Using AI to Evaluate Pre-K Instruction
Summary
High-quality early learning programs boost children’s kindergarten readiness and effects can last into adulthood. However, defining and measuring quality is difficult and expensive. Classroom video observations are a classic method used by education researchers to understand how the instructional environment relates to student learning. Although videos provide rich context on what occurs in the classroom, they require a significant amount of human labor and time to accurately code, which can lead to delays in acquiring information. Moreover, early education researchers have found that constructs from standard protocols do not always predict children’s outcomes, indicating the need for improved construct measurement.
This project proposes a transformational approach to using AI with videos to advance current knowledge of Pre-K instruction. While prior education projects have examined automating the coding of classroom videos using a predetermined set of constructs drawn from standard coding schemes and “off the shelf” AI solutions, this project aims to develop an innovative, dynamic AI system that uses audio, visual, and spatial components of classroom videos to answer both classic and novel education questions that may not have been considered or available in other coding methods. If successful, the flexibility of this AI system will allow it to return query results beyond what is captured in existing coding schemes and answer questions that emerge as the education field evolves.
Have any questions?
Contact