On June 6th, 2013 Houghton Mifflin Harcourt (HMH) and leading adaptive learning company Knewton announce a partnership to offer personalized learning to K-12 students. HMH will use Knewton’s adaptive learning technology to power their math, reading, and other core subject area K-12 products, with their Personal Math Trainer being the first to launch. http://www.knewton.com/about/press/houghton-mifflin-harcourt-and-knewton-announce-pioneering-partnership/
The power of Knewton’s product has been proven at the higher education level as shown in the Arizona State University case study in the Fall 2011 and Spring 2012 semesters. “After two semesters of use with over 2,000 developmental math students at ASU, withdrawal rates dropped by 60% and pass rates went from 64% to 75%. Forty-five percent of students finished the course four weeks early.” (“The Knewton Platform,” n.d.). However, the K-12 system has a dramatically different student audience, especially in the primary grades, so can this system work at that level?
Adaptive Learning can take on many different forms with many systems focusing on single point adaptivity, performance at one point in time to determine learning path, or adaptive testing, uses fixed number of questions to determine proficiency. Knewton’s adaptive learning system is continuously adaptive, which responds in real-time to each individual’s performance and activity on the system. Their system takes into account both personal proficiencies and course requirements to continuously recommend course materials to meet each student’s exact needs. Taking into account the course requirements is a key component to this system as it ensures that all students can be evaluated based on the same criteria, but they can each take their own path to achieve the results.
Knewton does not actually create content or have their own platform, as they have partnered with other learning companies, such as Pearson, Houghton Mifflin, MacMillan, and many others to supply them with the Knewton engine to enhance their learning applications. “Knewton Technology consolidates data science, statistics, psychometrics, content graphing, machine learning, tagging, and infrastructure in one place in order to enable personalized learning at massive scale.” (“The Knewton Platform,” n.d.). Since their API has an invisible connection with the other company products, they can offer adaptive learning to almost any subject area, not just Mathematics.
To provide this continuously adaptive system, Knewton analyzes materials based on thousands of data points, including concepts, structure, difficulty level, and media format. Knewton’s recommendation engine also works on three key theories, Item Response Theory (IRT), Probabilistic Graphical Models (PGMs), and Hierarchical Agglomerative Clustering.
The Knewton engine is designed to be used by students over many different courses or grade levels, so a student learning profile is maintained within the system. When a student moves to a new grade or course, they start “warm” in that course as they already have data in the system. To maintain this profile the system uses two approaches, including spaced reinforcement, “in which new concepts or skills are absorbed while previously taught concepts and skills are reinforced,” (“The Knewton Platform,” n.d.) and retention and learning curves. “The Knewton recommendation engine needs to be able to take the degradation or diminishment of skill (or forgetting) into account.” (“The Knewton Platform,” n.d.).
As demonstrated in the DreamBox case, adaptive learning products can, in some cases, make the teacher feel marginalized and less empowered to shape the direction of the class. In their white paper, Knewton addresses this point while describing their Math Readiness instructor dashboard. “Using the dashboard, teachers can also see how students are performing in individual subject areas; which segments of material are the most and least challenging for students; and what kinds of patterns in both performance and activity emerge across the class. After multiple years of teaching the same course, teachers will be able to compare data from year to year; Knewton analytics will help them home in on useful information while leaving them free to interpret the results.” (“The Knewton Platform,” n.d.). If the new product being released by Knewton and Houghton Mifflin can provide enhanced data and class level reporting, can this enable the teacher to remain involved and develop lessons specific to their class weaknesses? Might some teachers need some tutoring in terms of reading visual representations of data and making decisions based on their interpretations of the data?
Adaptive learning systems have also been known to individualize learning. Knewton has suggestions in their white paper to address this issue, “An adaptive system can improve student engagement by weaving a social component into coursework. Knewton Math Readiness, for instance, provides a dashboard that allows teachers to group students who are working on the same material together. Using the reporting features, teachers can also arrange peer review opportunities and form groups of students whose abilities complement each other.” (“The Knewton Platform,” n.d.). Can this design feature overcome the individualized nature of the system and truly enable collaborative work? If students are grouped with other students working at the same level, is there a danger of creating performance-driven silos in a classroom?