Problems come in four domains: simple, complicated, complex, and chaotic. Each calls for a different approach. This is a sense making framework, not a categorization method. I tried to capture this in my notes, which appear below.
* Simple. Hierarchy works in simple situations. In the complex system, you need to set the boundaries.
* Ordered system. Constrains agents, reductionism and rules, determinism, observer independence. Bring in an expert or analyze situation. Sense, analyze, respond. Good practice.
* Chaotic systems. Agents unconstrained and independent of one another, studied thru stats and probability. Act, sense, respond. Novel.
* Complex adaptive systems . Lightly constrains agents. Agents modify system thru interaction with it. They co-evolve (making things irreversible). Inherently unpredictable. CAS is non-causal. Some things just are. Can only understand by interacting with it. Probe, sense, respond. Do NOT bring in experts: they are the most conservative people on earth. Bring in diverse experts, if any. Emergent practice.
Learning Research Methods 