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

 
Independent Variables & Dependent Variables: 
In this study, the multinomial logistic regression has six 
independent variables. The name of each independent variable was labeled by using factor names derived from the 
factor analyses – AORERP, AOPERP, AEERP, BORERP, BOPERP, and BEERP. The value of each independent 
variable was measured by the average scores of items which were highly correlated with each factor. The dependent 
variable was WTP for ERP, which has three categories.
In the data from Greece, as a result of fitting a three-category logistic regression model, in the likelihood 
ratio test, AORERP (Chi-square =4.815, df=2, p=0.09) and BOPERP(Chi-square =8.30, df=2, p= 0.01) were found 
to be significant among the variables loaded from principal factor analyses. In the logit function illustrating the 
result of the comparison between category 1and 2, consumers with higher attitude toward external ERP are 1.4 times 
more likely to be in category 2 than are consumers without such higher attitude. However, this result was not 
statistically significant. The confidence interval indicates the odds could be as little as .698 times or as much as 2.7 
times larger with 95 % confidence. In the logit function illustrating the result of the comparison between category 2 
and category 3, consumers with higher behavior intention toward operational ERP are 2.297 times more likely to be 
in category 3 than are consumers without such higher behavior intention. The confidence interval indicates the odds 
could be as little as .987 times or as much as 5.345 times larger with 95 % confidence. From these results, the 
different degree of consumer attitude toward external ERP was the major determinant of consumers’ choice between 
paying nothing and paying 2%-6% for ERP. Whether consumers are WTP 2% - 6% or more than 6% for ERP 
depended on the level of their attitude toward organizational ERP. The following table 1 shows the result of 
multinomial regression in the data from Greece.
In the data from U.S., the
 
likelihood ratios test showed that only behavior intention toward operational ERP 
(Chi-square =8.957, df = 2, p=0.01) had a significant effect on WTP for ERP. In the logistic regression models 
comparing category 1 with category 2 and category 2 with category 3, the degree of consumers’ behavioral intention 
on operational ERP was found to be a major factor for determining customers’ choice between paying nothing and 
paying 2% - 6% and between paying 2% - 6% and more than 6%. However, these results were not statistically 
significant. When comparing category 1 and 3, consumer with higher behavior intention on operational ERP are 
2.684 times more likely to be in group3 than are consumers without such higher behavior intention. This result was 
statistically significant. The confidence interval indicates the odds could be as little as 1.206 times or as much as 5. 
99 times larger with 95 % confidence. These results suggest that the different degree of consumer behavior intention 
on operational ERP was the major determinant of the level of WTP. The following Table 2 shows the result of logit 
function 1 comparing category 1 vs. category 3 by using data from the U.S.