Ambidextrous leadership theory proposes that a leader's interplay between opening behaviors and closing behaviors enhances followers' exploration and exploitation behaviors, which ultimately increases innovative outcomes. Unfortunately, previous research suffers from problems with causal interpretation and endogeneity concerns threatening the validity of the theory. Our aim was to constructively replicate previous research with an experimental design, more rigorous measures, and state-of-the-art data analytical approaches (2SLS). In two randomized experiments (Study 1: N = 395, Study 2: N = 229), we manipulated four leadership styles (opening, closing, ambidextrous, and transformational leadership) and tested their effects on participants' exploration/exploitation behaviors as well as objective innovation outcomes. We only found partial support for the hypotheses from ambidextrous leadership theory. We discuss implications in terms of refining central concepts of the theory and offering more accurate assumptions about timing. We also elaborate on more general insights from our constructive replication studies for the leadership field. This project contains all data and materials used for a conceptual replication which has now been published: Klonek, F.E., Gerpott, F., Parker, S. (2020). A conceptual replication of ambidextrous leadership theory: An experimental approach. The Leadership Quarterly (Advanced online publication). https://doi.org/10.1016/j.leaqua.2020.101473 The article will appear in a special issue from Leadership Quarterly (see call for papers below): Clapp-Smith, R., Carsten, M., Gooty, J., Connelly, S., Haslam, A., Bastardoz, N., & Spain, S. (2018). Special registered report issue on replication and rigorous retesting of leadership models. The Leadership Quarterly, 88, V–VI.