Teaching Problem Solving Skills

Each of us only knows what is in our own heads. We can get a window into what other people think by listening to what they say. Unfortunately, there is often a fundamental difference in perspectives that remains unsaid. Getting others to think differently is a common struggle that can apply with your kids, friends, employees, really anyone!

When I boil down this entire blog and my idea of the “meta-problem” what I am really trying to share is a general framework for problem solving. The framework is simple enough: Understand what you are trying to do, and then assess what decisions you could make to help achieve those goals. It sounds simple, but there are a lot of layers I’ve needed to incorporate to have answers for all the kinds of “yes, but…” that come up. In the latter portion of this post, I’ll detail my own journey and the various disciplines I have pulled from.

But to the general framework, the foundation is that we are people discussing human action. In physics, we learn about forces that drive action. Forces are fundamental to all physical phenomena. Said another way, you can understand what will actually happen if you just understand the forces. Humans are the same, though those fundamental forces are different. People do things for reasons.

Those reasons are what the meta-problem is all about. Define what you care about and say what you believe your ability to impact those things is. A group can discuss those things directly and develop a shared set of reasons. Sure, I might think one bit of information is more important than you do and therefore come to a different conclusion than you. But if we don’t start from the reasons, everything looks random. If you didn’t know about gravity, how would you explain that what comes up generally-but-not-always comes down?

Human problem solving is then fundamentally about the forces we as people have for action. It’s confusing because we generally talk about teaching problem solving by teaching specific methods in school. But why would computing an equation be the same skill as deciding what to do in a situation? My simple answer: They aren’t the same.

The difference is why coming up with a new term for “deciding which problem to solve” has been such a breakthrough in my project. The meta-problem speaks to the fundamental forces for human action: reason.

My Journey

For as long as I can remember, I have been interested in the process of decision making. I remember coming across a book at the library in high school that described how to make decisions under uncertainty in daily life. (I tried to find it a decade ago but was not successful).

In college I took my first optimization course, and it presented a mathematical model for decision making. I took the whole idea to heart and concluded that decisions should be rational, and so the reasons always came down to math.

In an early class in graduate school taught by Olga Raskina, she shared the observation that people would often give you their requirements but leave things out. Initially I thought they were being unhelpful on purpose. My journey since then has been basically a circle. I learned bit by bit the weaknesses in the math model of decision making and ultimately how that same math model of decision making solves the very issues it created. Some choice examples:

• How can you balance dollars and not-dollars? Multi-objective decision making and the Pareto frontier says you can trade off one objective against another to make an overall decision.

• People seem to often get locked in to a single solution, but why? I took a systems engineering course and learned that collecting a diverse set of alternatives and comparing them was a really effective way to manage tradeoffs.

• Why do people seem to make really bad choices a lot of the time? I had a friend who came from the field of descriptive psychology. There are a few dozen “maxims” from descriptive psychology, and the one that has had the most lasting effect on my understanding is that basically, people’s behavior will always make sense in the context of their views of the world.

• How can you make a good decision when there are unknowns? My PhD was focused on decision making under uncertainty. I learned the math but also the intuition for how it can be difficult to make good decisions when you have unknowns (as well as what good means).

• Why do people play the lottery when the expected return is negative? I took a couple game theory courses and learned about mathematical models of preferences and how they play out in decision making.

• How can we figure out how people balance dollars versus risk? I’ve recently dug a bit into the field of “Decision analysis” and learned all the ways that field tries to elicit preferences from stakeholders.

• What should we be working towards when solving the meta-problem? Since graduate school I’ve done a ton of requirements gathering and have learned a lot about what works and what doesn’t.

Ultimately, I have come back to the intuition of learning algorithms and how our own brains work. If we don’t have a reason to do something, we won’t. Which means that without reason, the world is entirely unchanged. If you want to change something, that comes through action. Choosing an action is a decision.

With that, we’re all the way back to decisions come from reason. Therefore, we can follow a model of “if we believe x decision will have y result, then we use that information to decide if we should make that decision.” Back to problem solving, the kind of problem solving that has an impact on the world is about decisions, which means it’s fundamentally about goals and how we can realize them.