While the majority of published research on conceptual change has focused on how misconceptions can be abandoned or modified, some recent research findings support the hypothesis that acquired scientific knowledge does not necessarily erase or alter initial non-scientific knowledge but rather coexists with it. In keeping with this "coexistence claim," this article presents an analysis of scientific understanding in four groups of individuals with varying degrees of expertise (preschoolers, elementary students, secondary students, and science teachers) using a cognitive task on buoyancy. This task allowed us to determine the prevalence of certain conceptions and the interference caused by two possible conceptual distractors with regard to producing accurate answers. Results describe the progression of the desired (scientific) conception with age/expertise as well as the evolution or regression of the statuses of two misconceptions. Results also show that misconceptions continue to interfere with performance even when there is a higher degree of scientific expertise, and that patterns of such interference can be studied. In keeping with these conclusions, we argue for the use of a model of conceptual learning called "conceptual prevalence."