2015 AAAS Annual Meeting (12-16 February 2015)

February 12 - 16, 2015

Next-Generation Learning and Assessment Environments for Science Inquiry Practices

Many rich, interactive, learning environments have been developed to help students learn science concepts and inquiry skills, including hypothesizing, collecting and interpreting data, and forming arguments. These capture vast amounts of process data -- such as clickstreams of students’ interactions within the software -- and work products, such as students’ drawings or written essays. The data have potential to provide authentic, rigorous, and nuanced assessment of students’ skills that is not possible with traditional multiple choice-style approaches. However, to date, assessment of this type has been challenging to scale because of both theoretical and technical limitations. Specifically, there is a lack of guiding principles for parsing, aggregating, and analyzing logs, and difficulty auto-scoring students’ open responses, due in part to the irregular nature of students' writing. This session features new learning environments and cutting-edge methods that address these assessment challenges, using techniques from educational data mining, natural language processing, and item response theory. Discussion will include future directions for this field, and interactive demonstrations of each learning environment will be available for attendees.