Datapalooza = Yelp.gov?

This is a response to President Obama’s recent initiative regarding higher education costs, value and affordability, but I will start by talking about Yelp.

I don’t like Yelp.  I’m not as disestablishmentarian as Jaron Lanier (defining Wisdom of the Crowd as Mob Mentality), but I understand that multiple variables color the aggregation system on a company’s reviews, expertise perhaps one variable, perhaps not.  This concerns me.  The LA Times discussed Yelp’s business backlash yesterday.  Yelp continues to deny that failure to advertise with their website results in the site’s algorithm casting less favor on a business, but business owners are convinced that Yelp runs like the mafia, and see advertising as a necessary evil to cull favor with the site.

The most obvious solution to this standoff would be for Yelp to publish their algorithm, but that will not happen.  I imagine Yelp would say publishing such sensitive data would destroy their business, equating their success to the algorithm and not their developed branding and affiliation.  What publishing the algorithm would certainly do is show the numerous variables that dictate a company’s ranking and placement within a community of businesses, and those variables could spark discussions about the efficacy of crowdsourced commentary (for example, prolific users’ comments hold more weight than irregular users, meaning certain quantity is valued over potential quality).  Maybe Yelp is the mafia.  Or maybe Yelp runs its business to best ensure continued success through rewarding regular users (with higher weighted reviews) and advertising businesses (with featured spaces, regardless of how their ratings are measured), a perpetuating cycle less about quality of product and more about measurement of Yelp.

There are lots of points in President Obama’s College Affordability outline where I have a problem.

  • I think competency-based learning is a bad idea, whereas the policy memo celebrates it as cutting-edge pedagogy (it seems to enjoy a renaissance every 20 years)
  • The “use technology” remarks view tech as a magic elixir, a narrative not uncommon to historical and contemporary education discourse.  This is a terrible way to view technology in education, as it perpetuates its use as supplementing a status quo rather than transforming learning opportunities.
  • The focus on removing barriers to accreditation for edu startups and other service providers is a scary proposition, especially considering a great deal of student debt in America is due to the proliferation of for-profit colleges over the last 15 years.
  • Race to the Top and its foundation, No Child Left Behind, have not shown the statistically significant gains expected.  Creating a higher education version of it seems counter-intuitive.

All of these deserve more time and consideration, considering government policy is pushing in that direction.  My greatest difficulty with the President’s plan, however, is the datapalooza.  Part of this is because a word such as datapalooza, one that would be removed had I coined it for a research article, exists in public policy because of this initiative.  Joking aside, the implications of focusing so readily on Big Data and data aggregation again show a disconnect between what policymakers discuss as learning outcomes and what teachers and students consider to be learning outcomes.

What will the datapalooza look like?  Several points:

  • A collegiate ratings system developed in part on affordabilityaccess, value and outcomes, one that would provide the foundation for the government’s allocation of student aid as early as 2018
  • A $260 Million First in the World fund designed to spur technological innovation in higher education that can “yield dramatically better outcomes…and to develop new ways for colleges to demonstrate that they are helping their students learn.
  • Reducing regulatory waivers for experimental sites that focus on high-yield learning and potentially tie financial aid into how much a student learns.

The problems here are multiple.  First, if you are developing a Big Data system to rank colleges based on affordability and other issues, you are federalizing an aspect of governance and society that prior legislation has shunned.  Education is a right of the state, and while the Higher Education Act of 1965 and the GI Bill were federal initiatives designed to provide access to higher education for more people, both pieces of legislation explicitly state that their endeavors are geared at students and not at institutions, as it is not constitutional for them to do so.  I understand that increasing tuition and decreasing state subsidy is a problem that needs to be fixed, but an infusion of federal money based on a federalized ranking system is an attempt to have it both ways:  state-autonomous universities that receive money based on federal directives.

Will the algorithm of the Big Data system be public?  The problem I note with Yelp is in the privacy of the system — how can businesses trust a system that, from an owner’s perspective, seemingly harms them?  If the federal government looks at this ranking system from an Open Access/Open Education Resources perspective, then debate can focus on whether such a system is constitutional or whether the metrics focus on authentic learning or standardization and how various disciplines are weighted.  That’s the debate I wish to have, but if it remains secretive there will be enough distrust and animus at government to distrust any aspect of Big Technology’s oversight.

Diane Ravitch wonders how non-STEM courses will be weighed into a college’s value/access/affordability/opportunity quotient, and it’s a good question.  Bryan Alexander noted that the legislation does not mention STEM, but at the same time expresses surprise considering how much of President Obama’s education speaking lists STEM — one can assume that this legislation is supplemental to STEM focus, not a replacement of.

There are lots of opportunities to go down rabbit holes and predict dire consequences for this. If colleges are already puffing their chests out to gain an honorary designation from US News & World Report, won’t it only get worse if real $$$ is one the line in a federally mandated system?  Will the focus on future technologies and learning outcomes continue to gauge learning from a standardized, data-output driven model supported by somewhat archaic learning theory, or will the Department of Education reorganize their conceptual lens and view learning as contextual and environmental in assessing these outcomes?  This is just speculation, but considering the track record of education initiatives and government over the past several decades, trust in the federal institution is not at a high.

Mike Caulfield and I have gone back-and-forth on his blog in regards to the need for change, less about the positives of this plan and more about the problems of the status quo, problems Caulfield notes are more extreme than education pretends.  From that perspective, I can see how the potential of this policy brief can outweigh the existing negatives of a system designed to prop up US News and World Report rankings.  I think he makes a few wonderful points here, but I have a difficulty in substituting a private ranking system with no connection to subsidy with a federal program directly tied to subsidy and is ideologically against the purpose of the Higher Education Act (read the actual act; the Wikipedia page is very weak at the moment).

Advertisements

One thought on “Datapalooza = Yelp.gov?

  1. Audrey Watters (@audreywatters)

    Here’s what the Feds currently collect via IPEDS: http://nces.ed.gov/ipeds/about/

    Is this sufficient for them to make their ratings system work? (This is what generates some of their new College Navigator apps.) What’s missing? What data are colleges willing to hand over in order to create “better” and “more accurate” ratings?

    I totally agree about the dangers of a black-boxed algorithm dictating this. I’m hopeful that the open data folks at the Department of Education would resist this, but who knows…

    Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s