The more the merrier? A network analysis of construct content validity in school leadership literature

Abstract

School leadership research literature has a large number of widely used constructs. Could fewer constructs bring more clarity? This study evaluates construct content validity, defined as the extent to which a measure's items reflect a theoretical content domain, in school leadership literature. To do so, we reviewed 29 articles that used Teaching and Learning International Survey (TALIS) as data sources to study school leadership, and developed a construct co-occurrence network in which nodes represented constructs in the 29 reviewed articles and ties connected a pair of constructs that used the same TALIS survey items. Among the 82 constructs, 43 constructs had overlapping measures with one another in the TALIS literature on school leadership. Results of network analysis suggest three problematic issues with constructs of TALIS literature on school leadership: (1) same measures for different constructs, (2) different measures for the same constructs, and (3) missing alignment between theoretical and operational definitions. To strive for construct clarity, we provide four recommendations for future research: (1) efforts to prevent construct proliferation, (2) alignment of theoretical and operational definitions, (3) rigorous evaluation of construct validity, and (4) following a fundamental principle of parsimony.