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

 
2007 Annual International CHRIE Conference & Exposition 
460
The service climate measure (SERV*OR) was subjected to an exploratory factor analysis because of the 
relatively small sample size coupled with the large number of factors and dimensions found in the original 
application of the SERV*OR instrument, which suggested a low likelihood of factor replication (Tabachnick & 
Fidell, 2001). The 34 items of the SERV*OR instrument were subjected to a factor analysis using SPSS FACTOR, 
Principal Axis Factoring. The initial solution yielded a scree plot suggesting between 3 and 5 factors, accounting for 
between 59.5% and 64.9% of the variance, respectively. A substantive theoretical reading of the factor structures 
suggested that the 5-factor solution was the most meaningful of the set, indicating that while differing from the 
original factor set, the differences made conceptual sense. Accordingly, 5 factors were submitted for rotation 
because the results of factor extraction, if unaccompanied by rotation, are often hard to interpret.
In order to determine which rotation method would yield the most interpretable results, both orthogonal 
(i.e., varimax) and oblique (i.e., oblimin) methods were applied. Inspection of loadings suggested that the oblique 
solution yielded the most interpretable and simple structure of the two methods. Table 1 shows the 5 factors along 
with a sampling of factor loadings, values of Cronbach’s alpha for the resulting unit-weighted composites derived 
from the factors, and the total factor variances. Although the result did not directly replicate the original scale 
structure found by Lytle et al. (1998), the structure was similar to the original and the 5-factor structure made 
conceptual sense. 
Table 1