Upcoming Events
1. Call for Nominations
The National Assessment Governing Board (NAGB) is currently seeking qualified individuals to serve as Board members to help govern the National Assessment of Educational Progress (NAEP). One of the three seats designated explicitly for Testing and Measurement Experts will become available this year. Current NAGB and NCME member Scott Marion is eligible for reappointment, but NAGB and Secretary Cardona welcome new applicants. Board members serve a four-year term, with the potential for reappointment for a second term.
In 2019, as part of an NCME initiative to increase its impact on educational policy, the NCME Board began nominating members for consideration for the Testing and Measurement Expert seat on NAGB. To date, NCME members Suzanne Lane, Scott Marion, and Guillermo Solano-Flores have served in this role.
The NCME Board will be nominating candidates again this year. NCME members who are interested in being considered for nomination should send their vita to ncme@ncme.org by September 30, 2024. Letters of support are welcome but not required.
Nominations for NAGB are due by November 1. Those interested in self-nominating or nominating others independently can find additional details by visiting NAGB's website.
For more information, please contact the NCME Board at ncme@ncme.org.
2. Enhancing Automated Scoring Engines with Generative AI: A Dual-Study
October 9th, 2024 | 4:00 PM ET
Presenters: Justin O. Barber and Edward W. Wolfe, Pearson
Automated Scoring Engines (ASEs) are revolutionizing the scoring of constructed response (CR) assessment items through artificial intelligence. A major challenge in developing ASEs, however, is the limited number of training examples, particularly for infrequent score categories, driven by high scoring costs. Data augmentation, particularly using generative AI, is emerging as a solution to this issue. In our first study, we investigated multiple data augmentation methods and their impact on ASE performance. Simulated CR responses were created using these techniques to expand training datasets. The second study specifically focused on generative AI's effectiveness for data augmentation, evaluating ASE performance using metrics such as score point recall and quadratic weighted kappa to measure agreement between human and ASE scores. Our findings show that ASEs trained with generative AI-augmented datasets can achieve performance equal to or even surpass human raters. This presentation will delve into the methodologies, key results, and implications of both studies, highlighting how generative AI can enhance the reliability and accuracy of automated scoring systems.
Join us for the AIME meeting on October 9th at 4:00 PM ET!
Zoom link: Join here
Meeting ID: 817 2844 3502 | Passcode: 649132
To learn more about AIME, visit ncme-aime.org or sign up at NCME.org. We hope to see you there!
Chris Ormerod, Co-Chair | John Whitmer, Co-Chair | Maggie Beiting-Parish, Secretary