My 90 Minute Attempt to Teach Rationality to 5th/6th Graders:
[Remember that this is based off my subjective assessment of a sample size of ~30 students with no actual recording of data. Please draw conclusions with care and with your own judgment in my ability to give useful second-hand info.]
Things I attempted to cover in ~90 minutes:
- A 45 minute presentation on economics, game theory, and basic 80,000 Hours-esque considerations with an environmental slant. Also touched on cognitive biases.
- A 45 minute “smash class” on rationality where I tried to give some mental models of what rationality is, with some digressions into techniques.
More in-depth description of my classes:
- In Talk 1, I covered:
- Anchoring, overconfidence, scope insensitivity
- Scale, considerations of proportions
- Thinking on the margin, replaceability, comparative advantage
- Prisoner’s Dilemma, climate change
- Arguments, Double Crux, finding the truth
- In Talk 2, I covered:
- System 1 vs System 2
- Epistemic vs Instrumental distinction
- 2, 4, 6 game
- Reminders vs Motivation
- Internal Double Crux
(For context, I didn’t know how long I had and did not create any new materials. I merely hashed together two talks I’d given previously in different contexts, one at an environmental conference, and the other for two rationalist friends-in-training.)
The students were a mix of boys and girls (don’t remember gender ratios, seemed fairly even upon recollection). About half of them seemed to respond very well (read: excitedly) to my material.
I did not dumb things down; both talks I gave were essentially unchanged to their original versions (which I’ve given to high school students). I think that there were some points where I asked if X made sense and they said “yes” even though it probably didn’t make that much sense.
- Potential Query Point: Students aren’t actually sure about norms for assessing their own understanding and potentially default to saying “I know what’s happening” even if they don’t. Not just social norms about looking stupid, but it’s actually sort of a mental blind spot (How can you know what you don’t know?)
Students were high energy. Somewhat overly so at times. I generally felt like the attention was on me, and my approach got their respect / their attention in a way that didn’t exactly shift the power dynamic to the traditional slave-master student-teacher relationship. (More on this in the below section).
Today’s students are apparently more trendy than I think. Dabbing, “roasting”, and “triggering” were commonplace slang.
My first talk was very casual and focused on being loud, humorous, with nice-looking graphics. This obviously wasn’t all of it, but I think this went a long way to “bringing them onto my side”.
[This has interesting parallels to how I approach performances in magic, actually. Magicians that “challenge” their spectators to figure things out are setting themselves up for confrontational battles, and this is often a newbie/stereotypical way for magicians to act. I adopt a far more familiar stance where it feels like I’m sharing / partaking in the experience, which places more of a focus on the content and less about “figuring out the secret”. Obviously this can’t be done in all settings, but I think this parallels teaching in some regard, where it’s less about lording power over the students and recognizing that you both need to cooperate to get the Thing to work.]
So it felt like there was a sense in which the students recognized that I had put in effort and was trying to not suck, so they were willing to do things like wait before speaking, wait for me to call on them in the midst of things. I think it was a good amount of giving-and-getting.
Good transactional dynamics appear to pay dividends in them treating you like a human. I think the moral here is something like “Treat your students as humans and they will tend to give you a similar amount of respect back?”
I think humor I used, the sort of dry, simple kind (showed a picture of a stick figure on the presentation, said it was a picture of me) seemed to work well.
Also, I used minor swearing sporadically (“hell”, “bullshit”, etc.). Sorta unsure about this one point, but it seemed like it had some good effects in setting the tone / letting people know I was on their side.
I took questions somewhat periodically, and students seemed happy to wait a little bit before I called on them. That seemed like a good sign.
Not that many students had questions. I think that more of them may have wanted to ask things, maybe? And there seems to be a skill here necessary to “jumpstart” questions. EX: After one person asks a question, more start to trickle in. That was good.
The Actual Rationality Teaching:
I think I may have somewhat immunized these guys against bad epistemology. That seems pretty good. The Double Crux stuff had some people nodding along / writing down the term. The stuff on rationality being this cool thing should hopefully give them positive affect.
One student, after I talked about the IDC model, mentioned how it was like “you were a judge between the two sides who are arguing”, and I really liked that analogy.
I don’t think the students actually walked away with good ideas of what the heck rationality even is. I don’t know if that’s really a failure, though. (Base rates and priors were pretty low.)
However, if you now told me that I had 90 minute to teach rationality, I think the class I could come up with would be far better than today’s.
My actual rationality content was fundamentally based on the use of a tree to keep things organized. I started w/ the epistemic-instrumental divide, and then I moved into how instrumental can be seen as a combination of creating reminders and fighting motivation.
(I think having some sort of categorization is very important in teaching rationality in general. Otherwise I feel like it’s hard for people to chunk stuff well.)
Probable Actual Outcomes that Happened as a Result:
- Increased positive affect towards psychology, rationality, and game theory.
- Basic understanding that “rationality” is a Thing that has Sub-Things.
- Leaving all the lights on is totally fine if you turn off the thermostat (I repeated this point several times to drive home the idea of the Pareto Principle / thinking on the margin.)
- Owen is a pretty cool guy.
Things That Went Well:
- Thinking on the Margin:
- My impression is that this is a surprisingly easy concept to convey to students with some well-placed examples, like how training times in sports follow a log distribution.
- Getting Students to Teach:
- This went over pretty okay (not a lot of volunteers, only 3), but I felt like this was a powerful tool that could have been used to greater effect.
- Giving and Getting Respect:
- In future endeavors, i think this is a fairly satisfying relationship to have. I’m striking a balance between both being friendly, acknowledging the other side as being actual humans, and also trying to maintain the underlying structure of the teacher-student relationship which means I have things to offer.
What I’d Change for Next Time:
- The 2, 4, 6 Game:
- This was fun for the kids, I could tell, but I didn’t set it up well, and I ended up just encouraging them to keep going for examples, and then they actually got the rule correctly. So I missed a good teaching opportunity here, but it’s also good to know these students did the unlikely.
- Sys 1 and Sys 2
- I feel like this was a pretty intuitive idea, but I didn’t teach it that well. I want to focus more on examples of where this is. (I only gave color recognition and math problems as one example of each). I’d like to do more of these.
- More Examples of Biases:
- I gave some stats, but that obviously wasn’t convincing enough (even with pretty pictures). I’d want to take more stuff from Kahneman et al and give them some more demos. (Asch conformity, Bystander effect, Scope neglect.)
- More Examples of Rationality Kicking Ass:
- I don’t think I gave enough social proof / convincing awesomeness stuff that would convince them to want rationality. I think it’d be good to find more ways to tie this meta-skill into other domains so the students could source more desire to learn.
- Smaller Group:
- As I mentioned, something like half of the kids were really into it, and the rest seemed to be tangentially not-that-bored-at-least. I think if i just took half the kids, I’d be able to run it perhaps a little better. Decreased group size increases familiarity which in turn builds into the friendly-attitude I was trying to cultivate. (Also something about how people don’t want to be mean to their friends and reducing group sizes increases perceived closeness, so I’d be able to use that to my advantage as well?)