Personalized Computer Review Improves Student Performance
“Our research shows that data collected from a population of learners can be leveraged to personalize review for individual students, yielding significant benefits over one-size-fits-all review,” said researcher Robert Lindsey, a doctoral student at the University of Colorado, Boulder.
“And this systematic, comprehensive review can be integrated into the classroom in a practical and efficient manner.”
Their findings are published in Psychological Science, a journal of the Association for Psychological Science.
Lindsey and colleagues were interested in using computational models to predict the effect of spaced study on learning, but they also wanted to ensure the real-world validity of their work.
Collaborating with an eighth-grade Spanish language teacher, the researchers were able to collect data from 179 students over a semester.
The students were responsible for covering a new chapter of their book each week and they were provided with an online flashcard app that allowed them to practice new vocabulary and phrases as well as to review old material.
Unbeknownst to the students, the review material came in three different kinds of sets.
Some of the material was in a “massed” set, with questions drawn from just that week’s chapter. Another set of material was “generically spaced,” drawn from just the previous week’s chapter.
According to the researchers, massed review is typically preferred by students while spaced review has been recommended by past research in learning and memory.
In the study, however, a third review set was drawn from any of the chapters that had already been covered previously.
This review set was based on an algorithm that predicted which material would be most beneficial for the students to review.
Similar to the approach used by online retailers to recommend products, the algorithm incorporated data from all of the students to determine which material any particular student might need to practice.
Lindsey notes that teachers typically don’t have the time to set up a personal question set for each student, but the use of technology enabled this personalized review, yielding promising results.
On a cumulative exam taken a month after the semester’s end, personalized review boosted performance by 16.5 percent over massed review and by 10 percent over generic spaced review.
Importantly, personalized review proved most effective for material from the first few chapters of the semester — material that would have been easiest to forget after several months — boosting students’ scores by an average of two letter grades.
“A relatively modest intervention — roughly 30 minutes per week of strategically selected review — can yield significant benefits in long-term educational outcomes,” said Lindsey.
Importantly, the personalized-question set proved most effective for the first five chapters of the semester – those chapters that would have been easy to forget after several months.
These results are promising, the researchers noted, because they provide solid evidence for personalized practice over simple study strategies students and teachers have used in the past.
“It is surprising how resistant students generally are to review,” Lindsey said. “They see their job as to learn the week’s new material, and feel that explicit review of old material is getting in the way of their learning. This experiment argues otherwise.”
Based on the results of the study, the Spanish teacher restructured his own lesson plans the following semester to focus on cumulative exams.
Lindsey and colleagues plan to continue investigating which review strategies are most effective for improving students’ long-term outcomes.
Nauert PhD, R. (2018). Personalized Computer Review Improves Student Performance. Psych Central. Retrieved on August 5, 2020, from https://psychcentral.com/news/2014/01/22/personalized-computer-review-improves-student-performance/64857.html