Collaborative Research: Enabling Robust Learning with Conceptual Personalization Technologies

Principal Investigator: 
Project Overview
Background & Purpose: 

This study has both a computer science and a learning science objective. The computer science objective is to create empirically-validated, domain-independent software algorithms that support robust student learning through personalized instruction. Robust learning is characterized by deep understanding of science concepts and successful transfer of knowledge and metacognitive skills to new domains. The learning science objective is to measure the effectiveness of conceptual personalization for enabling science learning, developing effective metacognitive skills, and promoting transfer across two domains: Earth science (atmospheric processes) and biology (natural selection).

Setting: 

University of Colorado at Boulder; University of Utah; Digital Learning Sciences (Boulder, CO).

Research Design: 

This is a comparative study designed to generate causal evidence. This project collects original data using achievement tests, survey research, design workshops, think-aloud protocol, and human annotation. We are targeting learners with high-school level knowledge of earth science and biology. For our study sample, we will draw upon undergraduate learners at our respective universities. Using this population for our studies reduces the complexity of recruiting participants and simplifies the logistics of running the proposed experiments. Our prior research has demonstrated that the science knowledge of non-science majors at the freshman and sophomore levels is not significantly more advanced than high school students.

Data from the computer science part will be collected and annotated by experts and verified for its consistency using a measure of inter-annotator agreement. This gold standard data will be used to assess the output generated by the algorithms using the standard measures of recall, precision and f-score.

Data from the learning studies will be analyzed using MANCOVA, where pretest performance is used as the covariate and the number of deep and shallow cognitive and metacognitive statements, and performance on the domain assessments (memory and application tests) are used as dependent variables. The between-subjects factor will be the learning technology used (CLICK personalized interactions vs. digital library use). Additionally, we will correlate the verbal protocol data with logged interactions with the CLICK system (e.g., average time per resource, number of switches between resources and work products, etc.).

Findings: 

Findings will be added as they become available.

Publications & Presentations: 

Ahmad, F., de la Chica, S., Butcher, K. R., Sumner, T., & Martin, J. H. (2007). Towards Automatic Conceptual Personalization Tools. In E. Rasmussen, R. Larson, E. Toms & S. Sugimoto (Eds.), ACM/IEEE Joint Conference on Digital Libraries, JCDL 2007 (pp. 452-461). Vancouver, BC: ACM Press.

Butcher, K. R., & de la Chica, S. (in press). Supporting student learning with adaptive technology: Personalized conceptual assessment and remediation. In M. Banich & D. Caccamise (Eds.), Generalization of Knowledge: Multidisciplinary Perspectives. New York: Taylor & Francis.

Butcher, K. R., de la Chica, S., Ahmad, F., Gu, Q., Sumner, T., & Martin, J. H. (2008). Conceptual Customization for Learning with Multimedia: Developing Individual Instructional Experiences to Support Science Understanding. In R. Zheng (Ed.), Cognitive effects of multimedia learning. Hersehy, PA: IGI Global.

Butcher, K. R., & Sumner, T. (in press). Self-directed learning and the sensemaking paradox. Human Computer Interaction.

Butcher, K. R., & Sumner, T. (submitted). What predicts learning with web resources? Relationships between prior knowledge, approaches to online learning, and learning outcomes.

de la Chica, S., Ahmad, F., Martin, J. H., & Sumner, T. (2008). Pedagogically useful extractive summaries for science education. In Proceedings of the 22nd Meeting of the International Committee for Computational Linguistics (COLING 2008). (pp. 177-184). Manchester, UK (August).

de la Chica, S., Ahmad, F., Sumner, T., Martin, J. H., & Butcher, K. R. (2008). Computational foundations for personalizing instruction with digital libraries. International Journal of Digital Libraries, Special Issue on Educational Digital Libraries, 3-18.

Gu, Q., de la Chica, S., Ahmad, F., Khan, H., Sumner, T., Martin, J. H., et al. (2008). Personalizing the selection of digital library resources to support intentional learning. In B. Christensen-Dalsgaard, D. Castelli, B. A. Jurik & J. Lippincott (Eds.), Lecture Notes in computer science, vol. 5173: Proceedings of the 12th European conference on research and advanced technology for digital libraries (pp. 244-255). Berlin: Springer-Verlag.