In the information age, data will follow us from the time we first walk into kindergarten to well past retirement. As data is used to guide us in making all kinds of decisions, from what we consume to what health plan we follow, it’s also becoming a powerful tool in education.

As more schools and colleges use algorithms to determine a student’s path, the Amazon- and Netflix-style practice of data mining will soon be the norm in how schools and students operate.

But that might not be such a bad thing. Just as the two online behemoths — Amazon and Netflix — are able to use software to predict books, music, and movies you might like based on your past preferences, schools are using data to place students not only in their appropriate learning level, but even to recommend what subject to major in.

In K-12 education, it’s happening in classrooms and computer labs in both rich and blue-collar schools. In Covington Elementary, for example, the affluent Silicon Valley community where each fifth-grade student has a laptop and is learning math using Khan Academy videos and quizzes, teachers can track each student’s progress in real time on their iPads. When a student is stuck in one subject area, teachers can help the student one-on-one.

Likewise, at Rocketship’s Los Suenos Elementary school in a working class neighborhood in San Jose, teacher Alana Mednick can track her students’ progress based on how they score on their online computer games in their Learning Lab. And these examples are hardly rare these days.

On the college level, student data is being used for everything from recommending courses to picking majors. Austin Peay State University in Clarksville, Tenn., rolled out a program last year that uses data based on students’ majors, class history, grades, and similar student performance to help students decide on courses. (Students also still get advice from guidance counselors.) And according to the university’s provost, students who took the software-recommended classes received a half-point higher GPA than those who didn’t.

This spring, Austin Peay will take the experiment on a larger scale and use the computer algorithm to recommend a major for students who are undecided and for those who might choose one that’s not “right” for them.

But even before students apply to college, a company called Parchment will help them figure out which schools they’ll have the best chance to get accepted to. Parchment uses vast amounts of users’ data — GPA, SAT scores, extracurricular activities and so on —  to assess whether former applicants with similar profiles gained admission into certain schools. Parchment also says it can help point students towards schools that match their profiles, helping them find schools that are a good fit.

Getting that granular level of information to help guide decisions can help students bypass mistakes — but what happens then to serendipity, to the path that follows curiosity and experimentation what educators called “passion-based learning” that’s the antithesis to the data-driven definition of achievement and success?

That’s exactly what data can actually be used for, says Mark Milliron, chancellor of Western Governor University Texas, a nonprofit online university, said at the Big Ideas Fest last month.

Capturing data has turned into an expensive and convoluted proposition, he said. Schools, whether they’re K-12 or higher ed, will collect data, then share that information with certain faculty but not others (the latter of whom are then upset they weren’t included), then bring in administrators who are uncomfortable with the data they’ve seen and want to “make sure it’s clean,” at which point they hire auditors, then compare the numbers to other institutions to see how they rate against them. The auditors will make recommendations, like buying more software, which requires hiring consultants, sending out requests for proposals, then implementing the software.

“So what happened to the kids?” Milliron said.

If Amazon’s high-powered data engines can take just one split second to process data about a person that will allow them to make a good decision, why can’t that same power be applied towards education? he asked.

“What we’ve seen in the consumer world and healthcare world that’s made such a huge impact is what happens when you get data to the front lines,” he said.

To that end, new initiatives are being launched in different colleges to help students. A program called Course Signals at Purdue warns students who are at risk of poor grades and “facilitates intervention and support” that can help improve student grades by an average of one letter, according to the school.

Here’s how it works: The program uses information already available about each student to determine whether he or she is “at risk of failing or withdrawing from a course as early as the second week of the semester or quarter.  Based on the data, the solution displays a red, yellow or green signal to students and faculty, indicating a students status in a course in real time. A red light indicates a high likelihood of failing; yellow indicates a potential problem of succeeding; and green signals a high likelihood of succeeding.” Students receive an email with the progress report, along with suggested resources and recommendations from faculty on what to do next.

At WGU Texas, where Milliron is chancellor, the non-traditional online learning institution uses the same conceit for students, allowing them to create what he calls their own “learning journeys.” Though the school is technically based in Texas, only 1,600 of its 25,000 students are located in the state. What also makes this school different from others is that it’s “competency-based advancement,” which means students don’t have to take classes in subjects they’re already proficient in and progress at exactly their own level.

“The average student at WGU finishes in about 30 months as opposed to 60 months. So, they can get through in about half the time because many of them already have significant experience in their field and they can test out on competencies,” Milliron said in a recent Texas Tribune article. “They’d have to sit through classes that they could be teaching, which is often the challenge with adult learners.”






  • Gericar

    Apparently we are only the sum of what we can be tested on at any given
    moment…I thought being human meant that you had endless possibilities. You can
    be totally happy spending your life in a profession you didn’t learn about until
    you were in college or even later. How early does this “advice” on choosing a
    life path start? I think people are more than algorithms can prove. And if human
    behavior is anything it is unpredictable.

    In this endless testing
    model…doesn’t everything have to be based on the kid’s answers to
    questions…. and won’t the kid answer questions from their current experience
    and learning…so it seems that the “predictions” are based on only what the kid
    has been exposed to up to the point of the test; not on the vast array of
    information they might not currently have but that might be wonderful for them
    to stumble across in a class or somewhere. i.e. If your father is a lawyer…
    you are likely to come across a lot of legal kind of information just in your
    family life and you live with someone with a logical mind…thus, answer
    questions based on this knowledge or thought process….then the “tester” says
    you should head for a career in law or something like that because you are good
    at this…. Not a career in music because you don’t know anything about it…..
    at the moment…. Same old same old…. and…. is a better grade always the
    determinant of success? It is certainly a boon for schools in proving they are
    effective and spending the tax payer’s money well….but not always the end goal
    for the student. This is exactly why kids need to be exposed to a wide variety
    of experiences and be allowed to follow their fancy at times against the
    tide…and frankly encouraged to take classes in lots of different things.
    Leonardo Da Vinci in this model would have been given advice to take classes to
    be an undertaker because he was interested in dead bodies…. and there would
    never have been a Mona Lisa.

  • Garreth

    I don’t think test models limit any possibilities of a person, but completely agree that people are more than algorithms can prove. I think the algorithm programs are just part of the education curve in this day of technology. It’s hard to argue with results of some of these programs such as Perdue’s Course Signals. Where the program has been consistent with grade performance improvement at both the course and developmental levels. In some cases, students who were receiving Cs and Ds have pulled up half a letter grade or more to a B or C. As and Bs have increased by as much as 28% in some courses. So it seems to be beneficial. And it’s hard to believe that most of these students are complaining with their grade improvements due to the programs help or feel limited to any possibilities they may want to purse because of it.

  • Rockethound

    I can also tell whether a student is likely to fail or succeed in my classes by the end of the second week.  Students who don’t show up for classes, haven’t bought the textbooks, don’t log into the learning management system, don’t complete the first assignments, they’re at risk of failing.  They get multiple messages from me pointing out what they’ve missed and what the likely results of their actions are.  In nearly every case, they don’t believe it.  They’re sure that they just need a little more time, that things will be better next week, that they can catch up.  But they don’t.  I wonder if a message from a computer algorithm will be any more convincing.

  • Sjfone

    This system is a joke.

  • Katie

    See Western Governors University Texas Chancellor Mark Milliron’s talk on the Big Ideas Fest YouTube Channel at

Sponsored by

Become a KQED sponsor