We previously reported that students' concept-building approaches, identified a priori using a cognitive psychology laboratory task, extend to learning complex science, technology, engineering, and mathematics topics. This prior study examined student performance in both general and organic chemistry at a select research institution, after accounting for preparation. We found that abstraction learners (defined cognitively as learning the theory underlying related examples) performed higher on course exams than exemplar learners (defined cognitively as learning by memorizing examples). In the present paper, we further examined this initial finding by studying a general chemistry course using a different pedagogical approach (process-oriented guided-inquiry learning) at an institution focused on health science majors, and then extended our studies via think-aloud interviews to probe the effect concept-building approaches have on problem-solving behaviors of average exam performance students. From interviews with students in the average-achieving group, using problems at three transfer levels, we found that: 1) abstraction learners outperformed exemplar learners at all problem levels; 2) abstraction learners relied on understanding and exemplar learners dominantly relied on an algorithm without understanding at all problem levels; and 3) both concept-building-approach students had weaknesses in their metacognitive monitoring accuracy skills, specifically their postperformance confidence level in their solution accuracy.