News

net.wars: The personal connection

This week my previously rather blasé view of online education – known in its latest incarnation as MOOCs, for massive open online courses – got a massive shove toward the enthusiastic by the story of 17-year-old Daniel Bergmann's experience in one such course in modern poetry, a ten-week class given by Al Filreis at the University of Pennsylvania and offered by Coursera.

How Daniel, the son of two of my oldest friends, got there is a long story better told on Filreis's blog by Filreis and Daniel himself. The ultra-brief summary is this: Daniel is, as he writes to Filreis (reproduced in that blog entry), "emerging from autism"; he communicates by spelling out words – sentences - paragraphs – on a letterboard app on his iPad. That part of the story was told here in 2010. But the point is this: 36,000 students signed up for the class, and 2,200 finished with a certificate. Daniel was one of them.

The standard knocks on online education have been that:

- students lose most or all of the interaction with each other that really defines the benefit of the college experience; - the drop-out rate is staggering;

- there are issues surrounding accreditation and credentials;

- it's just not as good as "the real thing";

- there may not be a business model (!);

- but maybe for some people who don't have good access to traditional education and who have defined goals it will work.

To some extent all these things are true, but with caveats. For one thing, what do we mean by "as good"? If we mean that the credential from an online course isn't as likely to land you a high-paying or high-influence job as a diploma from Cornell or Cambridge, that's true of all but some handfuls of the world's institutions of higher learning. If we mean the quality of the connections you make with professors who write the textbooks and students who are tomorrow's stars, that's true of many real-world institutions as well. If you mean the quality of the education – which hardly anyone seems to mean these days – that's less clear. Only a small minority can get into – or afford - the top universities of this world; education you can have has to be better than education you can't.

The drop-out numbers are indeed high, but as The Atlantic points out, we're at the beginning of an experiment. The 160,000 people who signed up for Sebastian Thrun's Udacity course on AI aren't losing their only chance by not completing it; how you spend the four years between 18 and 22 is a zero-sum game, but education in other contexts is not.

In July 1999, when I wrote about the first push driving education online for Scientific American, a reader wrote in accusing me of elitism. She was only a little bit right: I was and am dubious that in the credential-obsessed United States any online education will be carry as much clout as the traditional degree from a good school. But the perceived value of that credential lies behind the grotesque inflation of tuition fees. The desire to learn is entirely different, and I cannot argue against anything that will give greater opportunities to exercise that.

At this year's Singularity Summit, Peter Norvig, the director of research at Google, recounted his experience of teaching Udacity's artificial intelligence class with Udacity founder Sebastian Thrum. One of the benefits of MOOCs, he said, is that the scale helps you improve the teaching. They found, for example, a test problem where the good students were not doing well; analysis showed the wording was ambiguous. Vernor Vinge, the retired mathematics professor and science fiction writer, at the same press conference, was impressed: you can do that in traditional education, but it would take 20 years to build up an adequate (though still comparatively tiny) sample size. Norvig also hopes that watching millions of people learn might help inform research in modeling intelligence. There's a certain elegance to this.

Of course in education you always hope for a meritocracy in which the best minds earn both the best grades and the best attention. But humans being what they are, we know from studies that prejudices apply here as elsewhere. In his recently published book, Oddly Nomal, John Schwartz of the New York Times recounts the many difficulties his gay son faced in navigating a childhood in which his personal differences from the norm sentenced him to being viewed as trouble. If on the Internet nobody knows you're a dog, equally, nobody has to know you're a ten-year-old boy wearing pink light-up shoes. Or, as in Daniel's case, a 17-year-old who has struggled with severe autism for nearly all his life and for whom traditional classroom-based education is out of reach physically – but not mentally.

In Filreis's blog entry, Daniel writes, "Your notion that digital learning need not be isolating is very right where I am concerned."

Norvig, from the other side of the teaching effort similarly said: "We thought it was all about recording flawless videos and then decided it was not that important. We made mistakes and students didn't care. What mattered was the personal connection." This is the priceless thing that online education has struggled to emulate from successful classrooms. Maybe we're finally getting there.

"Emerging from autism." Such a wonderful and hopeful phrase.