4.4.7. Statistical procedures
To test the external fit hypothesis, ordered probit analyses were computed for the
sub-samples of ‘quality-enhancer’ hotels and ‘cost-reducer’ hotels on HR outcomes
and organizational performance outcomes. ‘Other’ hotels were excluded from this
level of analysis because of limited sample size and equal variance on some HR and
performance outcomes. Ordered probit analyses were also used to test the universal
relevance of HRM and internal fit hypotheses. In order to test for the universal
relevance of HRM, ‘low-HRM quality-enhancers’, ‘low-HRM others’ and ‘high-
HRM others’ hotel categories were dropped from the analysis because of the limited
size of these categories; the ‘high-HRM quality-enhancer category was held
constant. In order to test for internal fit, the ‘low HRM’ category was held constant.