The idea of causality is central in science and has long given rise to debate among philosophers and scientists. While the tendency to avoid causality seems to have become dominant in science and philosophy, research in science education has shown the strong presence in common reasoning of causal explanations, often conceived as a 'mechanism' capable of accounting for physical transformations. Some researchers have proposed using this common causal reasoning as a basis for teaching-learning sequences, especially in electricity and mechanics. This paper analyses some features of causal reasoning used in physics by students, using questionnaires and interviews involving students and teachers. This study has shown three aspects which are related to one another: a confusion between efficient and contingent causes, between the conditions of occurrence of a phenomenon and the cause actually producing it; a tendency to 'displace' causes, skipping intermediate objects; and a difficulty in connecting local causes and global effects. The paper highlights the differences between common reasoning and scientific usage, and their effect on learning. In fact, these trends of reasoning must be taken into account in teaching: they should be considered not only as creating an obstacle to learning physics, but also as resources at the learner's disposal.