Purpose: The purpose of this paper is to investigate the factor structure and the measurement invariance of the Multifactor Leadership Questionnaire (MLQ) across gender of K-12 school principals (n=6,317) in the USA. Design/methodology/approach: Nine first-order factor models and four second-order factor models were tested using confirmatory factor analysis. Findings: The results suggested that the nine-factor model provided the best fit for the data. Further examination revealed that most constructs lacked convergent validity and discriminant validity. Second-order factor models were tested and the hierarchical model with two higher order factors (i.e. transformational and transactional leadership) was deemed the best fit and it was then tested for measurement invariance between females and males. The measurement model was found to be invariant across gender. Findings suggested that female school principals demonstrated significantly greater transformational leadership behaviour, while male school principals demonstrated significantly greater transactional leadership behaviour. Originality/value: This study addressed construct and factor issues previously associated with the MLQ in the measurement of transformational and transactional leadership among a variety of organizations. By using a sample of K-12 school principals across gender, this study has provided support that may ameliorate contextual doubts of transformational leadership behaviour when examining the relational aspects needed to improve schools.