Skip Navigation

Science, Technology, Engineering, and Mathematics (STEM) Education

Grantees

- OR -

Investigator

- OR -

Goals

- OR -

FY Awards

- OR -

Promoting Robust Understanding of Genetics with a Cognitive Tutor that Integrates Conceptual Learning with Problem Solving

Year: 2009
Name of Institution:
Carnegie Mellon University
Goal: Development and Innovation
Principal Investigator:
Corbett, Albert
Award Amount: $1,447,525
Award Period: 3 years (7/1/09-6/30/12)
Award Number: R305A090549

Description:

Co-Principal Investigators: Baker, Ryan S.; Mitchell, Aaron

Purpose: Genetics is a fundamental and unifying theme of biology, but genetics problem solving is very challenging for high school students. The goal of this project is to develop educational technology for Conceptually-Grounded Problem Solving, scaffolded by intelligent learning support. Abductive problem solving is a form of logical inference that goes from an observation to a hypothesis that accounts for the observation, seeking to find the simplest and most likely explanation.  A progression of activities in the software will support rich conceptual understanding of genetics processes and engage students in constructing the abductive reasoning that relates observable data to underlying processes. high school students will efficiently acquire robust knowledge of genetics.  The purpose of this project is to develop an intelligent tutoring system for learning genetics. This educational software will be formatively evaluated, with human subjects, in two situations.

Project Activities: In this project, researchers will work to develop knowledge about how educational software for genetics can be more educationally effective.  Researchers will show how concept grounding activities can be integrated with abductive problem solving.  The research team will pilot the seven existing Cognitive Tutor Abductive Problem Solving units to establish the preliminary feasibility of stand-alone problem-solving units in high school classes. In the second project year, researchers will implement and test the feasibility of the full set of Conceptually Grounded Problem Solving activities: Forward Modeling, Solution Methods Construction, Worked Example Explanation, and Abductive Problem Solving, for four genetics topics. In the third year of the project, researchers will test the full Cognitive Mastery version of the Conceptually-Grounded Problem Solving environment, designed to test the additional opportunities for students to learn skills and concepts that have not yet been mastered, for all seven of the genetics topics. In addition, for all seven genetics topics, researchers will compare measures gathered during this final test to the first-year baseline data, to evaluate the overall learning efficacy of the new Cognitive Mastery Conceptual Grounding activities.

Products: In this project, products include a fully developed intelligent tutoring system, The Genetics Cognitive Tutor, for use with high school students (ages 14-16).  Peer-reviewed publications for research and practitioner audiences will also be produced.

Structured Abstract

Setting: The Conceptually-Grounded Genetics Problem Solving environment will be piloted in biology courses in three high schools in the Pittsburgh area, an urban public school, an urban private school and a rural public school. Researchers will also recruit a diverse group of Pittsburgh area high school students to participate in formative laboratory studies at a local area university.

Sample: Participants include high school students (ages 14–16) for pilot studies. Pilot schools represent a racially and economically diverse group of students including 40% minority students, predominantly African-Americans, in the urban school. 44% of students in the rural public school qualify for free or reduced-price lunches.

Intervention: A progression of three new Cognitive Tutor Conceptual Grounding Activities will be developed: (1) Forward Modeling, (2) Solution Methods Construction, and (3) Worked Example Explanation. The new Conceptually-Grounded Problem Solving environment will leverage an existing Cognitive Tutor for abductive genetics problem solving. Student modeling methods will be developed to individualize the sequence of learning activities, enabling students to learn efficiently in the new environment.

Research Design and Methods: A process of iterative design will be used, consisting of three cycles of software development, formative laboratory studies, iterative redesign, and classroom studies of feasibility. Piloting in authentic classroom contexts will begin in the first project year. Researchers will pilot the seven existing Cognitive Tutor Abductive Problem Solving units, using student surveys, classroom observations and teacher interviews to establish the preliminary feasibility of these stand-alone problem-solving units in high school classes. In the first, second, and third years, researchers will also collect five learning efficiency measures (pretest/posttest learning gains, transfer test performance, learning rates, cognitive model predictive accuracy and gaming rates). These measures will serve as baseline data for classroom piloting of the full complement of learning activities in the following project years (in the same courses, taught by the same teachers). In the second project year, researchers will deploy the full set of Conceptually Grounded Problem Solving activities: Forward Modeling, Solution Methods Construction, Worked Example Explanation, and Abductive Problem Solving, for four genetics topics. Again, student surveys, classroom observations and teacher interviews will be used to establish the preliminary feasibility of the full Conceptually Grounded Problem Solving activity sequence in high school classes. In the third project year, researchers will implement the full Cognitive Mastery version of the Conceptually-Grounded Problem Solving environment for all seven of the genetics topics. Feasibility will be determined using surveys and interviews of the full Conceptually Grounded Problem Solving activity sequence in high school classes. For all seven genetics topics, researchers will compare these measures to the first-year baseline data, to evaluate the overall learning efficacy of the new Cognitive Mastery Conceptual Grounding activities. For the first four genetics topics, the research team will compare the data from the second and third years to evaluate the added value of Cognitive Mastery.  The formative evaluations will determine the potential impact of each of the new Conceptual Grounding activities on performance and learning, and the potential impact of different ways of individualizing the sequence of learning activities, on learning efficiency.

Control Condition: The control group will consist of students who only complete abductive problem solving using the existing Cognitive Tutor.

Key Measures: Multiple feasibility measures will be employed including student attitude surveys (Woodrow’s Computer Attitude Scale), as well as measures developed (or to be developed) by the research team designed to test student learning rates and “gaming the system” in the learning environment. The team will also measure student learning outcomes using problem solving pre-tests and post-tests, and transfer tests. The outcomes will be evidence on the software’s feasibility for use in classrooms, and on its potential for improving students’ learning of key concepts and skills in the domain of Genetics.

Data Analytic Strategy: Qualitative data from surveys and interviews will be coded and analyzed.  Researchers will confirm predictive validity of the Cognitive Tutor’s Knowledge tracing estimates, and will use inferential statistics (t-tests, ANOVAs) to examine any differences in test condition.

Publications

Book chapter

Baker, R.S.J.d. (2010). Mining Data for Student Models. In R. Nkambou, R. Mizoguchi, and J. Bourdeau (Eds.), Advances in intelligent tutoring systems, SCI 308(pp. 323–337). Berlin Heidelberg: Springer.

Journal article, monograph, or newsletter

Baker, R.S., Corbett, A.T., and Gowda, S.M. (2013). Generalizing Automated Detection of the Robustness of Student Learning in an Intelligent Tutor for Genetics. Journal of Educational Psychology, 105(4): 946–956.

Proceeding

Baker, R.S.J.d., Gowda, S.M., and Corbett, A.T. (2011). Automatically Detecting a Student's Preparation for Future Learning: Help Use is Key. In Proceedings of the 4th International Conference on Educational Data Mining (pp. 179–188). Eindhoven, the Netherlands: Educational Data Mining.

Baker, R.S.J.d., Gowda, S.M., and Corbett, A.T. (2011). Towards Predicting Future Transfer of Learning. In Proceedings of 15th International Conference on Artificial Intelligence in Education (pp. 23–30). Berlin Heidelberg: Springer.

Corbett, A., MacLaren, B., Wagner, A., Kauffman, L., Mitchell, A., and Baker, R.S.J.d. (2013). Differential Impact of Learning Activities Designed to Support Robust Learning in the Genetics Cognitive Tutor. In Proceedings of AIED 2013, LNAI 7926 (pp. 319–328). Berlin Heidelberg: Springer.

Corbett, A., MacLaren, B., Wagner, A., Kauffman, L., Mitchell, A., and Baker, R.S.J.d. (2013). Enhancing Robust Learning Through Problem Solving in the Genetics Cognitive Tutor. In Proceedings of the Thirty-fifth Annual Meeting of the Cognitive Science Society (pp. 2094–2099). Berlin: Cognitive Science Society.