Economics. I took a lot of econ, and loved it. It really is a powerful framework for understanding the world. Bastiat especially has a lot to teach aspiring rationalists. I would encourage anyone to study programming, even if it's just at the introductory level. Understanding what it means to explain things to a computer is just a great exercise in precise thinking.

Philosophy courses did, seminar-style analytic philosophy classes in particular. (I wouldn't say that history of philosophy classes altered the way I thought, though I can totally see how Hume might be shocking to someone very new to the subject.) Aside from the actual content I learned, I got the following out of them:

1. The mental habit of condensing complicated lines of reasoning into minimal, fairly linear syllogisms, so that all of the logical dependencies and likely points of failure among the premises/inferences become much more obvious.
2. Relatedly, an eagerness to search for ambiguities in arguments and to enumerate all their possible disambiguations, with an eye for the most charitable/defensible contenders.
3. An appreciation for fine distinctions underlying seemingly straightforward concepts. (E.g., there are several related but distinct concepts that map onto the notion of a word or sentence's meaning.) These often have unexpected implications and/or vitiate seemingly plausible inferences.
4. Not being allowed, on pain of embarrassment or a bad grade, to get away with BSing or relying on unacknowledged, controversial assumptions. You have to be up-front about precisely what you mean and what's at stake.
5. Realization of the extreme rarity of knock-down arguments for any view, and the subsequent adjustment to the fact that assessing pretty much every philosophical question involves a robust trade-off of good and bad consequences. Sometimes every view on the table seems to imply something crazy, and you have to learn to accept that. And to accept that sometimes reality is crazy. (Yes, I know that the map is not the territory, etc.)
6. If arguments are soldiers, then at least learning to let some of your soldiers die - and sometimes even putting them out of their misery, yourself! It's very common in philosophical writing to go through all of the failed arguments for your view before moving on to the ones you find more promising. Even then, it's expected that you highlight their most vulnerable spots.
7. Epistemic humility. I'm much slower to draw hasty conclusions with high certainty on a given topic before I find out the best of what all sides have to offer. I definitely still form fast and intuitive judgments before investigating disputed subjects deeply, but I don't pretend that they're likely to be the last word or even novel contributions lacking high-level criticism.
8. A much richer sense of the space of philosophical views. But then again, probably something analogous holds for most other disciplines (biologists are presumably better-tuned to the space of biological hypotheses). Still, though, philosophical-view-space intersects an unusually large number of things.

I don't know if you're in need of any of these things, or if you're likely to acquire them through a small handful of philosophy classes. Even if you are, whether or not you'd succeed greatly depends on the quality of your teachers and classmates.

I have a Ph.D. in physics, and, needless to say, during the long and painful process of achieving it, I've definitely become more rational (in the broader sense) and a clearer thinker. My achievements may sound very basics to LW people, but you can consider them like the foundations of rationality, I guess.

• Calculus and Algebra taught me how to clearly apply the hypothesis->proof->thesis process. It might look a not so great achievement, but it is indeed, since it reflects in how I treat everyday issues (it's like the first step toward rationality in some sense)
• Programming courses taught me to break down complex processes in simpler atomic steps.
• Lab courses have taught me how to treat the data. Again this skill has been since then applied on a day by day basis
• General Relativity and Quantum Mechanics taught me that reality in not what it intuitively appears to be. This is a very important teaching: don't always trust your intuition, even if you really feel that you're right.
• Doing research activity, finally, taught me to calibrate my judgement (I used to be more often wrong than I expected), collect and elaborate data by myself and that apparently all the referees of Physical Review love to nitpick.

Physics for me, of course. The upper-level physics labs (where you get to choose your own projects, start playing with NMR and cosmic-ray detectors, and are expected to display actual statistical rigor) were particularly valuable. Though of course you need most of the physics before and a maybe a statistics class to really appreciate them.

Computer science (not necessarily programming) teaches you to think about the world rigorously and gives you an idea of what you can and can't do and how to do the things you can. I'd recommend topics like theory of computation (computability/complexity), algorithms, randomness/randomized algorithms, plus cryptography / quantum computing if you have time. The lectures at http://www.scottaaronson.com/democritus/ cover a good collection of topics, although it may be useful to have a more in-depth understanding than just the lectures (certainly you should at least solve the problems he give).

Mathematics makes you better at rigorous thinking, problem solving, and developing intuition. The same goes for theory of computation / algorithms above, but math classes can be a good addition. At least for a while, the classes also provide you with some really good tools (they get increasingly divorced from reality after some point, usually at the graduate level). I'd recommend taking at least one class in each of combinatorics, analysis, algebra, and topology.

You said things that improve your general thinking, and for those I would give the above. Note that I'm not concentrating on usefulness for other things (like more direct usefulness) at all, although a lot of those classes are also generally useful. But here are some other topics that are just really really useful:

1. LINEAR ALGEBRA -- this is actually just the most used set of tools in almost all technical fields
2. statistics -- not crap about T-tests, although that is moderately useful to know; I mean how to manipulate/analyze randomness, statistical modeling, and probability theory (things like MIT's 18.440, 6.041, and 6.856, David Blei's Bayesian Nonparametrics class, and Martin Wainwright's high-dimensional statistics class; sorry, the last one doesn't have lecture notes, although the last two are also a little specific / advanced)
3. statistical machine learning

I think I began to think much more clearly after I started reading about microeconomics. Opportunity cost in particular was an eye-opener.

I think part of it has been the content (opportunity cost especially), but another part has been the presentation. The economics writing I've read (from a couple textbooks, several popular books, and online articles/blogs) has generally done a good job of separating positive analysis and normative recommendations.

Programming was mentioned here, but surprisingly enough some aspects of system administration provided some insights that were useful to me later when studying mathematical logic. In general, even without all the AI thing, computers seem to have allowed people to repeat all the philosophical concepts in a new context..

The 20-th century foundation-of-reasoning discussion of proofs as objects (and how proofs depend on the set of the natural numbers..) can be nicely illustrated with chroot management; the old Aristotle idea of "first substance" of every object (unique but devoid of properties useful to use) is easily translated as DB id field.

Physics and recreational mathematics + computer science improved my mental abilities.

I took Physics for Engineers as a freshman in college. It's clear in retrospect that this class was designed to accomplish several things:

• Force students who breezed through high school with little effort to work hard to maintain a good grade.

• Weed out the students who don't have the intellectual firepower or stamina for engineering.

• Teach a particular problem solving methodology. To get decent marks on a problem set, you had to always draw a diagram, always start with appropriate equations, always derive the answer correctly (no numbers, just algebra + calculus), line by line.

• Work quickly and accurately. Tests and problem sets were always difficult enough to require the full amount of time allotted.

• Teach physics. Only this goal was ever mentioned explicitly, of course.

This was a difficult class to do well in, and probably the class most responsible for people leaving the engineering program. The default schedule also had it co-requisite with Calculus II, another demanding class, and the two classes used each others' concepts - you'd have to understand one to do the other.

Going through this taught me all the intended lessons: I went through high school barely studying, but had to give up sleep to study and understand every homework problem to do well in this class. I wasn't sure I had an A in the course until after I got the final back.

Internalizing the lessons of that course set me up for success in later courses, and now professionally. Physics courses aren't about physics.

Project Euler is the other thing I think has changed my thought processes for the better. Being able to think of how to express an algorithm succinctly and correctly seems to help out in various situations, like training a new employee how to do a complex technical task.

I thought about this before posting: Evolutionary Psychology. I don't know that there are courses with that title, but look in bio and psych departments for offerings. Primate and Mammal psychology would also be valuable. Understanding things in evolutionary psychology have significantly enhanced my introspective attempts to become more accurate in understanding the world.

(my first thought was physics, 2nd philosophy of ethics where as a 17 year old I learned that there IS no "right" answer, that the best human minds ever are just giving their opinion.)

Math and computer science are the obvious choices. I'll second intro to microecon. It will dissolve a lot of confused thinking wrt psychology and interaction (transaction).

Think it will depend on what your other courses are. If you lean toward the mathemathical arena, then stuff like psychology or other social studies courses will give you an alternative to the way you think now. I have an idea that being a precise thinker is about being balanced in your areas of expertise. Anyway, it kind of depends on how you define precise thinking.