Mixed methods in management research: from juxtaposition to explanatory integration in the study of organizational behavior

Contemporary organizations increasingly confront multi-generational complexity that single-method designs struggle to capture, yet mixed methods studies remain more often juxtaposed than genuinely integrated, particularly in underdocumented African contexts. This article addresses that gap by documenting a reproducible explanatory sequential mixed methods protocol, in which a quantitative survey phase (137 participants across thirty companies in Madagascar) identified statistical patterns requiring explanation, which then guided a purposively sampled qualitative phase of interviews and focus groups, with triangulation matrices classifying each quantitative-qualitative pairing as convergent, complementary, or divergent. Divergent pairings revealed three behavioral paradoxes concerning loyalty, recognition, and autonomy. The study's principal contribution is methodological rather than empirical: a documented, transferable procedure for converting statistical divergence into theoretically grounded meta-inferences that neither method could produce alone. This procedure in turn equips organizations and researchers with a replicable diagnostic for identifying generational misunderstandings that conventional single-method approaches would leave undetected.