Applied Rationality: Group Problem Solving Session

post by apophenia · 2011-02-08T14:06:47.512Z · LW · GW · Legacy · 23 comments

Contents

23 comments

This is a discussion thread about applied rationality.

In the comments, please explain an actual problem you have in your life you want to solve.  Using their combined powers of rationality, the community can discuss the problem with the poster and eventually propose solutions.  My hope is that this will give Less Wrongers a better idea how to apply rationality to daily life.

A note to those posting problems:  Please stick around long enough to try the solutions and comment on how they worked.  Remember to include negative results (the solution didn't work), including mini-problems like if you meant to try a solution and didn't get around to it.

 

Edit: Ack!  I see Alicorn posted something similar about common-knowledge problems between when I wrote this and when I posted it.  Because I have a general policy against deleting my posts, I will leave this up here.  I do think rationality-specific solutions are useful, but let's wait a month or so.

23 comments

Comments sorted by top scores.

comment by lukeprog · 2011-02-08T13:32:21.575Z · LW(p) · GW(p)

apophenia,

To avoid overlap with Alicorn's post, why not edit your original post a bit so that the focus is more on applied rationality than on simple procedural knowledge that just needs to be memorized?

comment by apophenia · 2011-02-08T09:38:02.788Z · LW(p) · GW(p)

Here's a problem, for example formatting. Problems can be as easy or hard, and as complex or simple, as you think is reasonable.

Problem: I don't know how to find scientific data. I can get some common-knowledge stuff, but anything controversial (ex. global warming) or hard to locate by keyword (ex. how well normal distributions model actual data) I have a hard time finding. Especially for controversial items, I don't know how to evaluate the soundness of the study. What tools can I use to find and evaluate studies.

Progress/Effort made: I can find studies with specific keywords. My current evaluation strategy is to look at the statistical comparison in the experiment, or to find a blog that mentions reputable papers about a subject. This doesn't work in general, because blogs don't discuss obscure topics, and I have a hard time finding out if a blog writer is a reliable expert.

Parameters/Evaluation: I want to be able to evaluate studies in pharmacology, but I don't want to learn too much about pharmacology in general. The problem will be solved when I can find out whether a scientific issue is solved one way or the other, along with the degree of contention. I would prefer solutions that don't involve personally talking to experts.

Replies from: lukeprog, David_Gerard
comment by lukeprog · 2011-02-08T13:44:52.470Z · LW(p) · GW(p)

How to find scientific data:

  1. Do some research to find out what the two leading entry-level university textbooks are on the subject. To do this, check online course syllabi to see what they're using. Or, just do some searching. After 60 seconds on Google and Amazon, I found these two textbooks on global warming. I dunno if they're the best, but they'll probably do. Be sure to get textbooks and not single-author academic studies, which may be highly skewed toward a particular position that is not mainstream.

  2. Skim the textbooks so that you get a sense for the major concepts of the field and how they relate to each other. I buy a lot of textbooks, but maybe you don't want to spend the money. In that case, you can call around to find what university bookstores have the ones you want in stock, then go there and hang out and read it in the store. Or, buy it online and sell it back. If you keep it in good condition, you sometimes end up spending only $20-$30 on the textbook. If either of these two options are inconvenient, well... that's research, yo.

  3. Now that you know the standard terms involved, Google searches and Google scholar searches will do wonders.

  4. Even better, any textbook worth its salt will give citations for some of the major studies that support its basic claims (they usually mention lots of large meta-analyses, for example).

  5. Make a list of all the papers you want to read. Go to an on-campus library at a major university nearby (I drive 30 minutes to UCLA every couple of weeks). Usually you don't need a library card to sit down at one of their computers and download a bunch of papers (they'll have access to JSTOR, etc. on campus computers) to your flash drive. If they don't allow flash drives, upload the PDFs to your free (or paid) Dropbox account.

  6. But usually, the textbooks themselves will contain a good overview of which debates are relatively solved, and which ones remain open. They will also describe what major questions need to be answered to solve the open questions, and what the evidence currently suggests. Of course, it's important to get the very latest edition of the textbook. Be sure to check The Best Textbooks on Every Subject.

  7. Depending on what you want to learn, you may have to get a more advanced textbook on a narrower subject, but it will probably still help to skim a lower-level textbook to get a handle for the basic concepts involved in the larger field.

  8. Did I mention textbooks?

Replies from: Zachary_Kurtz, wedrifid
comment by Zachary_Kurtz · 2011-02-09T02:28:57.180Z · LW(p) · GW(p)

instead of buying textbooks check out library.nu

Largest collection of [illegal, mostly] free textbooks I've seen on the net.

Replies from: ata, blogospheroid, lukeprog
comment by ata · 2011-02-09T02:40:58.297Z · LW(p) · GW(p)

Library Genesis (mirrors: gen.lib.rus.ec, free-books.dontexist.com) is another very large collection, focusing mainly on math/science/tech/engineering textbooks.

Replies from: lukeprog
comment by lukeprog · 2011-02-09T02:43:20.818Z · LW(p) · GW(p)

ata,

Have you ever found any on these sites that aren't also on library.nu?

Replies from: ata
comment by ata · 2011-02-09T02:55:03.108Z · LW(p) · GW(p)

Several, I'm pretty sure, though I don't remember any particular ones off the top of my head.

(I usually check Library Genesis before library.nu these days anyway, as they usually have what I'm looking for and there are slightly fewer trivial inconveniences (no login, one click to download from the search result page instead of four clicks, no password-protected archives).)

Replies from: Anubhav
comment by Anubhav · 2012-03-03T06:41:17.573Z · LW(p) · GW(p)

And now that library.nu is dead....

comment by blogospheroid · 2011-02-12T13:36:51.919Z · LW(p) · GW(p)

I've downloaded a couple of files from there, but they are all encrypted, and nowhere on the site was it clear as to what was the decryption password.

Replies from: Zachary_Kurtz
comment by Zachary_Kurtz · 2011-02-15T20:21:47.013Z · LW(p) · GW(p)

the archive password is listed before each external link in every example I've seen. Usually the password is either ebooksclub.org or library.nu

comment by lukeprog · 2011-02-09T02:42:36.814Z · LW(p) · GW(p)

Agreed.

comment by wedrifid · 2011-02-15T20:54:09.888Z · LW(p) · GW(p)

In that case, you can call around to find what university bookstores have the ones you want in stock, then go there and hang out and read it in the store.

Libraries are good for that too - especially for high end textbooks that usually come shrinkwrapped. In fact, it is what libraries are there for. At universities sometimes the specific department libraries have more copies of a text than the general library - and also tend to have copies that can not be checked out so are always available in the building.

Replies from: lukeprog
comment by lukeprog · 2011-02-16T05:26:20.533Z · LW(p) · GW(p)

The reason I mentioned university bookstores is because libraries usually do not have the latest edition of any textbooks available.

Replies from: wedrifid
comment by wedrifid · 2011-02-16T07:51:24.192Z · LW(p) · GW(p)

University libraries do. At least, they do at all of the universities I have attended.

Replies from: lukeprog
comment by lukeprog · 2011-02-16T15:50:14.558Z · LW(p) · GW(p)

Huh. Well, I hope the libraries I've checked are the exceptions!

comment by David_Gerard · 2011-02-08T18:08:28.533Z · LW(p) · GW(p)

Wikipedia really is an excellent start, even as preparation for lukeprog's suggestions.

comment by jwhendy · 2011-02-08T18:47:53.690Z · LW(p) · GW(p)

I have some time to look into this more, but have been trying to wrap my head around the best way to use rationality on a decision about homeschooling vs. conventional schooling for my kids. Originally, homeschooling was primarily for religious reasons -- they would be home schooled by my wife and this increased the ability to control their environment and try to protect them from "negative" influences. I'm no longer religious and thus this is no longer a factor.

That said, I still see some advantages to one on one time with one's parent and have heard that home schooling can be more efficient.

My very basic list of variables or decision factors would be something like:

  • Finances: not just cost of homeschooling materials vs. cost of tuition at a school, but also some type of calculation that examines the impact on society as well. Teachers go to school for this specific task and I'm curious as to what impact "lay people" deciding to do that role has on society, education majors, the school system, etc. Hopefully that makes sense.
  • Behavioral: pros/cons of one-on-one time with a parent during education, benefit/hindrance of not being surrounded by a lot of peers for many hours each day, access to extra-curricular experiences, etc.
  • Child-raising, in general: would so much parental time increase/decrease likelihood of more difficult teen rebellion, etc.?
  • My wife: I think she's hoping to work eventually (she stays home with our kids at present) and my intuition says that the longer she's out of the work, the harder it will be to re-enter. Perhaps if she just stayed home until they were both in school, that would be the best maximization of motherly care at an early age, combined with the minimum time out of work to enable the highest likelihood of workforce re-entry.

I'm sure there's more. I'm making this post quite early on. I have a 2.5 year old and thus have some time on my hands, but it has been more and more coming to mind and I've been trying to wrap my mind around all the variables. If anyone has been through this decision or has input on research into such things, that would be fantastic.

Replies from: MinibearRex, saturn
comment by MinibearRex · 2011-02-08T21:57:34.733Z · LW(p) · GW(p)

Firstly, you didn't state whether you were considering public or private school. Most of my experience and knowledge is with public school, so some of my own points may not be applicable.

  • Finances: Public school would be cheaper than home schooling for you, but I don't know about a cost comparison to private schools. In turns of a social impact, I have no real knowledge either way.

  • Behavioral: I knew a lot of kids who were homeschooled growing up. Most of their parents tried to get together with other homeschooled kids as often as possible, so that their kids had social interactions. Those individuals were pretty well adjusted. I knew of a few kids whose parents didn't do that, and that weren't nearly as social. Availability bias may be affecting my judgement on this. If you do decide to homeschool your kids, and you do decide to get together with other homeschooled kids, you should consider the people who are likely to be in that social group. You are not religious, but the vast majority of them will be. Your kids' friends will probably not be learning about evolution. You may have to stave off conversion attempts. That's a possible downside. With regards to extracurriculars, I knew a lot of homeschooled kids that did participate in (mostly scholastic) competitions between schools. Again, availability bias is skewing my own perception of how common this is, because that's how I met most of them.

  • Child raising: Most homeschooled teenagers I knew weren't big rebels. Part of that may be religious; most of them were frequently going to church youth groups and bible studies and had parents that were fairly strict. I don't know how much connection that had to parents or religion, but those will probably be the children your kids are friends with.

  • Your wife: I tend to agree with your own take on this, assuming you can take care of teaching your kids all on your own.

  • Another point you didn't mention: Education. Are your kids going to get a better education at home or at a conventional school? You are reading and commenting on less wrong, so I assume that rationalist methods of thinking are fairly important, and that you probably want your children educated in them. You can teach them that, as they grow older, but a public school probably won't. You could still teach them those skills on the side, though, if you wanted to. Considering it the other way, I don't know what quality of school is in your area, but how educated are you in each subject compared to the average teacher? By the time your kid gets to high school, their teachers would be individuals mostly with masters' degrees in their respective fields. You may be good at the field you currently work in, but are you going to be as good at teaching the American revolution as a professional in the field of history? And as good at teaching chemistry as someone with an advanced degree in chemistry? A major advantage of having multiple teachers is the capability for specialization. Your kids are likely to get more knowledge in a public school, but might get better thinking skills with you.

Personally, I went to public school and did quite well. I was in all honors classes, which helped give me good thinking skills and knowledge. I definitely had complaints about the way the school system was run, but I did (eventually) develop good social skills because of the constant daily interaction with other people. However, my parents did teach me a lot outside of school. They bought me books, encouraged my own curiosity and originality, and did a pretty good job at finding a balance between discipline and liberalism. No matter what you choose to do, don't forget that you will still be immensely important to your child's education.

Replies from: jwhendy
comment by jwhendy · 2011-02-08T22:42:06.344Z · LW(p) · GW(p)

Great reply. To your points:

  • Public/private? Not sure. Haven't thought about that much, but I can't think of a good reason to pay more for private, particularly without the religious motivation.
  • Behavioral: great points and I can think of examples where I've known home school kids to interact with others through various sports organizations or clubs.
  • Rebellion possibility: good to know and thanks for sharing your own experience growing up as well.
  • Wife: she would be homeschooling, not me, so it comes down to deciding whether she continues to stay home, and if so... how long.
  • Education: great questions. I'm a mechanical engineer and would feel quite comfortable teaching up/through high school chemistry, physics, math (geo, alg, calc), control systems, [basic] computer programming, english, and probably some others I'm not thinking of. I should have mentioned this, but I think I'm only really considering this up through middle school anyway. I think high school is a big enough jump in several areas (responsibility, freedom, potential for more long-lasting friendships) that it would be a good thing for a kid to enter the school environment by that time. Also, as you point out, it's almost guaranteed that teachers there will be better than me at most subjects if not all of them.
  • In general: I agree that I will have a role no matter what, as well. I'd like to invest a lot in teaching far more than what goes on in the classroom and have learned a lot from reading LW about trying to emphasize practical applications of knowledge and tools, not just carrying out equations in a vacuum and having no idea what they mean.
  • Side note on religion: my wife is still very religious. She seems to have eased up on the determination to raise my kids religious... but they still pick up quite a lot from her as she prays at home when I'm not there. They pretend to read the Bible, bow their heads before meals, etc. I'm wary of this impact on their future ability to remain open minded to evidence based reasoning when it comes to this area. It makes me a little wary about them being home with her all the time.

I really appreciate the comments. They are very helpful in opening me up to the various facets of this topic as I pursue further research.

Replies from: MinibearRex
comment by MinibearRex · 2011-02-09T23:42:21.104Z · LW(p) · GW(p)

Happy to help.

With regards to the side note on religion, that sounds fairly similar to my own upbringing. My dad was fairly nonreligious, maybe deism is the right word. Haven't talked about it with him all that much, but definitely not Christian. My mom, on the other hand, is quite religious. Not a fundamentalist, she's a biologist and believes in evolution, etc, but still definitely gave me and my brother religion. I can't say that that was fantastic, but I started being a rationalist as I transitioned from a Christian to an atheist. If your wife is raising your kids religious, they might yet get some benefits from it, even if it's not what your wife intended. Emphasis on might, though.

Replies from: jwhendy
comment by jwhendy · 2011-02-10T16:13:24.841Z · LW(p) · GW(p)

Thanks for the additional comments. Yes, it's a tricky question. She seems to have shifted from a "Definitely will raise religious" to "Will simply do what I do and if they ask about it (e.g. what prayer is, god, etc.) then I'll tell them what I believe." We'll see how this plays out, but I think telling them about (if that's really what she does) will be a far stretch better than instructing them on or teaching as fact. Does that make sense?

I'm glad this happened, as I was in a tough spot. I tended to think that my options were a) both teach equally that our positions were true, b) me let my wife teach religion and I say nothing, and c) teach proven rational tools but not anything about religion.

I thought c) was by far the best option, but also thought a) would have been more harmful than b). I don't think whiplashing a kid between two sides at young ages would have been helpful. It seems we've migrated toward c, which is great.

There's my off-topic ramble for you -- thanks for your comments.

comment by saturn · 2011-02-08T21:48:08.654Z · LW(p) · GW(p)

How much research have you done? For example, there are secular homeschooling organizations that could probably give you some useful information.

Replies from: jwhendy
comment by jwhendy · 2011-02-08T22:33:18.627Z · LW(p) · GW(p)

Not much. I really only posted here because I'd been thinking about it and the post seemed to ask for this kind of thing... so I went for it. I'll be looking into it far more in depth myself but went for putting it out there anyway.

Thanks for the suggestion about various organizations. It ended up being quite fantastic!.