Results
The data were tested in a two staged process (Anderson and Gerbing, 1988).
First, the measurement model was assessed via confirmatory factor analysis
to evaluate the reliability and unidimensionality of the scales used in the
study. Once we were satisfied with the measurement model, a structural
model was estimated using the SAS Institute’s (1990) PROC CALIS causal
modeling procedure. This allowed for an assessment of the overall model fit
and the hypothesized relationships. We also tested for multicollinearity in
our data by examining the variance inflation factor (VIF) associated with
each independent variable (operationalized as summed scales for the VIF
analysis). The VIF values do not exceed 10.0 (with the highest being 1.08),
thus indicating that multicollinearity does not impact the dependent
measures.