H 15 and H 16 cannot be tested as Social Support did not resolve as a valid
construct; and
H 17 to H 20 could not be tested as Service Quality did not resolve as a valid
construct.
No. Description Outcome (Model no.)
(effect sizeb)
H 1 Motivational Leadership → Job Performance Confirmed (M5b) (0.315)
H 2 Motivational Leadership Behaviour →^ Discretionary Service Confirmed (M5b) (0.403)
H 3 Motivational Leadership → Work Meaning Confirmed (M2) (0.558)
H 4 Work Meaning → Job Performance Confirmed (M2) (0.258)
H 21 Motivational Leadership Attitudes →^ Employee Positive Confirmed (M5b) (0.213)
H 22 Employee Positive Attitudes → Job Performance Confirmed (M5b) (0.279)
H 23 Work Values →^ Employee Positive Attitudes^ NOT confirmed (M4/5)a
H 24 Empowerment → Employee Positive Attitudes Confirmed (M5b) (0.626)
a WV→EPA was confirmed in Model 4 (0.334) but the introduction of Employee
Empowerment in Model 5 reduced the effect of WV to a non-significant level (p>0.05)
b effect sizes are standardised regression coefficients and are referred to below as
beta weights () following Kline (2005: 31)
Table 7-35 Research hypotheses and research findings
The effect sizes for hypotheses 1, 2, 21, 22 and 24 are all for the total effects and
are taken from the final model (5b) as this is the model that explains the
relationships between the greatest number of constructs.
Hypotheses 3 and 4 and their respective effect sizes are drawn from Model 2 -
Models 3 to 5b went beyond the findings of Model 2 and the identity of the Work
Meaning (ME) construct was lost as it was merged into the broader Employee
Positive Attitudes (EPA) construct. Model 2, however, remains valid and
maintains a high degree of interest insofar as it empirically establishes the effect
of motivational leadership on employees’ work meaning – an effect that has been
theorised (Avolio and Bass 2004: 96; Bass and Riggio 2006: 6, 28, 91), but for
which empirical evidence was not found during the course of this research.
Regarding the interpretation of the effect sizes, the extent to which a given effect
size represents a small, medium or large level of practical importance is highly
dependent on the research context (Kline 2005: 121; Stevens 2009: 9).
Accordingly, in seeking to interpret the comparability and potential practical