We have, since the 1800’s and perhaps a bit before, lived and learned in the industrial age of education. Our learning institutions were built to produce minds fattened on facts in the feedlots, I mean classrooms. At the end of the process, we get the blurred blue stamp — pass, fail, grade A or not — that allows the consumers in the job market to decide our value.
That need for a more or less consistent means of evaluation for the consumers — colleges, businesses, education departments… has driven the need for tests. Unfortunately, the focus of these evaluations is not the students.
How traditional testing has failed students
Let’s consider the final exam — one of the most stressful, gut-wrenching moments of any semester. I’d argue that there is no more ineffective and possibly inhumane way to evaluate learning achieved during a class. It puts your reputation for learning prowess, intelligence, competence on the line.
Let’s add to the stressful and judgmental nature of exams, the question of whether they really do provide a good evaluation of learning. “Brain science” old and new shows that true learning occurs after repeated engagement with a topic or task. If a student ‘hears’ a fact in lecture and then crams it into their brain just prior to the final, the correctly retrieved answer on a test may be last time that student’s brain cells reference it.
If information is not embedded in long-term memory, is it truly learned? Can you name the capitals of all the states? We all ‘learned’ them in elementary school. You most likely learned about the Treaty of Versailles as a freshman or sophomore. Do you recall when it was signed or even why it was signed?Standard exams provide incomplete evaluation of student knowledge and fail to enhance learning.
Let’s add one final black mark against most exams and cumulative exams in particular. They are poor instruments of learning. As a PhD student, I would end up with piles of graded, but uncollected final exams at the end of the semester. I remain surprised at how unimportant learning the specifics of their answers were to the students, even though they could have served to reinforce their learning. Sure, some students were keen to get their hands on them and argue “fine points” or beg to bump up their grades. On the whole, the feeling is one of “file this class under ‘done’ and I won’t ever have to seriously engage with this material again.”
How technology can enhance learning through testing
But enough about what’s wrong. Let’s consider what can be right with testing if we leverage technology more effectively. For starters, technology is tireless and potentially infinitely scalable. Teachers, even when augmented with assistants, are limited. Sure, some universities have classes with hundreds of students in a single class, but consider the massive open online courses (MOOCs) that are reaching hundreds of thousands of students per class and allow repeated taking of a learning review (quiz) at 3 AM. Technology creates the possibility of a testing application that would allow any number of students to review their knowledge at any time.
In addition to removing the limitation of when and where students could evaluate their knowledge, they could do so more effectively. Students could quiz themselves to see where their knowledge was shaky and receive guidance on what content to review — with links to that material.
A testing application can provide spaced repetition, a way to test the ability to retrieve knowledge periodically in a way that reinforces knowledge retention. There are language learning and flashcard apps that provide this sort of tailored learning reinforcement today. Learning algorithms dredge ‘easy’ or learned topics out for review less and less frequently based on how often you answered correctly before.
Technology further allows immediate feedback on an answer — was my answer right or wrong and was this in part or in full? There is also the potential to immediately provide the answer with an explanation and then point to a resource for further review. This is again possible to do with traditional means to some degree, but not at scale or on demand.
Finally, tests or learning reviews (and educational content) that are periodically available for review by former students could prevent the erosion of learned material that happens as a course of time. When was that treaty signed again?
Putting the pieces together — tests that teach
I must take a moment to acknowledge that there are teachers and organizations that are doing right by their students. But even teachers that manually implement some of the features of student supporting tests suggested above are resource limited. An army of one is still — an army of one.
Acknowledging that there are some kinds of subjects that still do not lend themselves to automated evaluation, for those that are — math, language, history, coding… let’s consider how technology could still test students but also help teacher focus on teaching and students on learning.
First, testing would be posited on students achieving the learning goals stated in a curriculum. As students are exposed to those topics — presented by a teacher, read in a book, seen in a video, learning reviews/tests would start to evaluate student learning. Topics that are mastered based on tests could be checked off the learning goal list. Those that need review could allow the students to refocus on those areas in particular until mastered.
As the class progressed, earlier topics would occasionally resurface in reviews, strengthening their retention and allowing review if forgotten. By the end of the course, students would have had repeated review of the earliest material and should have it firmly embedded in their long-term memory. Rather than a final exam, students could continue to show their mastery of the subject, improve on it, and assure its long-term retention by occasional review. Less stress, more learning — tests we can all live with.
John Starmer, Director of Education @john_starmer