AI, standardised testing and student outcomes

There are good predictors of academic success, like the GPA, but these and all other measures do not provide the comparability that ACT and SAT do. Now, however, there are less intrusive alternatives to standardized testing.

The coronavirus pandemic has suspended testing, which gives us an opportunity to rethink the existing paradigm. What emerges is the ability to replace summative testing – tests that summarize a student’s knowledge at a given point in time – with a formative approach to learning – microquizzes and other measures of student progress built into the learning process. Artificial intelligence is at the heart of this new approach.

My kids took standardized tests in April and got the results back in late October. This can be useful in maintaining state and school district accountability. But it is of no use at all for the students, because one cannot learn from it. Academic measurements should inform learning so that we get better results.

The key is having a quick feedback loop so you can find out what the student knows and the teacher can redirect the student in real time.

AI-powered formative learning is based on mastery learning, a concept that has been around for half a century. The idea is that students need to master the material gradually to create a solid foundation before moving on to the next level. But it never really caught on because the fixed teacher-classroom paradigm does not allow the necessary flexibility. Of course, some students take longer and need more attention than others to achieve their goals.

AI helps overcome this problem by doing much of the work that is otherwise done by teachers. An AI system can combine test or quiz questions with learning resources on a very granular level in order to make the learning process dynamic and personalized. When the system receives data on student performance, it makes predictions about the student and based on that selects learning resources – a short video or a short text.

“An AI system can combine test or quiz questions with learning resources on a very granular level in order to make the learning process dynamic and personalized.”

This is very different from using a traditional textbook where you always go through the same way from start to finish. Instead of offering a standard size for all students, the AI-supported mastery of learning is very individual. It adapts to the individual.

Once a student has mastered a particular learning goal, the student moves on to the next goal, with the system asking questions and offering learning resources for that goal.

This way we can optimize a learning plan for the student. Suddenly the evaluation itself becomes formative. It’s not just a conclusive value judgment from a test. It becomes part of the learning process in which we micro-assess the learner several times during a school semester or school year with little effort.

It’s an exciting offer. We thought it would be a decade before people were ready to accept this type of methodology. But Covid-19 accelerated everything. Since people couldn’t congregate in one physical place and everything is out of sync anyway, we are finding more and more people who are happy to adopt this method of evaluation.

AI can now predict a student’s test result with over 90% accuracy and identify their weaknesses and strengths with around 10 minutes of interaction. We can predict which questions a student will get wrong before they even try to answer them. We can even predict when a student will get tired and loosen up.

When such a system tells a student early on that their grade will be at the end of grade C but tells them what they should do to improve and they follow the recommendations and see their predicted grade moving to C + , this becomes a motivational factor in their learning. Many, if not most, students will work to improve the predicted outcome. When they finally reach the end of the course and achieve the outcome that the system predicted, that success becomes an incentive for further learning.

When students interact with such a system on a daily or weekly basis, and that system can predict student outcomes at any point in time and recommend the learning path they should take for optimal results, it eliminates the need for standardized testing.

A standardized test is a snapshot that people need to prepare for. Once the picture is taken, there is little follow-up. But if you have a system that constantly scores the student, you don’t need this snapshot. You know at all times where the student is on their learning path and what their likely outcome will be.

Mastery learning tells the student, “Here you are. Here’s what you’re weak at. And here is what you need to do to improve. “

If you are evaluating for the sake of learning, rather than evaluating for the sake of evaluating, the process changes completely. And, to be honest, it’s more fun for the learner.

If we as a community, society, and nation agree that the learning process is valuable and says something about the student, we should be able to replace that final summative grade with masterful learning. AI makes that possible.

About the author: Marten Roorda is the Chief Measurement and Learning Officer at AI and education company Riiid Labs. He was CEO of ACT from 2015 to 2020 and prior to that he was CEO of Cito, an international organization for education measurement based in the Netherlands, for 13 years.

Leave a Comment

Your email address will not be published.