Pandora for learning is music to my ears

By:

May 20, 2010

How I love this Freakonomics Radio podcast (click on Episode 1). It compares public schools to a bad radio station, which broadcasts one stream of music to everyone in reach, without the ability to customize based on a listener’s preferences. Steven Levitt points out that in contrast, the Pandora music service streams individualized playlists to its listeners, matching each listener to music that fits his or her preferences for melody, harmony, rhythm, instrumentation, orchestration, arrangement, lyrics and other traits.

Levitt argues that technology makes a recommendation engine like Pandora possible in an education setting as well. Instead of students funneling through a factory-like classroom, technology exists to allow for one-by-one customization of how students are taught, based on what works best for each student. Sound familiar?

A few aspects of Levitt’s comparison are particularly apt:

First, I like Levitt’s comparison of choosing music content to choosing a learning modality. Just as music lovers use Pandora to find new music that suits their preferences, what if all kids were right-fitted to the mode of instruction that would best teach them each concept? A theme of the New York City School of One is “Choose your modality.” Kids there learn through teacher-led instruction, virtual tutoring, independent study, online games, etc. based on a daily assessment of what is working and what is not for each child.

Second, Pandora is possible because beginning in 2000, a group of musicians began the Music Genome Project to categorize each song based on its most fundamental characteristics (melody, harmony, rhythm, etc.). By establishing these categories and then grouping songs by each trait, Pandora made a recommendation algorithm possible. Pandora’s methodology can serve as a pattern for researchers in the education sector who are thinking about the right taxonomy to apply to learning. As this taxonomy is perfected, an algorithm that matches students to the right learning system for each concept is within reach.

Finally, Levitt’s vision for customized education implies major public policy shifts. The job of educating children through individualized, continually improving learning plans will require an emphasis on frequent assessment and top quality data management. People will need to open to the possibility that for some students with some concepts, a learning modality other than traditional teacher-led instruction works better. As that truth is accepted, the nature of teacher contracts, job descriptions and classrooms will change.

Do you think that in the future students will choose how to learn using a Pandora-like algorithm? Do you agree with Levitt’s comparison?

Heather Staker is an adjunct fellow at the Christensen Institute, specializing in K–12 student-centered teaching and blended learning. She is the co-author of "Blended" and "The Blended Workbook." She is the founder and president of Ready to Blend, and has authored six BloomBoard micro-credentials for the “Foundations of Blended Learning” educator micro-endorsement.