Learning Engineering is the application of methods from the learning sciences to create transformative tools that support the challenges of learners as they learn.
This field is still in its infancy, especially when it comes to K-12 education, but over the past decade, and particularly with the onset of COVID-19 and the new demands on the expectations of teachers, there has been an emergence of interest in combining leading research and evidence-based principles from educational technology and the learning sciences to create engaging and effective learning experiences.
This project sets out to use learning engineering to develop a tool that helps teachers by automatically generating comprehension questions based on imputed passages of text to help assess students’ reading comprehension in a number of core literal and inferential comprehension skills.
We have used various frameworks to test outputs of questions with varying degrees of success. Through the process, we have identified challenges in successfully developing questions automatically for specific skills noting that the inferential skills are much more difficult to generate predictably.
We are now presenting the system outputs, questions/answers and distractors, to our expert teacher teams at Eyeread to offer feedback that is intended to be used by the machine learning system to improve the framework. This cycle of feedback looped back into the system will be underway until the end of August 2021. In the fall 2021, we plan to have tools built into the Eyeread software that allow for more data, in the form of teacher generated questions and feedback from teachers on the automatically generated questions from a larger group of teachers from across North America.
