Skip Navigation

Literacy

Grantees

- OR -

Investigator

- OR -

Goals

- OR -

FY Awards

- OR -

Research on and With Novel Educational Technologies for Comprehension

Year: 2002
Name of Institution:
University of Colorado, Boulder
Goal: Exploration
Principal Investigator:
Landauer, Thomas
Award Amount: $799,884
Award Period: 3 years
Award Number: R305G020027

Description:

Purpose: In this project, the researchers proposed to develop a series of computer-managed instructional activities to help students who enter the learning environment with small vocabularies acquire the larger vocabularies necessary for high-level comprehension. These new automated assessment tools can be embedded in ongoing reading and writing activities within an automated literacy tutor.

Project Activities: This project leveraged Latent Semantic Analysis, a machine learning method that simulates human understanding of words and text, that members of this team helped develop. These novel, computer-managed activities engage, assess, and exemplify precise use of words in connected discourse. The technologies were evaluated in laboratory experiments and field trials through the ongoing IERI Colorado Literacy project (CLP). Researchers carried out laboratory research to analyze the cognitive processes measured and affected, and guide design of new interventions.

Key Outcomes: As described in the publications below, the project team developed the following tools:

  • Automation to construct and instantly score multiple-choice Cloze tests.  This tool was used as a beginning point to develop the more innovative constructed-response Open-Cloze activity described below.
  • A constructed-response "Open-Cloze" activity and assessment in which learners supply, rather than choose, missing words.  This activity provides learners with practice that supports the acquisition of word meaning, as well as the assessment of the learner's knowledge of the meaning(s) of a word.  This tool proved challenging to develop, but by the end of the project they were able to obtain adequate system-human agreement. They then used this tool as a summative assessment tool across multiple words, and also provided general feedback to the students after they completed the Open-Cloze activity. This development is in many ways the foundational achievement of this project and holds large promise in future uses.
  • An automated "Meaningful Sentence" activity and assessment. In this tool, students were asked to write sentences requiring the words-to-be learned (an authentic school-based task), and the system instantly evaluated the degree to which the word was used appropriately in the sentence.  This was completed early in the project and successfully.
  • A revised and extended version of Summary Street . This product was evaluated through an NSF award.  The IES award supported the development of tools integrated into the current system.  These tools provided graphic feedback both to students and teachers as to the progress of their summaries.
  • Tools to support individualized choice of practice words and reading passages. This objective was not met as originally intended, in part because it was contingent upon successful attainment of the tool development described above.  The time it took to accomplish that development work took them to the end of their project period.  However, the team did begin to explore how to use Lexile ratings and publisher-provided grade leveling in the choice of words and readings.
  • In addition, the project attained several goals that were not originally proposed, but which are central to the long-term success of this set of studies. As the team developed these new tools, they realized the demands of scoring and feedback using the new tools required the researchers to develop and test a new model of language that better captures sentences that are similar in meaning but which vary in their syntax and structure.  This addition of a syntactic dimension to the semantically driven LSA model has important implications for future work and was highly innovative.

PRODUCTS AND PUBLICATIONS

ERIC Citations: Find available citations in ERIC for this award here.

Select Publications:

Dennis, S. (2003). An alignment-based account of serial recall. In Proceedings of the Annual Meeting of the Cognitive Science Society (Vol. 25, No. 25).

Dennis, S. (2005). An Exemplar-Based Approach to Unsupervised Parsing. In Proceedings of the 27th Annual Conference of the Cognitive Science Society (pp. 583—588). Hillsdale, NJ: Erlbaum.

Dennis, S. (2004). An unsupervised method for the extraction of propositional information from text. Proceedings of the National Academy of Sciences, 101 (suppl_1), 5206–5213.

Dennis, S. (2005). A memory-based theory of verbal cognition. Cognitive Science, 29 (2), 145–193.

Dennis, S. (2007). Introducing Word Order in an LSA Framework. In T. Landauer, D. McNamara, S. Dennis, and W. Kintsch (Eds.), Handbook of Latent Semantic Analysis (pp. 449—466). Mahwah, NJ: Erlbaum.

Doxas, I., Dennis, S., and Oliver, W. (2007). The Dimensionality of Language. In Proceedings of the 29th Annual Conference of the Cognitive Science Society (pp. 227—232). New York: Erlbaum.

Harrington, M., & Dennis, S. (2003). Structural priming in sentence comprehension. In Proceedings of the Annual Meeting of the Cognitive Science Society (Vol. 25, No. 25).

Kintsch, W., McNamara, D.S., Dennis, S., and Landauer, T.K. (2007). LSA and Meaning: In Theory and Application. In T. Landauer, D. McNamara, S. Dennis, and W. Kintsch (Eds.), Handbook of Latent Semantic Analysis (pp. 467—480). Mahwah, NJ: Erlbaum.

Landauer, T.K. (2007). LSA as a Theory of Meaning. In T. Landauer, D. McNamara, S. Dennis, and W. Kintsch (Eds.), Handbook of Latent Semantic Analysis (pp. 3—35). Mahwah, NJ: Erlbaum.

Steyvers, M., Griffiths, T.L., and Dennis, S. (2006). Probabilistic Inference in Human Semantic Memory. Trends in Cognitive Sciences, 10 (7): 327—334.