TEP Charter: Teacher Pay, Student Scores

Wikipedia describes The Equity Project as

The Equity Project (TEP) is a New York charter school opened in September 2009. The school will instruct students in fifth through eighth grades and will pay its teachers a starting salary of $125,000 per year, more than double the national average. Teachers must also work double the hours a typical teacher works.

Now Vox reports that This school paid teachers $125,000 a year — and test scores went up

Four years later, students at TEP score better on state tests than similar students elsewhere. The differences were particularly pronounced in math, according to a study from Mathematica Policy Research released in October. (The study was funded by the Gates Foundation.) After four years at the school, students had learned as much math as they would have in 5.6 years elsewhere: ...

... Teachers at TEP also get more time to collaborate and played a bigger role in school decision-making than teachers in other jobs. Teachers were paired up to observe each others' lessons and provide feedback, collaboration that experts agree is important but happens too infrequently. During a six-week summer training, teachers also helped set school policy.

The workload at TEP, where teachers also take on administrative duties and had an average of 31 students per class, is fairly heavy even with the extra pay. But the school also had more teacher turnover than usual. Nearly half of first-year teachers didn't return for their second year, either because they resigned or because they were not rehired.

Intriguing...and a hopeful sign generally. It should be noted, however, that I got to that report from the link at statistics-prof Andrew Gelman's blog on Time-release pedagogy?? - Statistical Modeling, Causal Inference, and Social Science

...there’s something strange here or i’m missing something obvious. That jump from 3-year impact to 4-year seems excessive.

That does seem really odd. Still, it's a hopeful sign.

Girls vs. Tests, Boys vs. Classwork

Or perhaps boys vs. work, generally; Why Girls Tend to Get Better Grades Than Boys Do - The Atlantic

girls succeed over boys in school because they tend to be more mastery-oriented in their schoolwork habits. They are more apt to plan ahead, set academic goals, and put effort into achieving those goals. They also are more likely than boys to feel intrinsically satisfied with the whole enterprise of organizing their work, and more invested in impressing themselves and their teachers with their efforts.

These days, the whole school experience seems to play right into most girls’ strengths—and most boys’ weaknesses.

On the whole, boys approach schoolwork differently. They are more performance-oriented. Studying for and taking tests taps into their competitive instincts. For many boys, tests are quests that get their hearts pounding. Doing well on them is a public demonstration of excellence and an occasion for a high-five. In contrast, Kenney-Benson and some fellow academics provide evidence that the stress many girls experience in test situations can artificially lower their performance, giving a false reading of their true abilities. These researchers arrive at the following overarching conclusion: “The testing situation may underestimate girls’ abilities, but the classroom may underestimate boys’ abilities.”

Perhaps we need two-dimensional grades.

Other School Analogies: Day Jail?

School isn't just about learning; it's also about signaling that you've learned, of course. Beyond that, it's about making society work, in such a way that parents can do what they need to do all day. Arnold Kling, in The Dark View of Schooling, notes

Bryan Caplan thinks that schooling is not about education. He thinks instead it is about signaling.
Bryan’s view is benign compared with John Holt.

society demands of schools, among other things, that they be a place where, for many hours of the day, many days of the year, children or young people can be shut up and so got out of everyone else’s way. ... That is an important part of what schools are for. They are a kind of day jail for kids.

I do believe that technological improvements can make signaling a lot easier, in K-12 as much as in How testing could transform higher education:

having a well-calibrated measurement would put a lot more focus on learning the material and much less on how you learned it. No one will care too much which book you read, which online class you took, or what exercises you did. Those are all just ways to learn. Second, this would truly democratize education—the self taught Bangladeshi is on equal footing with the legacy Harvard admittee.

If you want to know if a kid has grasped 10th grade math, start him on the Khan Academy 11th grade sequence and see how quickly he reaches "mastery". Of course you probably need humans to rate English composition, if not reading comprehension, but you can go a long ways without adult humans involved -- let ten kids write short essays to summarize some information, each of them answers ten sets of multiple-choice questions, and they get comprehension scores based on how well they answered, but composition scores based on how well their readers answered. (Is this a bad idea? Probably; I just thought of it. There are lots of possibilities.) The testing tools for this probably won't come very quickly, and they won't come at the same pace in all subject/skill areas, but they're coming. I'm less confident that we'll get a substitute for the "day jail for kids." Many parents will want to know that their kids are somewhere being supervised -- the sort of "off in the woods, probably building a dam" that went over well enough in the late 50s (when I went along with my elder siblings) is probably gone forever.

Or on the other hand, maybe not.

Cornell Notes

On being systematic: I wonder how many variations on Cornell notes there are, and how to identify which students should use which method...What We Learned Today: Cornell Notes

the note-taking page is set up in a specific way. There is a wide left column for questions, keywords, or cues. The wider right column is where the main notes are taken. At the bottom of the page, there is an area to summarize the notes taken on that page.
Step One: When you’re taking notes in class or from another source, take notes on the right side of the page. Don’t use complete sentences, and abbreviate what you can. Write down the key ideas, and pay special attention to things written on the blackboard, if you are using this for school.
Step Two: Soon after you are finished taking notes, on the left side of the page write questions, keywords, main ideas, or cues to remind you about the notes. This will reinforce the learning in your mind, and help you make it all clear, connecting related bits of information. It is also a good opportunity to see if you have any questions about the material yourself.
Step Three: Within the next day, on the bottom of the page summarize the material in the notes column.

Maybe everybody should do things that way, or in some other specific way. I am more inclined to suspect that everybody (including me, at age 60) should spend part of each year trying to look for (and measure) a better way, even if it's a way that wasn't better before.
Or then again, maybe not.

Update 2014: Noticed that the wired.com link had succumbed to linkrot, asked Google to find the article on wired.com, and replaced the link target with the "archive.wired.com" value.

Marginal Revolution University

My favorite economics bloggers, being professors, have started a Khan-Academy-style online university with one course -- Development Economics; soon to grow. A Blog Hopes Its New Online Course Will Be More Than Marginal | The New Republic

one of Cowen’s many, many driving questions: what is education really for—what really makes it valuable? Mainstream economic thinking, Cowen noted in a lengthy 2006 blog post that foreshadowed MRU, views education primarily as a signaling device....
But Cowen posited that education is most efficiently used as a tool for self-improvement in the broadest sense. “Education is about self-acculturation,” he wrote, continuing:

education gives you a peer group, a self-image, and some skills as well. Getting an education is like becoming a Marine. Men need to be made into Marines. By choosing many years of education, you are telling yourself that you stand on one side of the social divide. The education itself drums that truth into you.

MRU is ultimately aiming for a better actual education, not a better means of signaling. ... “You can think of this,” Cowen says, laughing for the only time during our phone conversation and only lightly, “as a marginal attempt—a marginal revolution, so to speak—to get education to be more about learning.”
...a professor from the University of Oklahoma will lead a course on Mexico early next year—and Cowen is encouraging readers/viewers to make their own videos and send them in, with Tabarrok citing Wikipedia as MRU’s model here. Its optimistic motto is “Learn, Teach, Share.” This ad hoc philosophy is central to MRU’s method—and is potentially its greatest weakness.

As long as they're providing a resource for self-education and a resource for teachers in classrooms, that should work. If they need to evaluate students, I presume they'll have to develop some kind of peer grading; I've been thinking about networks of Google hangouts for that purpose. (You get assigned to a series of Google hangouts, working with different people and validating each other's actual existence...that it's actually you taking that test, for one thing.
Or then again, maybe not.

Investing in Education?

It's common to claim that when we spend more and more on education, the results tend to indicate that we've wasted our money. One more depressing set of graphs along those lines comes in looking at the SAT trends over the years Squandering Assessment Test

There are factors that make comparing year-to-year SAT scores imprecise. But the trend clearly reinforces what we should already know: we get almost no return for our education “investment.”

Well, maybe. Or maybe not. But this is data.

"American Graduate Week"

This week is apparently American Graduate Week, following up from the American Graduate Day Announcement

Unless graduation rates increase, nearly 12 million students will likely drop out over the next decade, resulting in a loss to the nation of $1.5 trillion in lost wages and increased social costs due to crime and healthcare. Among students who do graduate, one-third need remedial courses in college and far fewer will go on to earn a college degree. Yet, more than half of all new jobs in the next decade will require some postsecondary education.

As that phrasing suggests, the "crisis" is not an increase in dropout rates, but rather an increase in the need for the qualifications which graduation is an attempt to represent. Really we don't need high school graduates; we need adults whose education suffices for a life of production, consumption, creation, family, and citizenship. (Hm, is that all?) So we have an effort by the Corporation for Public Broadcasting which the NYT reports as Public Broadcasting Takes Role in Improving Graduation Rates

While graduation rates have inched up in recent years, nearly 25 percent of students over all drop out.

And Ray Suarez of PBS imagines our current system as a mechanism for generating dropouts, a mechanism in need of sabotage; Call is Out to Sabotage the Dropout Crisis | PBS NewsHour

The National Council of La Raza, the nation's largest Latino civil rights organization, helps run a charter school in Los Angeles? Who knew? A not-for-profit called BuildOn takes students who are on the verge of dropping out and gives them international service learning experiences, of the kind that has become commonplace for upper middle class kids, but is much less available to poor kids. Harlem RBI takes teens with school troubles, puts them on baseball and softball teams, and requires tutoring, workshops, and homework help to keep a player's average high enough to stay on the team.

The athletics approach is one I have to keep reminding myself about: it's not one that appealed to me as a kid. But it's not exactly a newfangled notion. In Plato's world,

Both boys and girls receive the same kind of education. Elementary education consisted of music and gymnastics, designed to ...create a harmonious person.

And I think archery and riding went in there too, although I also think story-telling comes even earlier, in character formation. But anyway athletics is important, somewhere in there, even if it might be minimized for people like me. But then, as Plato also put it,

“Do not train a child to learn by force or harshness; but direct them to it by what amuses their minds, so that you may be better able to discover with accuracy the peculiar bent of the genius of each.”

Fair enough. I suspect that somewhat greater attention to (and measurement of) what amuses kids' minds would do quite a bit towards raising graduation rates, and even the rate of achievement of any particular level of qualifications.

Or then again, maybe not.

Teacher Evaluation

Economics Nobelist Gary Becker seems to approve this summer's new teacher-evaluation research; in Good and Bad Teachers: How to Tell the Difference he says

Teacher unions all over the country have fought against using performance-based measures to evaluate teachers, but the unions are gradually losing this battle...
...the criterion used in evaluating teachers by many school systems and also by academic articles on school reform is the value added (VA) by teachers to student performance; namely, the improvements in students’ test scores as a result of taking classes of different teachers.
... the fundamental way to judge teachers is not how their students do on tests, but how different teachers affect the likelihood that their students finish high school and go to college, how teachers affect the earnings of their students after they enter the labor force, and whether their students get involved in gangs and crimes.
A small number of recent academic studies have tried to see how well VA measures predict how students do when they become adults. A summary of a good study along these lines by Raj Chetty, John Friedman, and Jonah Rockoff was published this summer in the journal Education Next under the title “Great Teaching”. ... 20 years... 2.5 million children ...
They find large effects on subsequent adult earnings when these young students had teachers who produced good improvements in test scores...

teachers with good VA ratings should be paid considerably more than teachers with bad ratings.

Or then again, maybe not.

Kids' Creativity Trending Down

It's common to focus on facts and skills, as being readily measurable. Still, there are ways to try to measure creativity, and they tell an uncomfortable story...

As Children’s Freedom Has Declined, So Has Their Creativity | Psychology Today

Kim, who is a professor of education at the College of William and Mary, analyzed scores on a battery of measures of creativity—called the Torrance Tests of Creative Thinking (TTCT)—collected from normative samples of schoolchildren in kindergarten through twelfth grade over several decades. ...

According to Kim’s research, all aspects of creativity have declined, but the biggest decline is in the measure called Creative Elaboration, which assesses the ability to take a particular idea and expand on it in an interesting and novel way. ... more than 85% of children in 2008 scored lower on this measure than did the average child in 1984.  Yikes.... ...

the TTCT seems to be the best predictor of lifetime achievement that has yet been invented. It is a better predictor than IQ, high-school grades, or peer judgments of who will achieve the most.[6]  The correlation coefficients found between childhood TTCT scores and real-world adult creative achievements have ranged from a low of about .25 to a high of about .60, depending on which tests are included and how adult creative achievements are assessed.[6]

So, the decline in TTCT scores among school-aged children indeed does appear to be cause for concern.  Kim herself calls it the “creativity crisis,” and that term has been picked up in a number of articles in popular magazines.

Well, surprise, surprise.  For several decades we as a society have been suppressing children’s freedom to ever-greater extents, and now we find that their creativity is declining.

Of course, maybe it's not freedom, maybe it's excessive fructose or excessive worry about fructose or too much or too little television. But it does seem like an issue worth considering, when we consider options for American schools.

Many to One Education

Economist Arnold Kling provides Many-to-One vs. One-to-Many: An Opinionated Guide to Educational Technology

The attempt to achieve large scale in college courses is misguided. Instead of trying to come up with a way to extend the same course to tens of thousands of students, educators should be asking the opposite question: How would I teach if I only had one student? Educators with just one student in their class would not teach by lecturing....
To put this another way, I believe that the future of teaching is not one-to-many. Instead, it is many-to-one. By many-to-one, I mean that one student receives personalized instruction that comes from many educators. To make that work, technology must act as an intermediary, taking the information from the educators and customizing it to fit the student's knowledge, ability, and even his or her emotional state.

I'd tentatively say he has a large part of a point here, but...hmm. The MOOC model, whether it's carried out for higher education or K-12, doesn't have to contradict his many-to-one ideal; if there are 100,000 students finishing a MOOC, then they all have something in common (they have passed "Intro to Calculus" or "Designing a Rocket Stove" or "Baroque Music as a Commentary on Galileo" or "Origami Polyhedra" or whatever it is.) That likely means that they have all done reasonably well on a series of (peer-graded + machine-graded) quizzes, exams, and projects. It means they share some kind of assessment. It doesn't mean that they have all gotten there the same way; it doesn't mean that they have all done what Stanford students do.

But I'm not disagreeing with him about what's important; education needs to be more customized, more personalized, not less, and specific MOOC proposals may try to save money by going the wrong way. (And I hereby propose that the "normal" course length should be a quarter, not a semester. Or maybe a week. Break everything into Pieces! Hulk SMASH!). And he's at least two-thirds right about tablets.

Or then again, maybe not.

Coursera TED talk--Daphne Koller

Lately I watch TED talks on a tablet, a Nexus 7, in the morning while walking quickly up and down stairs; it's 20 minutes of reasonable exercise for body and mind, and all that. This morning I watched That's Stanford CS prof Daphne Koller's talk on Coursera. Of course Coursera is higher education, but the techniques can certainly be applied to K-12.

• she wants to have lectures by the very best lecturers, broken into 8-12 minute conceptual units with students having to answer a question or two at the end of each unit (sounds like KhanAcademy or TED-Ed, so far).
• She wants personalized lectures, at least in providing optional units
• She also wants to have actual course scheduling, with tests and deadlines
• She wants peer grading, which apparently correlates extraordinarily well with teacher grading except that it's faster (22 minutes??), and it scales for huge courses
• She wants students to (usually) form study groups, both virtual and physical

She notes that personalization in this context can in several ways go farther than a physical classroom can offer, and it's not just choice of options; if 2% of your students share a misconception that causes them to get the same two test questions wrong in the same way, then as a classroom lecturer you'll never know, but as a MOOC (massive open online course) lecturer with 100,000 students, you'll see 2,000 students as one point on a graph and you can cause them to be automatically directed to the same extra units.

She also talks about the 2-sigma difference between (A) students who try to learn just from lectures, versus (B) students given individual tutoring. 98% of group B are "above average", if the average is set by looking at A. She thinks that personalized courses can come closer to B than to A.

I don't see anything there that doesn't work for K-12, except that you need to provide more physical supervision/security. (Well, I've known some small kids I would trust alone or even, maybe, in a group, more than some college-age students, but usually you need to provide more physical supervision/security.

Online Assessment

Everybody needs an IEP, especially me. But that's too hard, and we have no practical way to choose a program. But maybe we will: IBM Watson making progress to becoming a useful medical assistant for diagnosis and treatment planning

Given data on a new patient, Watson looks for information on those with similar symptoms, as well as the treatments that have been the most successful. The idea is it will give doctors a range of possible diagnoses and treatment options, each with an associated level of confidence. The result will be a system that its creators say can suggest nuanced treatment plans that take into account factors like drug interactions and a patient's medical history.

Let's try replacing "medical assistant" with "teaching assistant" and run that by again...

Will it work? Maybe. Will it be a waste of effort? Well, almost certainly in the first few times it's tried. Is it worth trying? I think so.