This research aims at validating a previously developed 5-factor model of changeoriented leadership capability scale in academic settings. Data were collected from Malaysian academic leaders in 25 public and private universities through an online survey distribution platform. The result of data screening procedure highlighted the existence of one outlying case, which was removed from the dataset. In addition, the assumption of multivariate normality was not met. Hence, to run a confirmatory analysis to test the factorial validity of the proposed scale, we applied the robust Maximum-Likelihood (ML) methodology in EQS 6.4 software package to correct the chi-square statistic and generate robust standard errors. The outcome of the analysis shed light on the fact that the change-oriented capability scale developed in academic settings was a 4-factor model. In addition, the fit indices and standard errors generated through conventional ML estimator and the robust method were compared to provide more insights regarding the undue influence of the multivariate non-normal data on fit indices and standard errors. The implications of the findings with respect to the practice, theory, and methodology were discussed as well.