Student Data, Algorithms, Ideology, and Identity-less-ness


Warning: less than fully formed thoughts here...

Democratizing Ideologies and Inequality Regimes in Digital Domains

My friend, the amazing Tressie McMillan Cottom delivered a fabulous talk yesterday at the Harvard Berkman Center: “Democratizing Ideologies and Inequality Regimes in Digital Domains.” (You can find her notes and slides here, a live-blog from PhD student Nathan Matias here, and a collection of Tweets from the talk here; it’s my understanding that the center will be posting the video-recording soon.)

As the title of her talk suggests, Tressie’s work deals with ideology, inequality, and higher education, with a specific focus on for-profit education through the lens of race, class, and gender. A sociologist, Tressie is interested in “inequality regimes,” that is, the practices, policies, processes – the systems – that perpetuate and extend inequalities.

I want to pick up on a couple of the points that she made about ideology, inequality, and digital education:

She noted that the hype about “democratization of education” through MOOCs and other online efforts has conflated “access” with “information” and “information” with “education.” She observed too that while technologies are quite good at measuring tasks, they are much less effective at measuring learning. But as a result, learning is increasingly defined by the demands of data collection and by the affordances of the digital platforms. Learning is reduced to tasks, and the individual learner’s context – their cultural context, their place, their identity – is stripped out.

This works in part, Tressie argued, because the student these new (cough) digital technologies are designed for is a “roaming autodidact” – an “ideal, self-motivated learner,” “embedded in the future but disembedded from place.” Disembedded from place, disembodied – this erasure of context and of identity is easily a re-inscription a “universality” of the white European male. There’s no real need to construe education as aspiration for the “roaming autodidact,” Tressie argued, because they already “live in the future.” There’s no need for education as upward mobility for these students; they’re already there.

These ideologies permeate new digital learning technologies. They reward the “roaming autodidact” while always judging others as inferior. (A lack of "intrinsic motivation," for example.)

“You don’t get to scale” with the “roaming autodidact,” Tressie argued. Her work this summer involves conducting ethnographies with women of color pursuing PhDs at for-profit universities. They don’t roam, she quipped, “they roost.” Place matters. Identity matters. You don’t scale with technologies designed for the “roaming autodidact,” Tressie contended, because if you look at the demographics and the global demand for post-secondary education, you’re unlikely to find a lot of “Privilege Unicorns” there (another great phrase from Tressie’s talk).

Ideologies, Identities, and Algorithms

But I do think there are efforts “get to scale” with this disembodied and disembedded student as a model. Not through online courses or MOOCs, but via “teaching machines” – or more accurately and less book-promotion-y, via algorithms.

Knewton CEO Jose Ferreira was involved in a brief exchange on Twitter yesterday with some parent-privacy activists accusing his company, among other things, of collecting and sharing PII (personally identifiable information). Ferreira insists the company does not (although for what it’s worth, Knewton does say on its website that it collects PII if a parent or student consents).

But what caught my eye, particularly in light of just having heard Tressie’s talk, was this response from Ferreira: “We can help students understand their learning history without knowing their identity.”

Is such a thing even possible? (Is such a thing desirable?) Can you have a learning history and not have an identity? Aren't these inextricable?

Granted, in this Twitter exchange, Ferreira is attempting to address concerns about data collection and privacy. (And that discussion opens up a whole other can of worms I'll save for another day.) But I think his response echoes some of the ideologies that Tressie’s talk identified as pervasive in ed-tech: building a model, a profile, an algorithm that claims to be liberated from "identity," that is grounded in a tradition of liberal humanism that ignores the body, privileges the mind.

So what does it mean when we talk about an “identity-less-ness” learning? What does it mean to build “learning sciences” and “learning technologies” on top of this sort of epistemology? What does “personalization” mean if there’s no “personally identifiable information” involved? And of course, what are the ideologies of algorithms – the “secret sauce” – that underlie adaptive technologies like Knewton?

I can’t help but invoke Donna Haraway here in her famous 1988 (!!) article “Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective” (PDF):

"The only people who end up actually believing and, goddess forbid, acting on the ideological doctrines of disembodied scientific objectivity – enshrined in elementary textbooks and technoscience booster literature – are nonscientists, including a few very trusting philosophers…

Only those occupying the positions of the dominators are self-identical, unmarked, disembodied, unmediated, transcendent, born again. Knowledge from the point of view of the unmarked is truly fantastic, distorted, and irrational. The only position from which objectivity could not possibly be practiced and honored is the standpoint of the master, the Man, the One God, whose Eye produces, appropriates, and orders all difference. No one ever accused the God of monotheism of objectivity, only of indifference. The god trick is self-identical, and we have mistaken that for creativity and knowledge, omniscience even."

Is there a "god trick" in ed-tech? (I think Tressie's talk at Berkman suggests "yes.") More importantly, more urgently, is this "trick" being hard-coded, hard-wired into the infrastructure of our schools? Remember: we aren't simply talking about learning content; we're talking about learning technologies. And technology is infrastructure now.

What happens to bodies – particularly bodies of marginalized people – when they're submittted to a new knowledge regime that claims to be identity-less, that privileges identity-less-ness? I'm not talking her about a "loss of privacy" – indeed, Tressie higlighted in her talk that the women she works with want to be able to find and get to know one another and build their own support networks. Bodies matter when we learn; communities and affinity and situatedness matter; digital learning, even though some of it is "virtual," does not – or should not – change that.

Image credits: Serge Bares and The Noun Project

Tags: , , , , ,