Does (problem-based) practice always make proficient?

The current study compared the rate at which problem-based practice increased the use of retrieval-based strategies for students identified as displaying accurate min-counting with students identified as displaying almost proficient performance. The findings supported the prediction that the rate at which problem-based practice promoted retrieval use was lower for students in the accurate min-counting group; in fact, it had no effect on their retrieval development. Implications for teaching practice are discussed, in particular, the notion that such students may require exposure to different problem representations (e.g., visual imagery) to move them away from conceptualising addition as counting.