Because of the focus of introductory physics courses on improving students' problem-solving and reasoning skills, researchers in physics education have been developing and refining theoretical frameworks for how students reason through physics problems. Recently, researchers have begun to apply dual-process theories of reasoning (DPToR), from cognitive science and psychology, to support mechanistic predictions of student reasoning in physics. In this article, we employ a novel methodology involving reasoning chain construction tasks in order to test DPToR-based predictions for two physics questions in which salient distracting features have been found to cue incorrect first-available mental models. In a reasoning chain construction task, students respond to a physics question by drawing from a list of reasoning elements (all of which are true) in order to assemble a chain of reasoning that leads to a conclusion. Two sets of experiments were conducted to test the hypothesis that students would be unlikely to abandon an incorrect first-available model unless they were provided with information that called into question the validity of that model. We found that providing increased access to information relevant to the correct line of reasoning did not produce large differences in student answering patterns. However, providing increased access to information refuting the first-available model did produce large differences in student answering patterns, but only among those students who demonstrated that they possessed the relevant mindware (i.e., conceptual understanding). Our findings are consistent with DPToR and further illustrate the applicability of such reasoning frameworks in the context of physics.