June 15, 2014
June 15, 2014
June 18, 2014
K-12 & Pre-College Engineering
24.186.1 - 24.186.8
Modeling Student Program Evolution in STEM DisciplinesThe 21st century workplace places a heavy emphasis on STEM disciplines making them anessential part of middle school and high school education. Unfortunately the US is laggingbehind the rest of the world in STEM competency at all levels. One of the ways STEM educationcan be made more interesting and relevant is by tying it to real-world problems. With this inmind, we have been developing a system called C3STEM (Challenge-based CollaborativeCommunity-centered STEM), where students can learn the fundamentals of STEM disciplineslike physics and mathematics by modeling real-world systems, simulating the models tounderstand the behavior of the system, and then running experiments with the simulation modelsto solve complex problems. One such domain that we are working on is the modeling of trafficflow in urban areas. C3STEM provides students with an agent-based modeling environment anda visual programming interface, complete with conditionals and mathematical operations, whichallows students to build vehicle models using computational thinking constructs. C3STEMallows students to simulate their models by incorporating model translation and execution withNetLogo. Students can view their simulation alone or side-by-side with an expert simulationrunning in lockstep and employing the same initial parameters. By building vehicle models inC3STEM, students learn science curriculum fundamentals and mathematical modeling principlesin realistic contexts. They can then use their individual vehicle models to build traffic flowmodels through city streets and intersections.We present preliminary work on modeling the evolution of students’ programs/models in theC3STEM environment. In this paper, we present an initial analysis of two small studiesperformed in summer 2013, one involving seven high school students who were consideringundergraduate studies and future careers in engineering, and the other involving six middleschool students, who attended a one week long science summer camp. In these studies, studentsfirst modeled basic vehicular deceleration to come to a halt at a stop sign and then accelerationaway from the stop sign. Subsequent tasks included modeling driver behavior at a stoplight anddriver behavior while attempting a left turn across traffic.Our initial analysis employs measures typically used to compare much larger segments ofsoftware code and modifies them to apply to the much smaller amount of code that studentscreate in C3STEM. Some of the more well-known comparison metrics, such as the bag of wordsscore and abstract syntax tree edit distance, allow us to assign a similarity score between eachversion of a student’s model and the expert model. Tracking model changes over time allows usto better understand students’ modeling progressions and hopefully their understanding ofphysics, mathematics, and computational thinking constructs. Detecting errors in student modelsallows us to scaffold the student’s learning to facilitate understanding of both computationalthinking skills and domain specific knowledge.
Dukeman, A., & Shekhar, S., & Caglar, F., & Gokhale, A., & Biswas, G., & Kinnebrew, J. S. (2014, June), Analyzing Students' Computational Models as They Learn in STEM Disciplines Paper presented at 2014 ASEE Annual Conference & Exposition, Indianapolis, Indiana. https://peer.asee.org/20077
ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2014 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015