Resisting Big Data and Adaptive Software
Even if adaptive software can do some good, we should probably resist it.
Every time I hear cheerleaders of adaptive-learning technologies or big data approaches to education say they have the potential to be innovative, I want to channel my inner Inigo Montoya and say, “You keep using that word. I do not think it means what you think it means.”
“Innovation” and “improvement” are not synonyms. Innovation means “new,” not necessarily “better,” and every time we talk about something being an innovation, we should examine the possible negative consequences.
As a real-world example, I give you the self-serve soda station. In theory, this is a labor-shifting device that improves worker efficiency. In practice, it means that I have to eat lunch after having a sticky mist spray one of my hands. It also creates more labor for employees who have to clean up the ungodly messes left behind by customers who care little about the mess they leave at the self-serve soda station.
I thought about this distinction between innovation and improvement when reading a recent Scientific American article titled “How Big Data Is Taking Teachers Out of the Lecturing Business.” Written and reported by Seth Fletcher, the article examines larger trends towards computerized learning through the lens of specific practices at Arizona St. University.
In Fletcher’s words, “Arizona State's decision to move to computerized learning was born, at least in part, of necessity.”
That “necessity” was massive enrollment (70,000 students) combined with cratering state funding (down 50% over five years). Factor in what Fletcher calls “alarmingly high numbers of students showing up on campus unprepared to do college-level work,” and ASU finds itself running into the arms of adaptive-learning software provider, Knewton. As reported by Fletcher, by Fall 2014, ASU will be enrolling 19,000 students per year in adaptive learning courses.
Fisher does a credible job of articulating the points of view of both adaptive-learning backers and its skeptics, and the piece is worth reading in full.
There’s little doubt that we’re looking at innovation. The question is, in what ways can adaptive-learning be considered an improvement?
Here’s some facts about adaptive-learning/data-driven education that I think everyone can agree on, regardless of where one stands in the debate:
1. Adaptive-learning software is an attempt to simulate the already-existent best practices of good teachers: individualized instruction that takes into account the needs of different students.
2. With some limited exceptions, primarily courses where a high-degree of standardization is desirable, adaptive-learning software is not, and likely never will be as effective a teacher (or teaching device) as a well-qualified human being in a face to face setting.
3. Therefore, the push towards the use of adaptive-learning technologies is primarily, though not exclusively rooted in greater “efficiency” and lower cost, rather than improved quality. (While there are educational initiatives engaged in the genuine study of computer-mediated instruction, the ASU move was predicated entirely on questions of decreased cost and increased efficiency.)
4. Data-driven education in the form of No Child Left Behind testing has proven to be ineffective at improving student performance, even as measured by tests geared towards the specific kind of content-heavy instruction high stakes testing demands. (“Alarmingly high numbers of students showing up on campus unprepared to do college-level work.”)
5. We will primarily or exclusively see the adoption of adaptive-learning/data driven education in public colleges and universities subject to declining state support. (You won’t see Harvard putting its freshman math students in front of computer terminals.)
6. The primary economic beneficiaries of an increase in the use of adaptive-learning technologies will be private corporations such as McGraw-Hill, Pearson, and Macmillan who will receive their revenue via public monies and student tuition dollars transferred through colleges and universities.
Now, some thoughts that are perhaps more arguable.
1. “Big Data” is not putting teachers out of the “lecturing” business, and we should stop making comparisons between “personalized software” and large lecture courses in order to frame adaptive software as some kind of improvement. ASU is not using software in the place of 1500 person lectures. They’re substituting the software for teachers in what are traditionally small classes, otherwise, there’s no improvement in efficiency.
In the words of Fisher, “Arizona State administrators went looking for a more efficient way to shepherd students through basic general-education requirements —particularly those courses, such as college math, that disproportionately cause students to drop out.” (Emphasis mine.)
2. Students are not sheep and teachers are not shepherds. The choice of words above is Fisher’s, but when you have 50 instructors “Coach[ing] 7,600 Arizona State students through three entry-level math courses running on Knewton software,” the description is apt. The notion that this is an acceptable substitute for teaching – no matter how great the software may be - chills me.
3. There are, obviously, ways that the integration of adaptive-software into teaching will benefit student learning.
4. However, once unleashed, these forces will not augment, but replace human teachers. Witness what’s happening at ASU.
It chills me because we’re turning the work of education into a process where students sit in front of screens and execute programs. This is like something out of the dystopia of Wall-E.
Look at the photo adaptive-software provider Knewton itself chooses to illustrate its applications and ask yourself if this bears a resemblance to the kind of learning and life we wish for future generations.
As reported by Steve Kolowich in IHE, one of the goals of Knewton founder and CEO Jose Ferreira “is to create individual, psychometric profiles that would presume to say, with statistical authority, what students know and how they learn.”
I do not embrace a future where we are the sum total of our data. This is not the experience of life and living as I understand it.
4. I personally prefer the imperfect human teacher to standardized adaptive software. The notion that all of education can be standardized restricts progress. While we would like all teachers to strive for excellence, the messiness of individual instruction and variation in teachers is not a bug, but a feature, that is unless we think the ups and downs of life itself is a bug.
Using tools developed by the adaptive-software providers to assess learning is not a test of student development, but of how well they learned to interact with the software. It does not necessarily correlate to any kind of real-world application. It is a credential without utility.
5. The embrace of adaptive-learning software in the short run may seal the doom of public universities in the long run. In the inevitable cycle of innovation, when Pearson, McGraw-Hill and Macmillan can provide the same “education” for less, they will cut out the middle-man.
They are not our educational partners. They are parasites that will, once possible, overwhelm the host.
6. The defunding of higher education (and primary and secondary education) is not an immutable fact of existence, but a choice. We could, at least in theory, choose differently.
7. By pretending (not an accidental word choice) that software can substitute for humans we are making it much harder to choose differently. This is a form of capitulation.
8. The Bill & Malinda Gates foundation, authors of the “Adaptive Learning Market Acceleration Program,” may be the most destructive force in education. The Obama administration (Race to the Top) may not be far behind. They are accelerating the pace at which we outsource a public good (education) to private corporations.
9. The embrace of adaptive-learning software as a substitute for face-to-face instruction is a vehicle that furthers the exacerbation of inequities in education. Students with access to “elite” universities will continue to be taught by humans while the rest will be trained by computer software.
10. This is the most personally frightening paragraph in the entire article:
Long before that happens, generational turnover could make these computerized methods of instruction and testing, so foreign now, unremarkable… Teachers could come around, too. Arizona State's executive vice provost Phil Regier believes they will, at least: “I think a good majority of the instructors would say this was a good move. And by the way, in three years 80 percent of them aren't going to know anything else.” (Emphasis mine.)
11. That a high-ranking administrator at a leading public university thinks this way should scare the bejeezus out of us.
Are we really prepared for a world where students are going to equate teaching with software because they “aren’t going to know anything else”? Phil Regier is complicit in destroying his own institution, even as he must imagine he's saving it.
A university that is no longer about the human exchange of ideas and knowledge doesn't deserve the title.
12. I’ve probably just ruined any chance of future employment at Arizona State, but I’m okay with that.
Reasonable minds can differ on this. Feel free to call me all manner of names on Twitter.
Read more by
Opinions on Inside Higher Ed
Inside Higher Ed’s Blog U
What Others Are Reading