How is a Therapist like a Modeler?

This paper argues that the process of modeling in science and the process of encountering and working with a client in clinical psychotherapy overlap. In briefer terms: what makes a good therapist is much like what makes a good scientific modeler. Both modeling and psychotherapy are iterative processes, requiring careful observation, generation and testing of hypotheses. Both processes also face similar epistemic and pragmatic trade-offs. Heuristics and biases can shape both practices, for better and worse. Implications are considered for both training in clinical psychotherapy, and for larger debates concerning evidence-based methods in assessment of mental health care, and categorical v. dimensional approaches to diagnosis.