Monthly Archives: September 2013

Raising the Bar for Teachers

The short version: New York state schools (the SUNY system) are requiring potential teachers to have a 3.0 GPA before applying to their teacher training programs.

I was initially against this. I thought, “We need more teachers! If someone wants to teach, let them at least try!” But taking a longer-term view, I think this is the right choice.

Teaching is a low-prestige job in the USA. The pay varies, but is often low. The hours are long, and most teachers work over the summer. Everyone respects someone who became a doctor or an engineer; not so much for someone who became a teacher. Everyone thinks it’s hard, but few people think they couldn’t do it themselves. “Those who can’t, teach.”

It’s easy to forget that here in the northeastern USA. Most of my friends have a healthy respect for teaching, and the pay for teachers (especially if you live in New York or Pennsylvania) is generally better than elsewhere in the country. Across most of the USA, however, teachers are respected individually rather than as a group.

If there’s one thing that unites the best school systems in the world, it’s that teachers and education are respected in that country. China, Korea, Finland, all of them have cultures where education is held in high esteem. Finland in particular has very high standards for its teachers, and actually turns away the majority of people who want to teach. If we want the US education system to get better, we need to work towards that sort of esteem. It’s a cultural matter as much as anything else, and this is a step toward that.

I’m interested to see what the short-term impact will be as well. I’m more than a little nervous about that. If there’s a big drop in applications, this initiative might fail. My guess is that someone has already run the numbers on it and, I’m just speculating, maybe only 25% of the applicants would fall under the 3.0 GPA minimum. In the long term, though, having people know that teachers need high scores to get in helps to raise their esteem of the program, and of the people who get through it.


College Ratings

It’s a linkstravaganza!

In case you hadn’t heard, every college except Harvard (and this year, Princeton) tells people to ignore the US News and World Reports ranking of universities and graduate schools. Prep schools tell people to ignore their high school rankings. Educational organizations are, in general, terrified of being compared to each other, because they know that someone will come out on top. Part of the rationale typically given is that each school is a unique and special snowflake that should be considered on the merits of how it fits with the individual student. Having seen a lot of schools, most of them are not. Don’t get me wrong, there are some unique schools out there (Sudbury, Hyde, Crane Union), and I love them for it, but most of them are not. Schools are indeed comparable entities.

Folks seem likewise terrified of a new initiative from Obama to rate schools and make financial aid dependent on that rating. I suspect this will not be a numerical ranking, the way USNWR does things – my guess is that they’ll give out A-F grades (because, ha ha, school joke, get it? so funny amirite?), but I could be wrong.

I’m of two minds about this. First I think that it could be a useful thing to do, and might encourage colleges to step things up. Then I remember that every system can be gamed, and that teachers and administrators are no better than students when it comes to gaming the system. I guess it’s another try-and-see sort of thing; I’m just worried as to whether a trial period and further reflection are actually built into the plan.

I sort of wish that things like the NEASC accreditation ratings could be used instead. They’re less a “how good is your school” rating and more of a “does your school actually do what you say you do” rating. They’re in individual categories rather than one overall score, and it’s quite a detailed report.!

Machine Learning in Online Environments

Machine learning techniques (including such things as Bayesian knowledge tracing and its successors) are a big deal in online education these days. You might not see it implemented right now, but unless things go very wrong, you’ll see it in the next five years. It will definitely show up behind the scenes; hopefully it will also show up as a major explicit feature.

The typical goal is personalization of learning, with the assumption (which seems fairly warranted to me) that such personalization can improve the learning process.

If you aren’t quite sure what could be done with machine learning and personalization in a course, let me give a few examples.

  • If Student S got the last three problems of this type right on the first attempt, don’t give S any more of this type of problem for a week.
  • If S hasn’t seen a problem of a particular type for three weeks, add one to the next assignment. If they get it right, mark them for mastery in that topic.
  • If S has mastery of a topic for which the current topic is considered a prerequisite, give S half the number of problems of this topic. If they get some wrong, give them the full set of problems and weaken the “prerequisite” connection between the topics.
  • If the average score on assignment A1 is too low, add assignment A2 to the course.
  • Randomize the order of resources R1 through R4 and see whether one particular order works better in terms of later performance.
  • Feed each student problems P1 through P5, but randomize who gets P6 through P10 so we can see which ones are most effective at helping students learn. Then fix up their grades afterwards so we aren’t costing anyone points.
  • If the successful guess rate is high for a problem, remove it.
  • Find out which categories of students are helped most by which sorts of exercises.
  • Adjust the wording of your resources to match the reading level of your students, or to push them toward higher reading levels.

All of these are fairly simple rules. You can work with individual students, with groups, with sections, with whole classes or even cohorts. You can recommend problems, book chapters, group member changes, even tell people which courses they’re likely to pass. Hell, some schools do that already.

The major barrier to including things like these in existing online education systems is that most people don’t know about them yet. Education researchers generally do, and people in AI research, but not the folks who program online education systems. There’s also a “just get it working” ethic that is fairly common in online education right now. That makes this a second- or third-generation technology – but a hell of a promising one.

The Student Loan Business

There is an interesting phenomenon that happens in the for-profit college business. As soon as a school qualifies for Pell grants, its tuition goes up… by the average amount of a Pell grant.

It will be interesting to see if there is a reverse effect when student loans become harder to come by. I sort of doubt it, but we might find out.

Pathwright follow-up

Two weeks ago, in my “Make your own online course” post, I mentioned a site named Pathwright. Over the weekend I got the chance to talk with Mark Johnson, one of the folks who created the site. He was interested in integrating some components from edX into their system. I, having worked with edX, was making some suggestions as to what components might be useful. Mark had some good ideas right off the bat, going for the pieces that made edX more modular and expandable.

It was also interesting to hear about their particular strategy. Pathwright definitely has a specific niche in mind. Let the big players gather their tens of thousands of students – these folks have a different approach. They’re specifically after the custom-learning market in a way that would be difficult to reach with automated learning systems.

Mark’s example for me was how he and his brother (also on the team) were both homeschooled, and how their mother saw them really getting into HTML at age 12 or so. She recognized that they were interested in it, saw that it might be useful in their futures, and let them do it as part of their science curriculum.

That’s the sort of decision that is very, very hard to automate. That’s like a Strong AI kind of decision right there, not the simpler machine learning that people are talking about incorporating into learning systems in the next 10 years. Don’t get me wrong, machine learning has the potential to absolutely revolutionize online learning, and I should talk about that some more on another day. But saying “Hey, let’s change up the curriculum because of what you’re interested in” is a whole other level. That takes not only a bit of courage, but a bit of foresight as well, and a willingness to take risks. I don’t know of anyone who is trying to build risk-taking into their machine learning.

Pathwright’s approach is to make every course adjustable, at the level of giving different assignments to individual students. It’s incredibly time-consuming compared to running a MOOC, so you can’t scale it easily, but scaling isn’t what they’re going for. It’s nice to see folks who are really going after a particular niche.

Common Core and Vocab

A quote from a friend of mine, who would prefer to remain anonymous:

This common core vocabulary book is literally ruining my classes.

Normally, the study of vocabulary is integrated into the books we are reading, so I don’t have to interrupt the study of the book we’re on to do vocab exercises. Also the exercises are decidedly NOT based on sound classroom practice. They require students to primarily fill in the blanks, rather than think about and use the words. Their quizzes reflect this deficiency: they use the words in an overly literal way related to the definitions, that are frequently awkward, incorrect, and not sound usage of the words. Moreover, as I’m sure we all remember from our vocab exercises growing up, we rarely remembered any of those words after the quiz. It encourages the “fill and dump” rote memorization. What we should be doing is having students read hard words, try to figure out the sentences, and then use the words in new ways. Using something sticks.

Anyway, the end result is that they’re not using the words, I’m wasting 1 and a half days per week on this, and not making any progress on the stories we’re reading, or the units we’re in the middle of. I can’t teach the way I want to, like by actually engaging them, because I’m sitting there with the book out drilling them on fill-in-the-blanks like I’m in a pre-1900’s schoolhouse with a grammar book.

Given the Common Core’s design considerations for ELA, I’m not sure why there even is a vocab book associated with it. It seems to almost run counter to their stated goals. I’m not saying that I’m surprised by such an implementation, just disappointed. More reasons why I wouldn’t want to teach in a public school.

Adjuncts and Teaching

An interesting new study:

This would not be unexpected. What’s interesting is that this actually contradicts previous research on the topic, in which there is typically no relationship found between research obligations and teaching performance:

The older studies tend to focus more on student evaluations, which may explain part of the difference. The newer one focuses on learning as measured by grades in that and future courses.

The newer study also focuses on just one institution, Northwestern. This may actually be a strength for the paper, as it may be possible to create an intervention study elsewhere. If there’s a school where the relationship has been shown to be null, then that school could adopt Northwestern’s practices and see if they can get a positive shift.


This also brings up the issue of adjunct pay, which is typically atrocious, and benefits, which are typically nonexistent. I imagine that a school could recruit teaching faculty much more effectively if they were willing to pay them what they’re worth. That’s the idea behind the Teaching Fellow positions in the UK, and one of the reasons the UK is so attractive to me right now.

Adjunct Infographic


Teaching Children vs. Training Adults

Most of my teaching experience has been at the high school and college levels. I taught middle-schoolers for a month or so one summer, and I’ve never taught high school. I’ve also done teacher training, tutor training, and radio station operator training for adults.

I’m looking at a few positions right now that are more adult-oriented, and so I wanted to check up on the best practices for adult education. After all, if I believe in the effectiveness of research, I should really see what it has to say.

A lot of places cite Malcolm Knowles‘ work from the ’70s, and talk about andragogy (as contrasted with pedagogy). To boil some of this down, there are a few principles that are generally agreed upon when it comes to adult learning. Things like, “Adults respond better to internal motivation,” and, “Adults appreciate being involved in the planning of their education,” and, “Adults want to know the relevance of the things they learn.” They want practical projects. They want respect in the classroom and outside it. They bring a set of experiences to bear on their learning.

In reading this, I have become frankly somewhat terrified about what we must think about children as learners.

Do people honestly believe that children, be they in grade 2 or grade 12, do not desire respect? Are there educators out there who think that high-schoolers don’t care about the relevance of what they learn? That they are better motivated by external forces than by internal interests and desires? Reading this stuff was painful, not because I found that any of it seemed wrong to me (it seems quite accurate), but because of what it implied about how children’s learning was, and probably still is, viewed.

Try teaching high school without addressing the relevance of the material you teach. Try doing it without a respect for your students. Try teaching things with no practical benefit, with no regard for the students’ prior knowledge, with no care for their own motivations. On second thought, don’t try any of that, because it would make you an awful teacher and your students would hate you.

I have no doubt that there are differences between primary school education, high school education, and adult education. This list of adult learner attributes? They’re not it. I’m still looking. If anyone has some good research-based suggestions for me, I’d love to hear them.

Crowdsourced Grading

The beautiful dream of every teacher: never having to grade. After many years of course design, I think that’s one of my main concerns when I create a new course. I say to myself, “I know I can teach the students well, but how can I do it with the minimum amount of grading?”

Grading is a pain. Grading is why a teacher’s day does not end when school ends. Grading is what makes it feel like your days just. Do. Not. End. Oh, classroom management can be a pain, and there’s a lot to be said for teaching students who genuinely want to learn, but six hours of grading on a Sunday can wear you down just as bad as a non-compliant classroom. I’ll look at basically any way to decrease that load.

The Crowdgrader tool mentioned in the above link is fairly simple – teachers create “assignments” (which means they write some instructions), and students upload their completed submissions, and then grade each other’s work, and then rate each other’s feedback. A nice touch is the ability to do group submissions.

I’ve done a few peer grading activities in the past, all of them in the classroom. The less structured these activities were, the worse they came out. When I set up a clear rubric, peer grading generally went very well. Students especially seem to appreciate rubric-based activities when they know that you’ll be using a similar rubric to grade exams. It’s always nice to open up the black box of the classroom so that students know how things work. In more technical or creative fields it can also be nice to see how your classmates approached their work.

In fact, I think that might be the most useful thing about all of this. Forget using it for courses where most things can be boiled down to right and wrong answers. That’s just as big a waste of student time as it is a waste of teacher time. Let’s push for this in more creative, more demanding courses where the approach can be just as important as the final answer. Let’s use this when there are multiple paths to the right answer, or when we’re in that fruitful area between purely objective and purely subjective, where justification and rational argument really count.

Using this in a MOOC would be another thing entirely. The ratio of participants to lurkers in MOOCs is so low that if you assigned everyone’s work to five other people during the first week of class, you’d get 1/2 of a person responding. However, later in the course, or with courses that have better “attendance” rates, there’s a lot of potential in this tool.