non-responses due to respondent question fatigue and (b) demonstrating the
brevity of the survey to hotel senior managers who in this case are ‘gatekeepers’
to the population of interest. The Service Quality construct stands somewhat
separately in terms of how it was developed. Specifically, this construct is
entirely exploratory having been developed from concepts and findings from a
non-psychometric source study (Bitner et al. 1 990, 1994).
It is entirely normal to make adaptions to existing scales to render them more
appropriate for a specific context, either in relation to reducing the number of
items or adjusting the substantive content. For examples of scale adaptions in
the hospitality organisational studies context see the discussions above on
previous work on Organisational Commitment and Employee Empowerment.
Nevertheless, such a course of action does bring epistemological / methodological
implications for the research. Specifically, an exploratory element is introduced to
the research and, as a consequence, the confirmatory factor analysis (which
provides the analytical method for the measurement model in SEM analyses) is no
longer strictly confirmatory and the broader SEM process enters a model
development mode.
Accordingly, under these conditions the researcher must prepare for the
eventuality that some of the factors may not converge on a statistically
acceptable solution and may therefore have to be removed from the model.
Somewhat less severely, it may the case that one or more indicators have to be
dropped from a latent factor in order to achieve statistical convergence and this
may affect the substantive content of the factor. Such modifications may also
have implications for the broader structural model where causal relations between
factors are evaluated. The issues are examined in greater detail in the following
section.