was necessary to test the interrater agreement and intraclass correlations before averaging across rater
groups (i.e. bosses, peers and subordinates) in order to obtain the scores for the focal managers on
both the strategic leadership and transformational leadership dimensions (see Table 7.2).
TABLE 7-2: INTRACLASS CORRELATIONS AND INTERRATER AGREEMENT.....................................................................................................
Similar to the findings in the previous chapter, Table 7-2 shows high interrater agreement scores
(ranging from .95 to .86) and acceptable intraclass correlation (ICC(I) scores (ranging from .18 to .49.
For 5 out of 9 dimensions the ICC(2) score did not reach the commonly mentioned criteria of .60
(Ostroff & Schmitt, 1993). However given the acceptable scores for Rwg and ICC(1) and the fact that
previous studies (de Hoogh et al., 2004; Simsek et al., 2010) have accepted ICC of .40, these results
can be considered as sufficient to aggregate.
7.3.2 COMPARING STRATEGIC LEADERSHIP WITH TRANSFORMATIONAL LEADERSHIP
In order to determine whether strategic leadership was different from transformational leadership an
exploratory factor analysis was conducted on the behavioral items underlying both strategic leadership
and transformational leadership. The exploratory factor analysis was conducted based on principal
component for extraction method and Varimax with Kaiser Normalization for rotation method.
Similar to the method adopted in the previous chapter, the first stage in analyzing the data with
exploratory factor analysis was checking the conditions for a stable factor structure (Ferguson & Cox,
1993). This required that the sample size was large enough. With 486 raters for 20 items, the subject-
to-variable ratio was 24:1, being significantly higher than the generally accepted minimum ratio (i.e.
5:1) (Ferguson & Cox, 1993).
The next step was to check whether the items were multivariate normally distributed by examining
their skewness and kurtosis. Following Ferguson and Cox (1993), the data was checked for whether