Investigations into perceptions of causality typically presume simple causal explanations. In contrast, this study allowed account complexity and structure to vary. Eighty-nine participants were asked to choose the causes necessary to explain one of four outcomes differing in severity and valence of consequences. They had eight possible causes to choose from that varied along dimensions of locus, stability, and controllability. Most explanations were three causes in length arranged in a temporal chain. Internal, stable, and controllable causes were used more than external, unstable, and uncontrollable causes. Results were consistent with the fundamental attribution error and the need for predictability and control. This suggests that complex explanations of necessary causality may be more common than previously assumed, but that the dimensions used for explanation remain consistent with established theory.