Hybrid Labs: Integrating Computational Modeling With Physical Experiments in Undergraduate Biology Labs

• In Event: Bridging Computational Modeling Tools and Practices Into the Existing Structures of K–16 Environments in Science Education

Abstract

Project Goals
In modern biology, experiments and computational models are two powerful and complementary approaches to conducting research on complex systems. Computational models allow students to rapidly explore multiple parameters, conduct simulated experiments and get immediate feedback. Experiments challenge students to confront the material challenges of designing experiments in a way that will yield interpretable data. Integration between the two methodologies has been shown to drive innovation in science (MacLeod & Nersessian, 2013) and learning (Blikstein, 2007).

The Hybrid Labs project aims to leverage the affordances of coordinating these methodologies to generate and pursue investigations into complex biological systems. Our research questions are broadly related to examining students’ reasoning and engagement in scientific practices afforded by these two modalities. In this paper, we will examine the role of the computational modeling tool in supporting students’ investigation of a complex biological system.

Design intervention & data
We use MacLeod & Nersessian’s construct (2013) of coupling methodologies as an orienting framework to design curricular lab units that integrate these two modalities for students to inquire into biological systems (See Figure 1). We have designed and developed 3 curricular labs units, each of which includes 3 3-hour long lab sessions. These lab units are implemented in an introductory undergraduate biology lab, which is taught by a graduate teaching assistant.
In this paper, we draw on analysis of data from case studies of 3 groups of 3-4 undergraduate students investigate the E.coli bacterial system. The question that drove their investigation was: Is it better to be a low or a high mutator? Students designed and conducted experiments, and explored a computational model of the system to compare tradeoffs between being a high or low mutator in different environments. Our data included video data of groups from 3 3-hour lab sessions, screen-capture videos of their work in the modeling environment, written lab reports of all group members and interviews with some of the focal students. We analyzed this data to examine the ways in which the computational model supported their engagement with this system. We first identified episodes in which students moved between the computational and experimental systems. We then analyzed the extent to which each of these links helped students make progress in their scientific inquiries.

Findings & conclusion
Our analysis revealed that students encountered productive tensions between their observations from and analysis of the computational modeling environment and results from their physical experiment. We found students using the experimental system to inform and constrain their investigations in the computational modeling environment, using the computational modeling environment to refine their research question and make predictions, using the modeling environment to explore possible mechanisms that could explain and extend their experimental results, and comparing model output and experimental data analysis to draw conclusions and propose new experiments.

Lessons learned
This analysis has led to the identification of design principles for ways to meaningfully integrate these modalities, and considerations for design in mapping between an experimental system and a computational model.

Authors

• Aditi Wagh, Tufts University

• Julia Svoboda Gouvea, Tufts University