However, the technique needs to be used appropriately
by adhering to some key recommendations on using and
reporting SEM. Typical shortcomings identified by Smith
and Langfield-Smith (2004) with respect to SEM are issues
regarding model fit, small sample sizes or the use of nonnormal
data. Several other reservations regarding the use
of survey data are expressed in the accounting literature
more generally. They refer to inadequate pre-testing,
response rates and application of econometric methods
(e.g., Van der Stede et al., 2007; Ittner and Larcker, 2001;
Zimmerman, 2001). In our research design as well as in
performing our tests, we heed these warnings appropriately.