to have good reliability if it has an alpha value greater than 0.60. Hair et al. (1998 recited in Hoang Trong and Chu Nguyen Mong Ngoc (2008) suggested that reliability estimates between 0.6 and 0.7 represent the lower limit of acceptability for reliability estimates. In this research, multi-item scales measuring all corresponding constructs under each of the four dimensions affecting e-readiness including: (i) management; (ii) processes; (iii) technology; and (iv) people) were checked for reliability by determining Cronbach’s alpha, and an alpha value of 0.60 or greater was considered acceptable.
Results of Cronbach’s Alpha of the e-readiness tool with 27 questions in 4 categories
were displayed in Table3.3and the scale reliability was shown in AppendixC.
TABLE3.3: Cronbach’s Alpha of the questionnaire to measure e-readiness Item Cronbach’s Alpha N of Items
Total .942 27
Technology .914 8
Management .953 8
People .847 5
Process .940 6
As can be seen from Table 3.3, group items in the survey are returned in accepted values (the total value of 0.942 > 0.700), all individual indicators have values bigger than 0.300, and the alpha value if the item deleted in accordance with indicator sur- veyed is smaller than the group alpha synthesized, so the questions are significantly meaningful. In testing for reliability of technology category, as can be seen from the table, Cronbach’s Alpha of groups is 0.914 bigger than 0.600. So this group indicator reliability is acceptable in academic research. Furthermore, Appendix VI showed that all element indicators have Corrected Item - Total Correlation greater than 0.300 and Cronbach’s Alpha if the item deleted smaller than Cronbach’s Alpha of groups and Cronbach’s Alpha if the item deleted smaller than Cronbach’s Alpha of groups so they have correlation with other factors and should be retained in the model.
For management category, as can be seen from the table, Cronbach’s alpha of groups is 0.953 greater than 0.600. It can be concluded that this group indicator reliability is acceptable in academic research. All elements indicators have Corrected Item - Total Correlation bigger than 0.300 and Cronbach’s Alpha if the item deleted smaller than Cronbach’s Alpha of groups and Cronbach’s Alpha if the item Deleted smaller than Cronbach’s Alpha of groups, so they have correlation with other factors and should be retained in the model. Similarly, in people category the Cronbach’s Alpha of groups is 0.847 bigger than 0.700 so this group indicator reliability is good. All elements indica- tors have Corrected Item - Total Correlation greater than 0.300 and Cronbach’s Alpha
if Item Deleted smaller than Cronbach’s Alpha of groups so they have good correlation with other factors and should be retained in model. Finally, with the process category, as can be seen from the table, Cronbach’s Alpha of groups is 0.94 more than 0.700 so this group factor reliability is good. All elements indicators have Corrected Item - Total Correlation more than 0.300 and Cronbach’s Alpha if the item deleted smaller than Cronbach’s Alpha of groups so they have good correlation with other factors and should be retained in model.
The same procedure was applied to test the reliability of the 20 questions in the third part of the questionnaire. The results of the Cronbach’s alpha were 0.96 for the total 20 questions and the scale reliability was shown in AppendixD.