to compare with 211 responses received from the second group mailing in terms of their
demographic information. The results are depicted in Appendix B, Table 3A as shown
statistically significant difference between early and late respondents at 95% confident
level, indicating that there is non-response bias between respondents and nonrespondents
in terms of demographic. The results indicated that there is no
systematically difference in who responded and who failed to (Armstrong and Overton,
1977) that is there is no key problem in this research.
Mahasarakham University
60
Measurements Development
The measure development procedures involve the multiple items development
for measuring each construct in the conceptual model. All constructs are the
abstractions that cannot be directly measured or observed and should be measured by
multiple items (Churchill, 1979). These constructs are transformed to the operational
variables for truly of measuring. All of variables are gained from the survey and are