Motivation scientists employing physiological measures to gather information about motivation-related states are at risk of committing two fundamental errors: overstating the inferences that can be drawn from their physiological measures and circular reasoning. We critically discuss two complementary approaches, Cacioppo and colleagues' model of psychophysiological relations and construct validation theory, to highlight the conditions under which these errors are committed and provide guidance on how to avoid them. In particular, we demonstrate that the direct inference from changes in a physiological measure to changes in a motivation-related state requires the demonstration that the measure is not related to other relevant psychological states. We also point out that circular reasoning can be avoided by separating the definition of the motivation-related state from the hypotheses that are empirically tested.