How Big Data Is Taking Teachers Out of the Lecturing Business | Hacker News Reader

How Big Data Is Taking Teachers Out of the Lecturing Business

scientificamerican.com

27 pointsm_class17 comments

The part of the article that alarmed me the most was that the state of Arizona has reduced state funding for the university by 50% over the past 5 years....

Wow. The baby boomers really are gutting the system that gave they benefited from.

It's worse than that. There are three public universities which receive funding: ASU, University of Arizona, and Northern Arizona University. So it's not just one school. And Arizona schools may not be front page news, but these are institutions with big contributions to Mars rovers, astronomy and biotech. For instance, Lawrence Krass works at ASU.

It's really tragic how little Arizona higher education is supported, given their accomplishments.

It's even worse than that. States are not only reducing their funding, but the funding they do provide is nowhere near the operating requirements for major universities. Consider my alma mater, the University of Virginia, which receives about $8500 in funding per full-time student from the state (so roughly $150m/yr). This is <6% of their $2.6b annual operating budget[1]. Compare this to the 45% of funds that comes from health center patient revenues, or the 10-11% that each come from research/grants and gifts.[2] Pretty ridiculous. Spending is out of control, of course, but even not-for-profit schools are competing against each other for talent & grants, and the best ways they seem to know how involve huge infrastructure & technology investments, not to mention extra perks for students/faculty/staff.

[1] http://www.virginia.edu/finance101/state.html [2] http://www.virginia.edu/finance101/budgeted.html

I don't think it's just Arizona that's pinching the public universities.

http://www.insidehighered.com/news/2012/01/23/state-funds-hi...

New Hampshire had a 41% one year drop!

In some ways I like it as I've always learned better on a self-paced system. But no computer will ever inspire anyone to reach outside themselves and become someone new. The best teachers are the ones who move their students in directions they never even considered. There is no big data algorithm for that.

Well, lets do some abstraction here and extract the behavioural and cognitive elements from the best teacher-student interaction. Then we dump the data and process it using some tweaked ML algorithm and test the models on hundreds of thousands learners (tens of thousands is normal for a Coursera MOOC). Then using monitoring of social media students and their work profiles we obtain data about their career changes. Voila, computer inspired someone to become someone new.

I think the trend will be differentiated by goals... There are core classes with known material that people don't look to for inspiration: entry level math, accounting, etc. In the right circumstance, this should free up the teachers to inspire for courses that do require more intense discussion and guidance. For example: philosophy, political science, lab science, etc.

Basic math has plenty of room for inspiration, from the Fibonacci series to the combinatorics of counting to the concept of infinity. Sometimes those topics come up spontaneously in class with a human teacher -- not nearly often enough, but now and then. Are these cool and interesting topics that aren't on some Common Core standards list ever going to come up if students simply work through a computer program written by a corporate consultant and vetted by a committee?

Fair enough. Perhaps I'm biased by personal experience. Through an advanced track in high school and a college minor in math, I didn't have one inspiring math teacher. I got more inspiration out of Project Euler than I ever did from formal math teaching. While it would be great to have inspirational math teachers, perhaps I'm jaded when I say, "Why pretend?" And this is coming from someone who believes Math should be appreciate for beauty in addition to practicality. (Enough to put it in my name!)

I'm biased by personal experience too, having taught math on a college/university level for... ten years now. I have wanted to be the inspiring prof. I've done it for maybe 20 students. I feel the failure acutely.

I haven't given up yet but people (mathematicians, others) need to start really thinking about what education should be for and how we can achieve that. I love computers, science, (big) data. The folks implementing the current innovations are not mathematicians/chemists/economists, and the people choosing the current innovations are administrators or politicians rather than researchers or practitioners. I am not willing to cede control to these forces without a fight. What are we going to do to make sure that edtech is not just a way to expedite credentialing while making students dumb and efficient machines?

I appreciate where you are coming from, as you bring much more real experience into this than me.

Here's an open question, and I'm going to pull some numbers out of the air to just make a point...

Let's say that you have a given math department with 10 faculty charged with education of 4000 students, and your costs are fixed.

Which of these scenarios is best? 1) Each of the 10 teachers teaches 4 classes of 50 people per year. Each student gets 2 ((10450)/4000 = 1/2 per year) large courses over 4 years.

2) Each of the 10 teachers supports 1 large online classes of 400 people each, and teaches 3 classes of 33. Each student gets 4 impersonal online classes ((101400)/4000 = 1 per year) and 1 smaller class per year ((10 * 3 * 33)/4000 = 1/4).

The #s are made up, but the idea is that with a fixed amount of resources, you can have more small group classes with experts if you're willing to accept very large online classes. Or - you can get a lot more classes for the same cost if you're willing to take them online.

Again, this isn't a panacea, and doesn't work for every subject. It also doesn't increase creativity, or encourage research.

That's why for this class the teacher was a 'roving mentor'. The students still can use him for motivation and inspiration (if they so choose). The computer component takes care of the knowledge transfer. I go to WGU, and this is how all of their classes are. All resources are online (books, videos, practice quizzes, etc). If I have a question, I email the class teacher or ask my student mentor.

Well, there's a computer game[1] for people with depression that teaches them to change though patterns, and at least according to preliminary results, it success rates are much better than that of a therapy.

And, there's the field of persuasive technology that works on building technology to help people change behavior. According to wikipedia, there are numerous studies showing changes in health behaviours.

So i wouldn't rule out computers inspiring people(which seems like an interesting problem to think about/work on).

[1]http://pandodaily.com/2013/06/25/a-game-thats-better-than-th...

Maybe not, but that's why online and software based learning needs to be made exemplary. Teachers should be pooling their notes and putting together scripts that people with excellent charisma can execute on camera. Casually filmed lectures will rarely if ever be as inspirational as a powerful teacher can be face to face, but high production movie-like videos might just stand a chance.

"Sufficiently advanced testing is indistinguishable from instruction. In a fully adaptive classroom, students will be continually assessed, with every keystroke and mouse click feeding a learner profile. High-stakes exams could eventually disappear, replaced by the calculus of perpetual monitoring."

I see this, combined with more adaptive selection of learning goals, natural learner exploration, and a little bit of 'gamification' as the true future of education.

I didn't know folks had taken the khan academy model, given it the fancy name of 'Adaptive Learning' and raking in the dough with proprietary software.

The field of Adaptive Learning predates Khan Acadamy. In particular, one of the key predecessor fields - Programmed Learning [1] - predates Salmon Khan's birth.

[1] https://en.wikipedia.org/wiki/Programmed_learning

Reply on news.ycombinator.com