Bryan Caplan's blog post argues that behaviorism is effective because it implicitly relies on assumptions about mental states like preferences and beliefs. via Econlib