The purpose of this paper is to address the statistical issues associated with the hierarchically structured data in previous studies that focused on servant leadership. To resolve these issues, multilevel modeling methods were applied to re-visit the construct validity of the servant leadership questionnaire developed by Barbuto and Wheeler (2006) and investigate the relationship between servant leadership and job satisfaction under a multilevel framework.
The survey data was obtained from a sample of 2,089 teachers from 117 primary and secondary schools in Hong Kong. The analyses were conducted using multilevel confirmatory factor analysis (MLCFA) and multilevel structural equation modeling (MLSEM).
The results revealed the significant and non-trivial variances that were explained at the organization level in the items measuring servant leadership, which justified the use of MLCFA and MLSEM. The results of MLCFA provided empirical support for the multidimensional construct as well as the second-order factorial structure of servant leadership measures at both the individual and organization levels. In addition, the positive relationships between servant leadership and the followers’ job satisfaction were found to vary at different levels.
This study reiterates the importance of using appropriate methods to capture a solid definition of the construct of servant leadership and provides new insights into the conceptual framework of servant leadership as well as the effects of servant leadership on individual and organizational outcomes.
