Microsoft Word 2007 ichrie conference Proceedings Final-Final 06-06-07. doc

 
2007 Annual International CHRIE Conference & Exposition 
298
were deemed unusable, therefore resulting in a usable sample of 333. The low response rate is a limitation of this 
study and is attributed to many factors including the prevalence of spam, fear of email virus, out-of-office reply’s, 
and the plethora of e-surveys in the marketplace. Non-response bias was checked by comparing the responses of 
early and late respondents and no significant differences were found. To test hypotheses, hierarchical regression 
model was tested. Hierarchical regression allows the investigator to enter individual predictors or sets of predictors 
in a specified order each of the independent variables can be entered cumulatively in some specified hierarchy, at 
each stage of which and R
2
is determined. The moderating effect was tested by examining the change in R
2 
(Hair, 
Black, Babin, Anderson, & Tatham, 2006). Moderator effects occur when the moderator variable, a second 
independent variable, changes the form of the relationship between another independent variable and the dependent 
variable. Since the loyalty program use was non-metric variable, it was coded as a dummy variable in a regression 
analysis; yes is 0 and no is 1.