Markets and organizations are shaped by prices, expectations, incentives, institutions, laws, technology, and people. I prefer to begin with that wider context, identify the mechanism that matters, and then narrow the analysis.
Evidence before certainty
Confidence should reflect the quality of the evidence. A forceful argument is useful only if its assumptions and supporting facts can stand up to examination.
History broadens the comparison
The past offers examples of institutions under pressure, decisions made with incomplete information, and consequences that took time to appear. Those examples can improve a question without supplying an automatic answer.
Incentives shape behavior
A stated objective does not fully explain a policy or an organization. The incentives faced by the people involved often provide a more useful guide to what will happen in practice.
Look beyond the immediate effect
People adapt to changes in prices, rules, and technology. The consequences of that adaptation may matter more than the first visible effect of a decision.
Time horizon changes the question
A short-term development and a long-term structural change can point in different directions. An analysis should be clear about the period over which its conclusion is meant to apply.
Judgment remains necessary
Models organize information. They do not decide which assumptions deserve confidence or which omissions matter. Judgment includes recognizing when a model is useful and when its context has changed.
Keep conclusions open to revision
What evidence would change the view? Stating that test makes a conclusion easier to assess and helps prevent commitment to an idea from becoming a substitute for analysis.