Studio 4: Declare and diagnose provisionally
This studio is part of a book in active development and has not yet been through the author’s review. Content may change as the review advances.
What you can defend when you leave. Write your design down as something specific enough to be diagnosed, then find out how it behaves before it costs you anything.
The milestone ahead
This is the hinge of the book. Everything before it shaped a question; everything after it executes a design. The contract you write here (objective, estimand, population, data strategy, a provisional operationalization, warrant, answer strategy, uncertainty) is declared before any result exists, so that later changes are visibly design or visibly drift. Diagnosing it on simulated data now, while changing your mind is still free, is what makes the declaration honest rather than decorative. The four MIDA parts you declare here, and the declare-and-diagnose loop you run on them, were introduced by Blair, Cooper, Coppock, and Humphreys (Blair et al. 2019) and developed in RDSS (Blair et al. 2023); this book puts them in plain language and joins them to a governed AI workflow.
This studio closes with Milestone 4: Your research contract, v0, a short chapter of its own after the lessons. What it asks you to produce. Research Contract v0 — objective, target estimand, population, setting and time, data strategy, a provisional operationalization you mark for revision, warrant, answer strategy and uncertainty statement — plus the design properties you diagnosed (its bias, its wobble, and how often it would detect what you are looking for), your permission status, and a redesign record. Measurement proper is taught and assessed in Studio 6; here you only commit to a starting choice and say why it is defensible for now.
Milestone 5 commits the contract to one research pathway and writes what that pathway can and cannot establish. The milestone chapter keeps the working details: the practice steps, the versioned record, and how the artifact is assessed.
The lessons in this studio
Read them in order, working each lesson’s It is your turn as you go. Each one builds a piece of Milestone 4; beside each lesson is the piece it hands you.
- Lesson 10 — Model, Inquiry, Data Strategy, and Answer Strategy: your MIDA declaration: model, inquiry, data strategy, and answer strategy, with the alignment sentence that ties them together — carrying population, setting and time, the provisional operationalization, and the warrant into Contract v0.
- Lesson 11 — Uncertainty Before You Need It: your estimand and estimator, the repeat-run story behind them, your dependence structure, and your uncertainty sentence next to its wrong twin.
- Lesson 12 — Declaring and Diagnosing a Research Design: your design diagnosis: bias, wobble, and power read from simulation, one redesign, and the honest call that followed.
- Lesson 13 — Research Ethics and Data Governance: your declared permission status, your minimised column list, your AI boundary, and your four governance decisions.
When the last lesson closes, the milestone chapter is next. Nothing you write in a lesson is busywork: every piece above is on the milestone’s checklist.