Prior research suggests that the effectiveness of retrieval practice may be moderated by response format: overt retrieval (e.g., typing a response) outperforms covert retrieval (e.g., mentally recalling a response) for complex materials like definitions, but both forms of retrieval are equally effective for simple materials like single words. These differences are explained by the "retrieval dynamics hypothesis," which posits that retrieving a single unit of information is relatively automatic and easy, while retrieving multiple interconnected units (e.g., definitions) is more demanding and may not be exhaustive when done covertly. In two experiments, undergraduates studied either key terms or definitions of declarative concepts and practiced via overt retrieval, covert retrieval, or restudy. Two days later, students completed a criterion test on either the key terms or definitions. Results supported the retrieval dynamics hypothesis: performance was higher for overt compared to covert retrieval, with a larger advantage for definitions than key terms. Both retrieval formats outperformed restudy. These findings suggest that overt retrieval practice is particularly beneficial for learning demanding material, whereas covert retrieval may be more efficient, though less thorough. Future research should explore methods to enhance the effectiveness of covert retrieval. Until then, we recommend students prioritize overt retrieval when mastering complex information.