Some of the hardest decision problems we (people, business, government) face are policy designs. Policies are, in effect, what rules or incentives are put into place to shape many subsequent decisions. Public policy problems are complex are in ways that sustainability problems are complex, with the additional dimension that policies in effect regulate human behavior. Stakes can be existential, literally or perceptually.

Lots of people exercise power and make policy without much preparation. And, they may do it without data. Not all of these policies are bad, some are great. In its way, democracy is a systems design for citizen delegates to make policies that are robust over time. And, very well educated, experienced, well informed and well intentioned people have produced catastrophes.

My belief is that we get better, more broadly beneficial policies in all realms on average if we have on average solid fundamentals in managing, deciding, reasoning critically, and applying evidence of all forms, including from data. These fundamentals, to me, seem like they should be part of a college curriculum core. Some schools would certainly say they are.

In practice, the prospect of organizational leadership merits training – leadership is not just for business school, and not purely an experiential skill. In practice, data science for application to judgment and decision making problems is different than for example machine learning for automating judgment or experiment design in an interventional study.

And, when we look at complex problems, deep specific skill in policy subject areas needs to be complemented with the means to look at impacts in a multi-systems context and the means to look frankly at the obligations and consequences of the exercise of administrative power. The problem with complex problems is that everything seems to matter all at once. The job of analytic decision making is to help the policy maker accommodate the surface complexity through focusing and reducing into salient models, data-informed judgment, and useful simulation.

In a way, the policy design problem boils down to “what will happen to whom if we do this?” The collateral questions of “Is it fair?” “Is it sustainable?” shouldn’t seem hard. It’s the unintended consequences, and the complexity the combination of dimensions of stakeholders and underlying processes that make a robust answer hard.

Curriculum

How would I package all this stuff in a college curriculum? Here’s a sketch. I am posting this to support discussions about other frameworks.

Foundational Skills for policy design

In terms of the “Fundamentals”, or core,

  • Model thinking is critical thinking for a world viewed through a data lens.
  • Management arts are the skills of business leadership. Leading problem group solving processes. Leading teams.
  • Decision Making is a competency that is not just for managers – one can set it as a family of technique for problem solving, as a peer to planning and to design.

To my mind, an undergraduate business or management or public administration degree would feature all three core skills. I would like to see model thinking as a quantitative reasoning component of a Liberal Arts degree. A grail, to me, would be a packaging of the “core” that is a useful and attractive offering at the community college level.

“Practices” – Data Science for Judgments and Decision Making or Applied DM – I see this as a concentrations in a containing major.

“Complex Problems” I see as graduate, or professional concentrations. Coursework may draw on active research.

The two tracks highlight the complementary relationship between management decision making and policy design and supporting analysis. Both involve thinking hard about the sponsoring service mission. The analysis serves the judgment and decision to be made. Analysis needs a problem solving context in order to be meaningful. In practice, it is the analyst who is learning enough of the material of the decision makers domain to construct sensible and interpretable analyses for the client decision maker. In practice, there is a short of analysts who can function this way. Decision analytic skill (with an emphasis on framing) is a way for decision makers to cogently reduce the complexity of a problem space without sacrificing consideration of essential characteristics. The ability to think both in terms of models and in terms of the real stakes and interests of those adjacent to a policy is the real superpower of a successful leader. It should be more available.

Aside

As an aside, I should distinguish “model thinking” from Scott E Page’s “The Model Thinker”, which for my convenience I would retitle “The Many-Model Thinker” if I could. My thing is about the idea of models (abstractions with a purpose), what they can do for you (reduce complexity, support communication) and how to understand their limitations (reductiveness, bias). Page’s book, which is excellent but a little more technical than I see for my audience, demonstrates the virtues of applying different kinds of models – ensembles – to complex phenomena. Ideally my collateral on model thinking would be a strong foundation for understanding Page’s “The Model Thinker”.

Reference programs?

My figure is a sketch of a hypothetical program. Is this realistic? I can make an argument about societal benefits – but is there a market for a program like this? If I were to go shopping for something already out there, what might a casual survey find? Some programs are below. I am not mentioning these programs an endorsement - simply that they look like what I am proposing.

Stanford Management Science & Engineering. As an aside, MS&E is what Stanford’s Engineering-Economic Systems has evolved into. E-ES was the spawning ground for a leading generation of decision analysts.

Stanford Doerr School of Sustainability. Notably, the Change Leadership for Sustainability Program

Carnegie-Mellon Engineering and Public Policy (EPP), notably the Science, Technology and Public Policy complementary major and the center for Technology Innovation Policy. EPP has been at CMU for at least 40 years. When I was an undergrad at CMU, in MechE, I had no appreciation whatsoever for “policy” - it seemed very remote from design, which is where my interest was.

Minerva’s graduate decision analysis program. Minerva, both undergrad and in this grad program, has an emphasis on what I’ll call crudely soft skills and as well experiential, small cohort learning. The business side of Minerva has a significant investment in on-line learning.

These are expensive programs at highly selective schools. I’ve not looked at public administration programs, or schools of government, like Harvard’s. So, I have more investigation to do. What am I overlooking? Nonetheless I claim there should be many offerings out there, with access to many more people. The question is why not?