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

Factor 7. Brand trust 
V38 My satisfaction with a store (web-based or otherwise) is the company's 
most important concern 
.861* 5.15 
V39 My satisfaction with the product/service is a company's most important 
concern 
.839* - 
Dropped variables: None 
Note:
13 of original 39 observed variables were removed for the revised model of Consumer Experience; Standardized loadings indicate 
relationships between observed variables and their associated factors in order to examine convergent validity; * Factor loadings were all 
significant at p < .001. 
 
Cross-validity 
Cross-validity was assessed to see if there are differences in factor loadings in the tested measurement 
model across groups. In other words, factor loadings were tested for their equivalence across groups. Both data 
obtained in the first-split sample for EFA (Group 1; N =198) and the second-split sample for CFA (Group 2; N = 
199) were analyzed via multiple group analysis in SEM. Multiple group analysis is to evaluate a whole model across 
independent samples. Equal factor loadings imply that the degree of correspondence between each observed variable 
and its respective factor is the same for the first-split sample and the second-split sample. The 
χ
2
difference test was 
performed to compare the unconstrained and fully constrained model. An unconstrained structural model is 
estimated, in which factor loadings are allowed to vary across the cross-group datasets. Next, a fully constrained 
structural model is estimated by requiring that all factor loadings are constrained to be equal for cross-group 
datasets. The 
χ
2 
statistics for the unconstrained model and constrained factor models were 834.8 (
df 
= 556) and 
897.1 (df = 603), respectively. The difference is 62.3, with 47 degrees of freedom. There was no significant 
difference (p = .067) in factor loadings across groups. Based on this result, all measures of experience are operating 
in the same way for both groups. Thus, this data support cross-validity of the experience construct.