Education Red Alert: Astral Codex Ten Alpha School & School Reviews

A few thoughts - and questions for my visit.

These two reviews have dropped like a bomb onto the educational blogosphere. They are the first two reviews posted in Astral Codex Ten’s annual book review, except that this year participants can review anything except books. Everybody has thoughts, myself included. I also have a lot of questions. I will assume you’ve read the reviews, which can be found here: #1 Alpha School Review, and here: #2 School, the Institution Review.¹ They are both excellent, thought-provoking pieces that complement each other well.

I’ll start with the second review, “School”, first and then talk about Alpha School. I will follow with a call for questions that I can ask when I visit Alpha School later this month. I’ll finish with some thoughts on incentives and what each review holds in common.

School Reviewed: What is it good for?

I am impressed with this writer’s clarity, command of sources, and their ability to break down the key information into digestible and yet still-appetizing chunks. In other words, they write like a schoolteacher.²

The author claims that school isn’t primarily designed for learning, but to create motivation. The motivation is necessary to get children to learn and typical public school models basically provides two forms of incentives: structure and conformity.

The structure of school is habit-forming. You have to be there, you have certain expectations of behavior, you are supposed to be learning, etc. This structure provides behavioral nudges that get students to engage in the learning process.

The conformity aspect is why people are congregated in classes of same-aged peers. You are incentivized to fit in with the group. If you fall behind you’ll be ashamed and want to catch up. If you get ahead, you may be bored but at least you’ll be recognized as competent by your peers.

For which kids?

The author acknowledges that school doesn’t work for everyone. They divide learners into three groups. Some kids don’t need extrinsic factors or a formal school setting to learn, they’re no-structure learners. Autodidacts fall into this category. It’s small- probably no more than 5%. A second group learns reasonably well as long as they are surrounded by the structure and conformity of school, but they aren’t very sensitive to factors like teacher quality or pedagogy- these are the low-structure learners. The last group tends to struggle without significant support. These high-structure learners are more sensitive to teacher quality and need a carefully scaffolded curriculum in order to succeed.

According to the author these categories are fluid and will change over time and between different subjects, but they help to describe why school fails to satisfy so many and yet why it still may be the best option available. Some students don’t need organized school and it is restrictive to them: the no-structure learners are bored. The high-structure learners can keep up for much of primary but tend to fall behind at some point and become demotivated. The majority of students, however, tend to need the structure of school to get them to continue learning even if they’re bored because the material is easy. Thus, the system continues.

There are a few other claims in this post that deserve interrogation:

Claim: Personalized education has failed. I will admit that Khan academy and many of these online personalized courses only work for a small sub-group of the population (especially no-structure learners). I also think that the author is basically correct in suggesting that one key element they are missing is social incentives. Still, I am not sure that I can conclude that personalized learning has failed because I don’t think it has been employed at anywhere near its potential yet. Tutoring has proven effects and personalized learning is aiming at similar benefits but lower costs.

Claim: Tracking and ability-grouping are ineffectual. The author actually says that effects, whether positive or negative are very small. I’m not sure about this claim either. Could there be a way to implement tracking or ability grouping that still allows for appropriate social incentives? I’m optimistic.³

There is also a claim that tracking and ability-grouping don’t scale because the minute you get a higher-status track, parents oversubscribe to it and it becomes too large and general to meet the accelerated needs of its most-able constituents.

I respect what the author is trying to do here. It’s a rare, brave soul who looks at the status of public education today and tries to defend it. I do think that learning is happening and that there are other benefits that schools provide that aren’t captured in the measurable learning outcomes.⁴ The fact that education has seemed to resist massive changes in its structure for so long does provide some evidence that it may be more efficient than reformers would admit. That said, the reviews claims can come across as downright Panglossian. I don’t think we should tear down Chesterton’s fence public schools without any replacement but surely we can aim for improvements. Maybe a hedge?

Alpha School Brownsville

Alpha School: has the educational messiah arrived?

Does it work? I have no doubt that it works for those attending it right now and find it plausible that it has cracked some proportion of the code of education.

Some of the most exciting bits, for me:

• personalized, immediate feedback with built-in schedules of rehearsal

• the idea of finishing things so darn quickly

• the inclusion of both an academic component and an agency-focused component

• the transparency, both by the reviewer and on the part of the school - nobody publishes so much data and the data looks good

• the incentive system and the reviewer’s model of motivation

• the possibility of using eye-tracking and AI feedback to aid learning

• the focus on legible results

Some things I found a bit off-putting:

• I get a little ick factor when everyone goes to their own area and is on their own screens

• lessons involving ‘curated YouTube videos’ also gives me an involuntary negative response

• $$ for learning - though I am very open to the idea that I’m wrong about this

• the possibilities of using eye-tracking and AI feedback to aid learning - yes, I also listed this above - it could be a revolutionary or a hellscape depending on implementation

• The costs seem prohibitive for most students

• The focus on multiple-choice questions makes me skeptical of the depth of thinking required in the 2-hour sessions

Questions after reading about Alpha School:

Everything in education is selection effects. Even if Alpha school isn’t necessarily selecting on IQ, they are certainly selecting on income. How much of this is selection effects? I mean this is the obvious point of skepticism going in, but the reported numbers do seem strong enough merit investigation.

How sustainable is this? Grinding out lessons on personalized learning apps certainly doesn’t seems to work for most students on its own. The incentives, both social and the monetary ‘Alpha bucks’ certainly add more on top of it. Alpha school seems to run mostly from K-8, though it appears there is one high school in Austin. Something seems to change in the motivation profile between primary and secondary schools and I wonder whether this model will translate well and for whom.

How much does range restriction affect our conception of the effectiveness of personalized learning and ability-grouping? What if we went for this with more dakka? Success for All, one of the more evidence-backed literacy programs out there has been doing this since the 80s with positive results. This is a genuine question.

What are the teachers at Alpha School doing? The 1:5 teacher guide-student ratios seem great, but there’s relatively little in the review about what teachers are doing while students are engaging in their 2-hour learning time.

I would love to get a couple of days to see what the school is like, to see how transitions between break time and academic time are handled, to observe the quality of the interactions among students and between students and teachers. Fortunately…

I’m going. Do you have questions for me to ask?

I am traveling to Austin with my family at the end of this month and I will try to book a tour of Alpha School (or GT School) and ask a bunch of questions. Do you have any burning questions that I should ask?

A final note about incentives

I think that the common thread with both of these reviews is that education is really about motivation. The model of motivation that is in the ascendant right now is that students are motivated by mastery. They feel good about learning when they see their efforts rewarded with success.

The School review mentions conformity as the core of public school’s motivation model and in the Alpha School review lists positive peer regard as a key aspect of motivation. Both of these are important aspects of social incentives. I think that, to the extent that online personalized learning has failed thus far, much of the problem stems from neglect of social incentives. People perform because their actions derive positive attention from others. As young kids, that’s typically a parent or teacher. Older children and adults tend to perform for peers and status. Having a real person who is paying attention to what you are doing is really important.

The final and perhaps most controversial incentive in Alpha School is the cash reward system and gamification of school. Instantaneous feedback and rewards are very powerful, but that power really does need to be used thoughtfully because it has such a strong effect on behavior. That said, if it works is any more coercive than mandatory attendance? I’m not sure. There are legitimate worries about extrinsic rewards devaluing intrinsic motivation. However, the argument in the review- that extrinsic rewards can be used to help build lasting positive habits - may bear out in practice.⁵ What seems clearly true is that using multiple types of incentives to build motivation to learn could be a part of improving education beyond what we have today.

I’m left with a lot to explore: How can I apply these ideas about educational incentives with my children at home and with my students at school? Am I using enough of the right incentives?

What about Kieran Egan? How do these paradigms relate to Egan’s approach to education? Is there anything that Egan’s approach offers that these programs are overlooking?⁶

1. I’m going to publish this less as a long-form post and more as a series of reactions, because there is a definite speed premium for this one.

2. In fact, I’m going to go on record with a guess that this is written by Michael Pershan. 2nd guess is Dylan Kane.

3. See below.

4. Socialization, shared experience, babysitting (yes), soft skills, etc. I’ve been rewriting this post in my head for ages now.

5. For more on this, read Zvi’s post on Alpha School:

On Alpha School

6. Spoilers: I think yes, or at least yes until I’m proven wrong.