Variational Item Response Theory: Fast, Accurate, and Expressive
Saved in:
| Title: | Variational Item Response Theory: Fast, Accurate, and Expressive |
|---|---|
| Language: | English |
| Authors: | Wu, Mike, Davis, Richard L., Domingue, Benjamin W., Piech, Chris, Goodman, Noah |
| Source: | International Educational Data Mining Society. 2020. |
| Availability: | International Educational Data Mining Society. e-mail: admin@educationaldatamining.org; Web site: http://www.educationaldatamining.org |
| Peer Reviewed: | Y |
| Page Count: | 12 |
| Publication Date: | 2020 |
| Sponsoring Agency: | Defense Advanced Research Projects Agency (DARPA) (DOD) Office of Naval Research (ONR) |
| Contract Number: | FA865019C7923 MURIN000141612007 |
| Document Type: | Speeches/Meeting Papers Reports - Research |
| Education Level: | Secondary Education |
| Descriptors: | Item Response Theory, Accuracy, Data Analysis, Public Policy, Bayesian Statistics, Computation, Responses, Open Source Technology, Scoring, Inferences, Monte Carlo Methods, Grading, Computer Software, Second Language Learning, Second Language Instruction, Achievement Tests, International Assessment, Secondary School Students, Foreign Countries, Measures (Individuals), Language Skills, Language Tests, Vocabulary Development, English (Second Language) |
| Assessment and Survey Identifiers: | Program for International Student Assessment, MacArthur Communicative Development Inventory |
| Abstract: | Item Response Theory (IRT) is a ubiquitous model for understanding humans based on their responses to questions, used in fields as diverse as education, medicine and psychology. Large modern datasets offer opportunities to capture more nuances in human behavior, potentially improving test scoring and better informing public policy. Yet larger datasets pose a difficult speed/accuracy challenge to contemporary algorithms for fitting IRT models. We introduce a variational Bayesian inference algorithm for IRT, and show that it is fast and scaleable without sacrificing accuracy. Using this inference approach we then extend classic IRT with expressive Bayesian models of responses. Applying this method to five large-scale item response datasets from cognitive science and education yields higher log likelihoods and improvements in imputing missing data. The algorithm implementation is open-source, and easily usable. [For the full proceedings, see ED607784.] |
| Abstractor: | As Provided |
| Entry Date: | 2020 |
| Accession Number: | ED608051 |
| Database: | ERIC |
| Abstract: | Item Response Theory (IRT) is a ubiquitous model for understanding humans based on their responses to questions, used in fields as diverse as education, medicine and psychology. Large modern datasets offer opportunities to capture more nuances in human behavior, potentially improving test scoring and better informing public policy. Yet larger datasets pose a difficult speed/accuracy challenge to contemporary algorithms for fitting IRT models. We introduce a variational Bayesian inference algorithm for IRT, and show that it is fast and scaleable without sacrificing accuracy. Using this inference approach we then extend classic IRT with expressive Bayesian models of responses. Applying this method to five large-scale item response datasets from cognitive science and education yields higher log likelihoods and improvements in imputing missing data. The algorithm implementation is open-source, and easily usable. [For the full proceedings, see ED607784.] |
|---|
