Transparency about the discipline
Students read unsettled questions as their own failure to understand
A student cannot reconcile three papers and concludes they are not cut out for this, when they are standing at the frontier.
What it is
Being explicit about which parts of the field are settled and which remain actively argued, and saying where you stand when you present a disagreement.
Syllabi present every question as though it has an answer, because that is what a reading list looks like. A student who then meets three papers that contradict each other has no way to tell whether the field is unresolved or whether they have misread. Naming it removes that ambiguity and tells them uncertainty is a feature of advanced scholarship rather than evidence of their inadequacy.
Run it
- Open each major topic by separating what is settled from what is debated. One slide is enough.
- When you present competing views, say which one you hold and why reasonable people disagree.
- Where a disagreement is genuinely live, hand it to students rather than resolving it for them. See structured debate.
Say this
Opening a contested topic
“Two things before we start. The first half of this is settled, and if you disagree with it after the reading, you are probably misreading it, so come and find me. The second half is not settled. Serious people disagree, I hold one of the positions, and I will tell you which. If you finish today confused about that half, you have understood the state of the field correctly.”
In my courses
Planned for EPI 601, odd weeks throughout.
I will be explicit about which topics in epidemiologic methods remain actively debated (e.g., causal inference frameworks, the role of p-values, controversies in matching). When presenting competing viewpoints, I will acknowledge my own position and explain why reasonable epidemiologists disagree.
EPI 601 is built around unresolved controversies, so this is a closer fit there than in either of the other two courses. The specific disagreements named in the plan are p-values against confidence intervals, matching against adjustment, additive against multiplicative interaction, and the competing causal-inference frameworks. New for 601; not used before the fellowship.
Evidence
No study supports this as a discrete practice, and none of the fellowship materials offers one. The strongest thing that can be said for it is structural rather than empirical. A student who knows which kind of difficulty they are in can decide whether to push or to ask, and a student who does not know defaults to assuming the difficulty is theirs.
The related move of naming which kind of hard a task is sits under reducing negative stress.