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

 
2007 Annual International CHRIE Conference & Exposition 
394
(.69 and .72). Additionally, the results of the reliability analysis showed that the consistency of the measurement 
increases to the excellent level if the item is deleted. Thus, the problematic measurement item for CS was removed.
Following Anderson and Gerbing’s (1988) two-step approach, a measurement model was estimated before 
the structural model using AMOS 4. The chi-square value (
χ
2
) of the measurement model was 39.987 (
df
= 21, 
p 
< 
.01). Other practical fit indices demonstrated that the measurement model fits the data very well (RMSEA = 0.055; 
CFI = 0.998; NFI = 0.996). Cronbach’s alpha values were satisfactory for all four constructs (OSQ = .96; CS = .93; 
RVI = .95; WOMI = .88), indicating acceptable internal consistency (Nunnally, 1978). Further, all standardized 
factor loadings ranged from .88 to .97, which exceed Ford 
et al
.’s (1986) suggested cut-off of .40. The measurement 
model was assessed for construct validity. All indicator loadings for constructs were significant (p < .01). Also, all 
average variance extracted estimates in each construct (AVE; 
ρ
ve (
η
)
) exceeded .50, indicating adequate convergent 
validity (Hair 
et al
., 1998). These AVE values were greater than the square of the coefficient representing its 
correlation with other constructs, showing evidence of discriminant validity (Fornell and Larcker, 1981).