Analysis and Results

Common Method Variance

To check for common method variance, and thus to assess whether variance in the data could be attributed to a single factor, the Harman's one-factor test was run. This test, which involves an exploratory factor analysis of the data, revealed an unrotated factor solution involving 12 factors with eigenvalues >1. Because the first factor explained less than half (36%) of the total variance (78%), common method variance does not appear to affect the study findings.

We also ran Lindell and Whitney's marker variable technique to confirm these results. In essence, the marker variable technique uses a (marker) variable that is not theoretically related to any substantive variable of the study to calculate common method variance, and thus adjust the correlations among the study constructs. Although demographic variables are found not to be the most desirable option, there are many examples in the literature that use this option as a marker variable. Thus, in running the marker variable technique, we used the respondent's job type (0 = non-supervision role; 1 = supervision role) as a variable that was theoretically unrelated to any substantive study variable and could meet the necessary conditions to be a marker variable. As expected, this marker variable was not significantly correlated with any of the study variables. Furthermore, following Lindell and Whitney's recommendations, the lowest absolute correlation between the marker variable and the substantive study variables (rm = 0.05) was partialled out from the uncorrected correlations to check for the magnitude and significance of common method variance. After controlling for common method variance, all correlations that were previously significant remained significant, so we can conclude that common method variance is unlikely to have affected our findings in the current study.

TABLE 2. Descriptive statistics and correlation matrix (N = 270).

Descriptive statistics Correlation matrix. Cronbach’s alphas in bold (in the diagonal)
  Mean SD 1 2 3 4 5 6 7 8
1. Ethical leadership 3.61 0.52 0.90              
2. OCE 3.49 0.61 0.86 ∗∗ 0.88            
3. Years of experience in the job 2.57 0.67 –0.29 ∗∗ –0.28 ∗∗ n.a.          
4. Gender n.a. n.a. 0.12 ∗ 0.17 ∗∗ 0.05 n.a.        
5. Age 2.36 0.84 –0.00 0.11 0.38 ∗∗ –0.06 n.a.      
6. Level of education 1.71 0.57 –0.28 ∗∗ –0.28 ∗∗ 0.15 ∗ –0.08 –0.04 n.a.    
7. Job type n.a. n.a. –0.06 –0.10 –0.02 –0.06 0.26 ∗∗ –0.32 ∗∗ n.a.  
8. Employee readiness to change 3.32 0.83 0.74 ∗∗ 0.77 ∗∗ –0.22 ∗∗ 0.13 ∗ 0.03 –0.18 ∗∗ –0.05 0.72

∗∗p < 0.01; ∗ p < 0.05 (two-tailed test). SD , standard deviation; n.a., not applicable; OCE, organizational culture of effectiveness. Gender and job type were coded as dummy variables (1 = male, 2 = female; 0 = supervision role; 1 = non-supervision role), so mean and SD were not applicable; percentages were instead calculated, with values of 64.1% for males and 67% for supervision roles. Interval scales were used for measuring years of experience in the job (1 ≤ 5 years, 2 = between 5 and 10 years, 3 ≥ 10 years), age (1 = up to 24 years old, 2 = between 25 and 34 years old; 3 = between 35 and 44 years old; 4 = between 45 and 54 years old; 5 = 55 or more than 55 years old), and level of education (1 = primary studies; 2 = secondary studies; 3 = bachelor’s degree; 4 = master’s degree; 5 = Ph.D./doctorate).