DescriptionThis exploratory study investigated the impact of causal beliefs on how educational leaders explain and solve a complex problem (the mathematics achievement gap). In the first part of the study, individuals' causal beliefs were examined from a systems perspective and patterns of causal understanding, ranging from less to more systemic, were defined. Causal understanding was defined as a function of beliefs about causal agency, breadth of causation, system levels and connectedness among levels, and some system archetypes. In the second part of the study, variations in problem solving as a function of individuals' different levels of causal understanding were examined. The study sample involved educational leaders (district leaders and school principals) and teachers (no leadership position). Data collection employed a structured interview protocol, allowing for verbal and pictorial representation of thought. Data analysis involved the use of quantitative and qualitative methods. Qualitative analyses defined major categories and themes of answers as well as different levels of systematicity in participants' causal beliefs, which in turn served to determine different patterns of causal understanding (from less to more systemic). Quantitative analyses employed causal beliefs and patterns of causal understanding as independent variables to investigate implications for problem solving. In terms of causal beliefs, findings from this study corroborated much of what has been documented in the science education literature regarding individuals' failure to understand causality in a system, suggesting that major barriers to systemic causal understanding may be pervasive across different age groups and fields of knowledge and experience. In terms of problem solving, causal thinking patterns were correlated with types of solutions and ways to involve others. More systems-oriented causal thinking was associated with system change and empowering ways to involve others. Other findings also described (a) individuals' dispositions to change their mental models when faced with contradiction and (b) which pedagogical changes individuals believed were necessary to improve math achievement. These descriptions supported discussions on how individuals' beliefs and problem-orientations might create self-reinforcing loops that worsen the problem and prevent productive system change. Implications for instruction and educational leadership training were discussed.