the RMSEA is satisfactory at 0.031 with the upper limit of 0.072 still within
Browne and Cudeck’s (1989, 1993) >0.05 and <0.08 range of ‘reasonable fit’
and a satisfactory pclose value;
CFI is satisfactory at 0.996 (i.e. >0.96);
SRMR (0.0491) is satisfactory at <0.09; and
Hoelter’s CN of 285 is above the recommended 200.
The structural paths are both statistically significant at the 0.001 level (meaning
that we can reject the null hypotheses that the described relationships do not
exist) and the coefficients show that Motivational Leadership has an effect size of
0.414 on Job Performance and 0.396 on Discretionary Service Behaviour. In
statistical terms, these figures mean that for every (standardised) change of
value in Motivational Leadership, Job Performance will change by 0.414 of that
value and Discretionary Service Behaviour will change by 0.396 of that value. In
more practical terms, the effect size (or strength of association) describes how
much of the variance of the dependent variable is predictable from knowledge of
levels of the independent variable (Tabachnick and Fidell 1989: 54).
The absence of any empirical precedents for these effect sizes makes it difficult to
judge the importance of these effect sizes. No identified research has measured
the effect of motivation leadership on employee performance and, while Simons
and Roberson’s study employed the DSB construct, their leadership focus was on
employee satisfaction with their leader rather than leader behaviour.
Kline (2005: 121-122) writes that without any theoretical or empirical guidance
on what magnitude of effect represents an ‘important’ effect for any particular
relationship, it can be difficult to interpret effect sizes. This research is largely
exploratory (insofar as it is examining research questions hitherto not addressed
in hospitality contexts) and therefore falls into what Kline (p. 122) refers to as a
“new research area”. In such circumstances, Kline recommends the guidelines on
interpreting effect sizes provided by (Cohen 1988). Specifically, these guidelines
suggest that:
small effect = effects sizes less than 0.1;
medium effect = effect sizes around 0.3; and
large effect = effect sizes around 0.5 or greater.
Based on these guidelines the effects of the structural coefficients of 0.414 and
0.396 are interpreted as medium to large effects. These findings represent
important contributions for this research in: