Concept-Building in STEM Courses Study

Do students who use a rule-based approach to learning outperform rote-based learners in General Chemistry and other STEM courses that encourage creative problem solving and integration of complex concepts?

In an ongoing research program, Gina Frey (The Teaching Center, CIRCLE, and Chemistry) and Mark McDaniel (CIRCLE and Psychology), have been identifying students’ concept-building approaches (rote or rule-based) with an online function learning task, to examine the association between concept-building approach and performance in General Chemistry.

Project Overview

IMG_3405_Cropped_900_GenChem_Nov_2016We believe that one reason many talented students struggle and eventually migrate out of STEM fields is that they use a concept-building approach that is not compatible with the creative problem solving and integration of complex concepts necessary for success in many STEM fields. Basic cognitive psychology research (McDaniel, Cahill, Robbins, & Wiener, 2014) has identified stable tendencies for individuals to approach concept-learning tasks by either memorizing examples and features (rote-based learning) or by extracting rule-based information from examples (rule-based learning). Rule-based learners are able to use extracted rules and apply them to new problems and situations, whereas rote-based learners struggle to solve problems that are substantially different than those they have learned before.

In our research to date, we have identified approximately equal numbers of rote and rule-based learners and found that rule-based learners generally outperform rote-based learners in General Chemistry at Washington University and six collaborating institutions, with advantages ranging from about 3% to 9%. Further, when examining performance in a more fine-grained way, we find equivalent performance across concept-building approaches on near transfer problems—problems that are structurally similar to problems seen in class, homework or problem sets. The significant advantage of rule-based learners only emerged on far transfer problems, which require novel solution methods or integration across concepts.

Finally, at Washington University, we have tracked students into their second year to measure the influence of concept-building approach in organic chemistry and found an 11% advantage for rule-based learners over rote learners in overall exam performance.

Students collaborating in Gen ChemCurrently we are exploring the stability of the concept-building approach by administering the function-learning task twice across the two semesters of General Chemistry. Our initial results suggest that the concept-building approach is stable across some students, while others appear to switch between rote and rule-based approaches (adaptive learners). Our initial findings suggest that stable rote learners struggle in General Chemistry, relative to rule-based learners and adaptive learners, and that adaptive learners may benefit greatly from exposure to active-learning pedagogies.

In the near future, we plan to expand this project in several ways. First, we plan to expand our research programs into other STEM disciplines, such as biology, at the University. We also plan to bring the project to a wider range of institutions. Second, in addition to examining performance data, our future studies will directly examine, via questionnaires, the approaches and strategies that students use to learn in particular course. This will allow us to see how rote-based learners and rule-based learners approach courses differently and how those approaches influence performance. Finally, we will work to develop interventions that can move students toward rule-based learning or otherwise provide rote learners with strategies that can boost their performance to the level of rule-based and adaptive learners.


At Washington University, this research is being conducted through collaboration by CIRCLE, the Department of Chemistry. CIRCLE researchers include Gina Frey, Mark McDaniel, Michael Cahill, and Jiuqing Zhao.

Participating with Washington University in this investigation are six other colleges and universities that form the Luce consortium:

Berea College (KY)
Drexel University (PA)
Georgia Regents University (GA)
Ouachita Baptist University (AR)
University of Guelph (Canada)
Weber State University (UT)


This project has been supported by a four-year grant from the Henry Luce Foundation.


McDaniel, M. A., Cahill M.J., Robbins, M., Wiener, C. (2014). Individual differences in learning and transfer: Stable tendencies for learning exemplars versus abstracting rules. Journal of Experimental Psychology: General, Vol 143(2), 668-693.