Long-term knowledge and skill retention in undergraduate biology students

Background/Question/Methods

Instructors often interpret high scores on final exams as evidence that students have mastered the skills and content of the course. We reformed 3 sections of an introductory biology course (n=517) to include models as an important pedagogical component. Talk-aloud interviews (n=30, 10 per tritile based on incoming GPA) were conducted 2.5 years after course completion to determine whether students retained their abilities to connect genetic variation to evolutionary change in a box-and-arrow model. Most (22) of the students were in STEM majors at the time of the interview. During the interview, students created a model similar to the final exam. Models were analyzed for their biological correctness and level of complexity (linear to fully interconnected). In a second interview task, students ranked sample models that varied systematically with 3 levels of biological correctness and 3 levels of complexity along a spectrum of “best”, “correctness”, and “complexity”. 

Results/Conclusions

During the interview, high and middle tritile students began from a point of strength: they were familiar with more genetics and evolution terms and started their models with higher quality relationships. However, the middle tritile students maintained their higher quality and constructed significantly more correct (p<0.001) overall models than low and high tritile students. On average, students were unable to accurately represent the connection between gene and allele in their models but could correctly describe the relationship when prompted verbally. There were also differences associated with experience in biology; STEM majors were more proficient at correctly defining fitness and connecting it to phenotype. In the ranking task, high and middle tritile students always identified a model with the highest correctness as “most correct” while low tritile students selected various models. Students generally selected high correctness models as “best” and selected high complexity models as “worst”, with most students (97%) stating that correctness was more important than complexity, but that a certain amount of complexity may be necessary. All students selected “best” models that were more biologically correct than their own model (p<0.001), which suggests students are better able to evaluate the quality of models for biological correctness than to create them. Since modeling has been identified as a core competency for 21^st century biologists, instructors should consider how to provide opportunities for students to move beyond model evaluation towards practicing model construction.