Reasoning about curvilinear motion: Using principles or analogy

People possess implicit theories about the motion of objects, theories that are often incorrect. When asked to predict the path of an object emerging from a curved tube, for example, people often say that the object will continue to follow a curved path. However, when solving a problem that reminds them of a familiar previous instance, people often reason by analogy to the instance. In this study, we show that a previous instance must be very superficially similar to a problem in order to be used as an analogy; otherwise, people will use their implicit theories as the basis of their reasoning.