Appendix B — For Instructors
EDR|AI is a self-contained manual: a student working alone needs nothing beyond the chapters and their companion Colab notebooks. This appendix is for the other reader — the instructor who wants to teach with the book. It presents the companion course the book grew out of and makes all of its materials freely available.
B.1 Two artifacts, one loop
The book and the course are different artifacts, and keeping them distinct is what lets each travel on its own.
- The book is the manual. Every chapter ends with an It is your turn section, and every chapter has its own companion Colab notebook (the badge at the top of the chapter) where a reader runs the chapter’s code and completes that section. Worked in order, the 40 sections chain into a complete research project; the book stands on its own as a self-study path.
- The course wraps around the book. Its weekly lab notebooks are classroom artifacts derived from the book’s chapters and shaped for live 50-minute sessions, and each one is handed in at the end of its week; its readings are the book’s chapters (required) with matching RDSS chapters recommended; and each chapter’s It is your turn section is collected on the day that chapter’s reading is due, graded for completion. Its milestones then name their book anchor and carry that same work into the semester project, so each graded step of the project is a piece of the book’s chain.
When the book changes, the course material is resynchronized from it: the English edition of the book is the source of truth for everything downstream.
B.2 The companion course
The reference implementation is HONR 46400 — Evidence-Driven Research, a semester-long Honors seminar at Purdue University (three 50-minute meetings a week, individual research projects, a public poster conference, and an oral evidence defense). What students learn is the working structure of contemporary scientific research: turning curiosity into an answerable question, designing a study whose claims are honest by construction, producing and verifying evidence along the main pathways, stating uncertainty like a professional, and defending a claim in public, with AI agents doing the legwork inside that structure. Everything student-facing is open:
- Course website: davi-moreira.github.io/2026F_evidence_driven_research_purdue_HONR464 (syllabus, schedule, material)
- Repository: github.com/davi-moreira/2026F_evidence_driven_research_purdue_HONR464
- Course datasets, one bundle: data.zip
The weekly rhythm: from the second week on, every Monday and Wednesday lecture opens with a ten-minute lab meeting. The instructor asks the room how the projects are going, and the room answers: what someone decided since the last meeting, what the evidence looks like now, where someone is stuck. Nobody is designated to speak, nothing is prepared beforehand by anyone, and nothing said in those ten minutes carries a grade. The instructor leads from minute ten, runs the Socratic investigation on the week’s lab notebook, and owns accuracy, the AI tooling, and the clock: the AI does the legwork live, and every claim it produces is verified before it survives. Friday is a studio: a research stand-up, then a long sprint on the week’s milestone, worked with the student’s AI assistant (reviewer roles from an AI bench — at Purdue, GenAI Studio — are encouraged) and submitted by the Sunday that follows.
An open round needs no machinery. There is no draw, no packet, no announcement of who is up, and nothing to reschedule when someone is away. It reaches every student at every lecture rather than one student per lecture, and by the time the poster session grades a student for describing their own evidence in public, they have been doing it out loud since the second week of the term.
The course was first designed the other way around. A student took the day’s concept and taught it, in the role of Student Research Lead, running the Socratic investigation and assessed on a nine-criterion rubric worth a quarter of the grade. A later version kept the ten minutes but assigned them: one student per lecture held a reporting slot, drawn at random at the start of the term, with a preparation cell to fill in before class. The teaching role and the assigned slot are both retired for this edition, and both are kept intact for a future one, with the handbook, preparation template, rubric, per-lecture question bank, and the slot-draw script, at project/srl/. The question bank is still in daily use; it now belongs to the instructor. If your students are ready to teach each other, that material is there and it works.
B.3 Milestones: grading the book’s chain
Two things share the word “milestone”, and they are different instruments. The book’s Milestone chapters (1–12) close each studio: the reader concludes one by producing the studio’s versioned artifact, and the book’s own rubric sits on that chapter. The course milestones below (M1–M16) are the weekly graded deliverables of the companion course. They are aligned but not identical; when this page says “milestone” unqualified, it means the course’s.
The project runs through sixteen milestones, M1–M16, one per week, from a first curiosity map to a reproducible package a stranger can rerun. Each milestone brief names its book anchor: the chapters whose It is your turn sections that stretch of the project is built on. Those sections are handed in earlier, each on the day its own chapter’s reading is due, and graded for completion. The milestone then grades what the student’s own project made of them, so grading it is grading a stretch of the book’s chain. The briefs are open in the repository: _research_project/2026Fall/.
B.4 The course lab notebooks
One lab per week, nb01–nb16, each derived from the chapters it teaches. These are the classroom counterparts of the book’s companion notebooks — richer simulations, the lab-meeting opener, and session structure — and any of them opens in free Colab.
The lab is also what the course collects each week, and the rule is completion. Students hand in the notebook they worked in class by 11:59 PM on the Sunday that ends the week, and the last week closes on the final class day instead: full credit on time, half credit within seven days, none after that. Sixteen submissions, the two lowest credits dropped, and the block carries 20% of the course grade. Nothing is scored on whether the answers came out right, and nothing said at the lab meeting is scored at all, so the hour itself stays a place where a student can be wrong out loud.
That block is one of three completion contracts, and together they set the shape of the grade: attendance 1%, participation 9%, the chapters’ It is your turn sections 15%, the weekly lab notebooks 20%, and the final project 55%. Forty-five of those points go to showing up and doing the work. The research itself is judged in the remaining 55, across the milestone chain, the peer evaluations, the peer review, the poster, and the instructor’s evaluation.
Every lab is linked, with what it does, in the lesson-to-lab adoption table below — that table is generated from the course crosswalk, so it cannot drift from the book’s structure the way a hand-kept list can.
B.5 Adopting or adapting
The book is deliberately institution-agnostic, and the course material above is open: take the whole architecture, one lab, or nothing but the reading sequence. Instructor-side material — solution notebooks and per-meeting session guides — lives in a private repository linked from the Instructor tab of the course website; write to the author for access (About the Author). If you teach with EDR|AI, the author would genuinely like to hear how it went.
B.6 Grading rubrics — the “It is your turn” sections
One rubric per chapter, derived from the chapter’s numbered steps. The same rubric appears in the chapter’s companion notebook, so the standard the reader worked to is the standard you apply when the project carries that work. The scale, for every rubric: 0 missing · 1 attempted · 2 complete, project-specific, and verified.
B.6.1 Ch. 1 — From Curiosity to a Research Problem
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 18 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Copy your opening move’s four lines here, date them, and label them version zero: the first entry of your project’s record | |
| Step 2 | Take the strongest curiosity you now hold and walk it down the funnel in writing: experience, curiosity, topic, research problem | |
| Step 3 | Run the ownership test on the problem you are converging on: would you still chase it if every AI tool disappeared tomorrow, and is the scope one you… | |
| Step 4 | Score it on importance, feasibility, and contribution, one honest sentence each | |
| Step 5 | Write a second candidate problem from the same curiosity and scope it the same way | |
| Step 6 | Under the problem you kept, write your expected answer in one sentence, and name one real person or role for whom a different answer would be… | |
| Step 7 | Ask an AI to attack your chosen problem on all three tests, and keep only the objections that survive your own reasoning | |
| Step 8 | Finish by handing your problem sentence to someone who knows nothing about your project | |
| + | Craft and verification record: every AI activity has one line in your log, and every fact you kept was checked before you relied on it |
B.6.2 Ch. 2 — AI Is Your Arm, Not Your Brain
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 12 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Open a blank spreadsheet or document and title it AI Research Ledger | |
| Step 2 | Open your ledger’s first rows backward, one for each Studio 1 activity you logged: the brainstorm, the source search, the candidate list, and the… | |
| Step 3 | Under your chosen problem, write the single fact you would need to know before you could take it further | |
| Step 4 | Take the fact you just named, and ask an AI tool for three published sources that speak to it | |
| Step 5 | Log the exchange in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.3 Ch. 3 — You as a Research Director
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Take the research problem you committed to, and list every task the project around it would need: finding sources, gathering or cleaning data,… | |
| Step 2 | Sort every line into safe to delegate, delegate then verify, or never delegate, with a one-line reason beside it | |
| Step 3 | For each “delegate then verify” line, name the check now, while it is cheap: which document you will open, which number you will recompute, which… | |
| Step 4 | Star the two lines you would defend hardest as never-delegate, and write one sentence each on why those calls are what make the project yours | |
| Step 5 | Hand the same list to an AI tool, ask it to argue with your placements, and keep only the changes you can justify out loud | |
| Step 6 | Log that exchange in your AI Research Ledger, and verify at least one placement with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.4 Ch. 4 — Specify, Delegate, Interrogate, Inspect, Verify, Document, Defend
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Pick one genuine, checkable claim — your opening move’s starting belief is the natural choice, because checking it sharpens the question you will… | |
| Step 2 | Specify. Before you open any tool, write your own expected answer and what a fair way of getting it would look like | |
| Step 3 | Delegate, Interrogate, Inspect. Hand the task over | |
| Step 4 | Verify. Run one independent check and write its outcome down either way — “nothing changed” is a result worth recording | |
| Step 5 | Defend. With the tool closed, write one sentence stating the result in your own words, and one sentence naming what it does not establish | |
| Step 6 | Document the whole run in your AI Research Ledger, naming the verification method you used from the Verification Guide and the check’s outcome —… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.5 Ch. 5 — Research Responsibility and Intellectual Ownership
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your ownership statement: three or four sentences naming the work you intend to put your name on, what you will personally answer for, and… | |
| Step 2 | Write your never-delegate list for this specific project | |
| Step 3 | Draft the AI-use disclosure you would attach to your work as it stands today, built from your ledger rows rather than from memory | |
| Step 4 | Read the disclosure and the ledger side by side | |
| Step 5 | Find one sentence you have already written that you could not defend if someone asked how you know it is true | |
| Step 6 | Log all of it in your AI Research Ledger, and verify at least one surviving claim with a named method from the Verification Guide: primary-source… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.6 Ch. 6 — Choose Your Question’s Kind and Reach
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your research problem out as questions, three or four of them, each a sentence evidence could answer and be wrong about | |
| Step 2 | Classify each question on both axes, kind and reach, and name the position | |
| Step 3 | Hunt the double-barrels | |
| Step 4 | Pick your lead question, the one your project will be built around | |
| Step 5 | If your lead question is causal, describe the comparison world it needs: the version of events that did not happen and that your design will have to… | |
| Step 6 | Log the classification in your AI Research Ledger, and verify it with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.7 Ch. 7 — Declare Your Research Question
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 16 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your field card first, from your own lead question as it stands: objective, unit of analysis, outcome, conditions, kind, reach | |
| Step 2 | Delegate the wording, under show-what-changed: Candidate wordings (you choose; the tool accounts for itself). ```text Here is my research… | |
| Step 3 | Choose the final wording and write one sentence on why it beat the others, including yours if yours lost | |
| Step 4 | Write the boundary pair: the sentence you hope to defend, and the stronger sentence you will not be able to defend, each in one line | |
| Step 5 | Write the declaration’s uncertainty-and-limitations line: the key limitation you already foresee in answering these words with the evidence you can… | |
| Step 6 | Run the stranger test on the declaration: hand the question and the card to someone who knows nothing about the project, and ask them to say what one… | |
| Step 7 | Log the declaration in your AI Research Ledger and date it: this is version zero of the sentence your whole project answers to, and the milestone… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.8 Ch. 8 — Research Builds on Research
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your lead question at the top of a fresh page | |
| Step 2 | Search once yourself before you delegate | |
| Step 3 | Ask an AI for five more candidates on the same question, each with authors, year, exact title, and venue | |
| Step 4 | Snowball your seed by hand | |
| Step 5 | Put all of them in one candidate list with a status column, and write lead in every row |
|
| Step 6 | Log the delegation in your AI Research Ledger, and verify at least one candidate now with a named method from the Verification Guide; primary-source… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.9 Ch. 9 — Finding and Verifying Prior Evidence
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 16 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Try to retrieve every candidate on your list | |
| Step 2 | Open the ones that resolve and read them | |
| Step 3 | Criticize each one | |
| Step 4 | Draw your evidence map: one node per verified source, lines between the ones that agree, and a marked line between any two that disagree | |
| Step 5 | Find the quiet spot on the map and write your gap as one bounded sentence: “Across the sources I retrieved and verified, ___ has not been established… | |
| Step 6 | Close the loop on your declaration: rewrite your declared question in light of the map, or defend leaving it unchanged with a reason tied to a… | |
| Step 7 | Log the verification work in your AI Research Ledger, naming the method you used from the Verification Guide; primary-source reading is the one that… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.10 Ch. 10 — Model, Inquiry, Data Strategy, and Answer Strategy
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your model in three or four sentences of plain language | |
| Step 2 | Write your inquiry as one sentence naming one quantity: an average, a difference, a rate, a share | |
| Step 3 | Write your data strategy: who or what gets sampled, from which list, and who gets which condition, if any | |
| Step 4 | Write your answer strategy: the arithmetic that turns your data into your inquiry | |
| Step 5 | Audit the alignment yourself and write the one sentence that matters: “Because my data strategy is ___, the strongest honest wording of my result is… | |
| Step 6 | Log the draft in your AI Research Ledger, then verify the alignment claim with a named method from the Verification Guide; peer reasoning or a causal… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.11 Ch. 11 — Uncertainty Before You Need It
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | State your estimand in one sentence, as a quantity in the world with a population, a setting, and a time | |
| Step 2 | State your estimator as a recipe: what you will collect, and exactly what you will compute from it | |
| Step 3 | Name what would differ on a repeat. Write the two or three things that would come out differently if selection and measurement ran again for the… | |
| Step 4 | Name your dependence structure. List the levels at which your observations were sampled, repeated, or connected, and how many units you have at… | |
| Step 5 | Write your uncertainty sentence, and then write the wrong version of it. State what your interval will and will not claim | |
| Step 6 | Log it. Add your AI Research Ledger rows for anything you delegated here, and record which decisions you kept | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.12 Ch. 12 — Declaring and Diagnosing a Research Design
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Declare your design as something that can be run | |
| Step 2 | Diagnose it | |
| Step 3 | Name the worst of the three, and say which kind of problem it is | |
| Step 4 | Redesign once | |
| Step 5 | Make the honest call in one sentence: run it as redesigned, redesign again, or narrow the claim to what this design can actually deliver | |
| Step 6 | Log both diagnoses in your AI Research Ledger, and verify the key number with a named method from the Verification Guide; simulation is the method… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.13 Ch. 13 — Research Ethics and Data Governance
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 16 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Run the determination. Answer the three questions in writing: is your knowledge meant to travel, do identifiable living people enter your study,… | |
| Step 2 | Declare one permission status — cleared, formal determination required, pending, or not authorized — and name the competent authority at your… | |
| Step 3 | List your columns and minimise them. Write every variable you plan to collect, then cross out the ones your declared analysis does not require | |
| Step 4 | Run the uniqueness check on your planned quasi-identifiers, using the companion notebook | |
| Step 5 | Write your AI boundary in two lines: what you will send to an AI tool on this project, and what you will never send | |
| Step 6 | Write your four governance decisions — where the data live, who can open them, how long you keep them, and what happens at the end | |
| Step 7 | Write your stop plan. In one sentence: if your determination comes back “not authorized”, what is the version of this question you would ask… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.14 Ch. 14 — Observational Descriptive Research
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Decide which case you are in | |
| Step 2 | If this is your pathway, write your four boxes, one line each: target population, accessible population, sampling frame, sample | |
| Step 3 | Draw your selection diagram: an arrow from each box to the next, with the filter that removes people written on the arrow | |
| Step 4 | Write the two sentences that define your claim | |
| Step 5 | If this is not your pathway, run the classification drill instead | |
| Step 6 | Log the work in your AI Research Ledger, and verify the key claim with a named method from the Verification Guide; simulation fits here, because the… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.15 Ch. 15 — Observational Causal Research
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 16 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Decide which case you are in | |
| Step 2 | If this is your pathway, write your causal question as exactly two things: a treatment (the thing that varies) and an outcome (the thing you… | |
| Step 3 | Draw the causal diagram | |
| Step 4 | Name your leverage and its price | |
| Step 5 | Write down one number you would recompute by a second route to check any estimate a collaborator or an AI hands you, and say what result would make… | |
| Step 6 | If this is not your pathway, run the classification drill | |
| Step 7 | Log all of it in your AI Research Ledger, and verify with a named method from the Verification Guide; the causal diagram is the method built for this… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.16 Ch. 16 — Experimental Descriptive Research
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Read your lead question aloud and underline the verb | |
| Step 2 | If it is yours, name the latent characteristic your project has to reveal, the property nobody publishes anywhere, and the controlled stimulus you… | |
| Step 3 | Write your estimand in one sentence: the exact quantity, for which units, described as it stands | |
| Step 4 | Name the one demand effect and the one instrument effect you most fear, each with the redesign that rules it out | |
| Step 5 | If this pathway is not yours, run steps 2 to 4 as a drill anyway, treating your question as if you had to measure it rather than intervene | |
| Step 6 | Log the step in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.17 Ch. 17 — Prediction and Generalization
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Ask the blunt version of your question: does your project need to guess an outcome for a case whose outcome does not exist yet? | |
| Step 2 | If yes, sign the four-part contract, in order | |
| Step 3 | List your features and write beside each one the moment its value is settled | |
| Step 4 | Write the boundary in one sentence: the cases your forecast covers, and the ones a reader should not assume it covers | |
| Step 5 | If prediction is not your question, write the line that says so plainly, then run step 3 anyway | |
| Step 6 | Log the step in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.18 Ch. 18 — Experimental Causal Research
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your treatment and your outcome as two short phrases | |
| Step 2 | Take one real unit from your study and write Y(1) and Y(0) for it in plain words | |
| Step 3 | Name the quantity your estimate targets and for whom, in one sentence a stranger could repeat back to you correctly | |
| Step 4 | Name the threat nearest your design, whether attrition, noncompliance, or spillover, and the check you will run in your own data to see how bad it… | |
| Step 5 | If you cannot randomize, run steps 1 to 4 as a drill, then name the confounder that random assignment would have killed for you and the observational… | |
| Step 6 | Log the step in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.19 Ch. 19 — Hybrid and Complex Designs
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your design as an ordered list of the measurements and stages it takes to get from raw data to your answer | |
| Step 2 | For each line, name which MIDA part it lives in and the single quantity it targets | |
| Step 3 | Pick the piece whose error would hurt most and trace how its uncertainty reaches your final answer | |
| Step 4 | Decide, and write the one-sentence justification | |
| Step 5 | Name the systematic error in what remains that more data will never shrink, and what you would have to change instead: a different source, a wider… | |
| Step 6 | Log the step in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.20 Ch. 20 — Data Provenance and Data Quality
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 16 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | List every dataset, table, and borrowed number your project uses | |
| Step 2 | For each, write four things: who produced it, when, from what original record, and every hand it passed through on the way to you | |
| Step 3 | Open the primary source behind the entry your headline claim leans on hardest | |
| Step 4 | Beside each entry, write its definition in your own words and mark whether it matches the definition your question needs | |
| Step 5 | Save the record as a file that travels with your project, and note the date you retrieved each source | |
| Step 6 | Close with the studio’s governance trio: write and date your acquisition route, update where the data live and who can open them, and recheck the… | |
| Step 7 | Log the step in your AI Research Ledger, and verify at least one entry with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.21 Ch. 21 — Measurement and Operationalization
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 16 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | List every concept your question contains and circle the abstract ones, the words no instrument reads: engagement, health, participation, quality,… | |
| Step 2 | Under each circled concept write the construct you will really measure, plus one line on why that facet and not a neighboring one | |
| Step 3 | Under each construct write the indicator: the exact procedure and the exact number it produces, specific enough that a stranger could repeat it and… | |
| Step 4 | Under each indicator finish this sentence in your own words: “this number does not capture ___.” Use the facet that worries you most, and add who… | |
| Step 5 | Write one sentence saying what your number means and one saying what you will do with it | |
| Step 6 | Plan one reliability check that matches an error source you actually worry about, then run it: the same units measured twice (occasions), two coders… | |
| Step 7 | Log the step in your AI Research Ledger, and verify at least one measure with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.22 Ch. 22 — AI as Programmer
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your quantity of interest in one sentence, and the frame it runs over in a second | |
| Step 2 | Prompt an AI tool for the code that computes that quantity over that frame | |
| Step 3 | Read the returned code line by line and ask what each line removes | |
| Step 4 | Print the shape of the data the number was computed over, and the row count before and after anything that can drop rows | |
| Step 5 | Write one sentence naming what this number does not cover: the cases your frame left out, and what you therefore cannot say about them | |
| Step 6 | Log this first number in your AI Research Ledger, and verify it with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.23 Ch. 23 — AI as Analytical Assistant
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 16 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Before the assistant touches anything, write your headline estimate in one sentence and the answer you expect | |
| Step 2 | Hand over one well-specified task at a time | |
| Step 3 | Keep a cycle log as you go: one line per turn with what you asked, what came back, and what you changed and why | |
| Step 4 | Re-derive at least two numbers yourself, by hand or with a second simple expression, before any of them enter a claim | |
| Step 5 | Run a placebo: shuffle your group labels, re-run the same code unchanged, and confirm the fake gap lands in the ordinary part of that pile | |
| Step 6 | Close with the milestone’s three checks: restart and run everything from a clean state and confirm the headline numbers match, record your… | |
| Step 7 | Log the analysis in your AI Research Ledger, cycle log attached, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.24 Ch. 24 — Robustness and Sensitivity
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 18 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your headline estimate in one sentence | |
| Step 2 | Pre-list the checks you will run, at least two, aimed at the handles a hostile reader would attack first | |
| Step 3 | Before you run anything, run your list through the gate | |
| Step 4 | Run every check on your list | |
| Step 5 | Ask what every version on your list has in common | |
| Step 6 | Add one null check, and name it with the chapter’s three labels, because each label comes with its own procedure and its own reading | |
| Step 7 | Read back your AI cycle log and mark any re-prompt that followed a disappointing number | |
| Step 8 | Log the pre-listed plan and its full results in your AI Research Ledger, and verify at least one output with a named method from the Verification… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.25 Ch. 25 — Diagnostics and Negative Tests
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Name the artifact you are most afraid of: the one way your procedure, rather than your subject, could have manufactured your result | |
| Step 2 | Choose the negative test that could catch it, whether that is a placebo label, a falsification outcome measured where the effect cannot yet exist, or… | |
| Step 3 | Before you run anything, write down two things | |
| Step 4 | Run one diagnostic beside it | |
| Step 5 | Record what you got against what you predicted, and if the estimate moved, say by how much | |
| Step 6 | Log both tests in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.26 Ch. 26 — AI as Adversarial Reviewer
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write a one-paragraph summary of your design, your headline claim, and the checks you have already run | |
| Step 2 | Commission the review: ask for the single most serious flaw and the exact measurement that would confirm or refute it | |
| Step 3 | Take the three hardest points you got back | |
| Step 4 | Run all three checks against your own data | |
| Step 5 | Find the flaw that no check can fix, the one that is a boundary problem rather than a measurement problem, and narrow your claim until the claim is… | |
| Step 6 | Log the review and every adjudicated flag in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.27 Ch. 27 — Recognizing False Confidence
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | List every number currently in your project that you have not personally recomputed | |
| Step 2 | For your headline number, write down where your confidence in it actually comes from: your own recomputation, a tool’s fluent summary, or the fact… | |
| Step 3 | Write your boundary sentence, the claim your design licenses and nothing beyond it | |
| Step 4 | Name the one place you are most likely to let a confident answer past unchecked, and write it down | |
| Step 5 | Adopt a standing rule for the rest of the project: no AI-reported number reaches your paper, note, or poster until you have recomputed it by a second… | |
| Step 6 | Log the audit in your AI Research Ledger, and verify your headline number with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.28 Ch. 28 — From Results to Claims
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 12 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | List every result you might put in front of a reader, then write the claim sentence each one licenses | |
| Step 2 | Check every verb against what your design established, and downgrade the ones that reach too far | |
| Step 3 | Write a boundary line beside each claim naming what it does not establish, and put the uncertainty in the same eye-span as the number | |
| Step 4 | Trace every number back to the cell that produced it and cut the ones you cannot walk back | |
| Step 5 | Log it in your AI Research Ledger, and verify at least one claim with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.29 Ch. 29 — Claim–Evidence Tables
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 12 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write one row per claim with four columns: claim, evidence, trace, boundary | |
| Step 2 | Fill the evidence column with the actual result and its uncertainty, not a description of the result | |
| Step 3 | Fill the trace column by opening the cell or the source and reading the number off the screen | |
| Step 4 | Write each boundary as a sentence about what the row does not establish, then hunt for blank cells | |
| Step 5 | Log it in your AI Research Ledger, and verify at least one traced number with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.30 Ch. 30 — AI Disclosure and Research Integrity
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 12 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Read your AI Research Ledger end to end and list every use, including the one you would rather not mention | |
| Step 2 | For each use, write the three parts a disclosure needs: which tool, for which task, and how you verified the output | |
| Step 3 | Turn the rows into a short statement in your own words, then check every sentence against a row | |
| Step 4 | Say what you did not verify | |
| Step 5 | Log it in your AI Research Ledger, and verify the finished statement with a named method from the Verification Guide, such as peer reasoning | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.31 Ch. 31 — From Dossier to Research Note
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 16 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your problem–gap–question spine as three short paragraphs | |
| Step 2 | Order your findings, best-supported first | |
| Step 3 | Write your headline claim in one sentence, with its boundary inside the same sentence: the setting, the sample, and the population it does not cover | |
| Step 4 | Size your limitations to what actually threatens the claim | |
| Step 5 | Assemble the full note, not only its spine: methods (what you did, so a stranger could follow), results with their uncertainty, a discussion sized to… | |
| Step 6 | Attach your folder to the note and confirm the two agree: the number in your lead sentence is the number a cold restart-and-run-all prints | |
| Step 7 | Log the drafting round in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.32 Ch. 32 — Research Posters
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 12 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Choose the single figure that carries your lead claim, and draw it with the axis running the full scale of the measure | |
| Step 2 | Print each value as a label on the mark it belongs to, and show the uncertainty the design licenses right on the figure so the uncertainty arrives in… | |
| Step 3 | Carry every meaningful distinction in at least two channels, so color is never the only signal a reader can lose | |
| Step 4 | Write the headline above the figure using the verb your claim boundary allows, and lay out the rest of the page so a stranger finds that headline in… | |
| Step 5 | Log it in your AI Research Ledger, and verify at least one number on the page with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.33 Ch. 33 — Poster Criticism
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 12 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Run the gallery-walk audit on your own draft as if you had never seen it: claim boundary, figure honesty, read path, uncertainty, accessibility | |
| Step 2 | Rank the hits by how far the error travels | |
| Step 3 | Field the three lenses out loud: an outsider, a specialist, and a skeptic with a rival explanation | |
| Step 4 | Revise once, fixing the hits in ranked order, then make the lock call | |
| Step 5 | Log it in your AI Research Ledger, and verify at least one fixed number with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.34 Ch. 34 — Research Pitches, Talks, and Seminars
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your boundary sentence first, on its own line, and mark it | |
| Step 2 | Build the 30-second version: hook, claim, boundary, invitation, with no evidence beats yet | |
| Step 3 | Build the 3-minute version: the same skeleton with two or three evidence beats and the same boundary sentence, word for word | |
| Step 4 | Expand to your venue’s actual length: grow the 3-minute spine section by section (problem and gap, question, method, evidence, boundary and… | |
| Step 5 | Say the versions out loud to someone outside your field and have them repeat your claim back to you | |
| Step 6 | Log it in your AI Research Ledger, and verify at least one compressed sentence with a named method from the Verification Guide, such as peer reasoning | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.35 Ch. 35 — Difficult Questions and Uncertainty
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 12 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write the five hardest fair questions someone could ask you, one for each type: method, alternative explanation, why-it-matters, generalization, AI… | |
| Step 2 | Draft a two-sentence answer to each and mark it defend or concede | |
| Step 3 | Write your uncertainty-and-limitations statement in ULN form: how much your number could wobble and why, the one thing your design cannot show, and… | |
| Step 4 | Read it aloud and cut every self-erasing word, replacing each with the exact boundary it was hiding | |
| Step 5 | Log it in your AI Research Ledger, and verify at least one prepared answer with a named method from the Verification Guide, such as peer reasoning | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.36 Ch. 36 — Replication and Reproduction
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 16 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Gather everything into one folder: the data, the code, the run order, the random seed, and a short write-up that states your headline number in words | |
| Step 2 | Reproduce yourself cold | |
| Step 3 | Line every sentence of your write-up against a number your code actually prints | |
| Step 4 | Change one defensible choice you never disclosed, an exclusion rule or a cutoff or a subset, and record how far your headline moves | |
| Step 5 | Classify every missing value before you fill any of them | |
| Step 6 | Name the assumption your result would collapse without, and rank everything you found by threat to the claim rather than by ease of repair | |
| Step 7 | Log the audit in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.37 Ch. 37 — Open and Reusable Research Packages
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Assemble the five parts for your own project: the runnable notebook, the data-provenance note, the fixed seed, the decision log, and your AI-use… | |
| Step 2 | Run the chapter’s audit on your own key lines and fix what it flags, starting with the sin you were most tempted to leave alone | |
| Step 3 | Write the README a stranger reads first: what the project asks, what the headline number is, which file produces it, and in what order to run things | |
| Step 4 | Rerun it cold in a second environment, a clean Colab session or a machine that is not yours, and check the headline returns within rounding | |
| Step 5 | Write one honest sentence about the limit of all this: your capsule is runnable, which is not the same as correct, and name the choice inside it you… | |
| Step 6 | Log the packaging round in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.38 Ch. 38 — Managing Multiple AI Agents
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Pick one real section of your project to revise: your methods, your results, or your limitations | |
| Step 2 | Set up three loops, each with a one-sentence job you could grade | |
| Step 3 | Wire them, and declare each role’s capability while you do: who executes it, who chooses its next step (you, a script, or the tool), what tools it… | |
| Step 4 | Name the one node you keep human, the step where the loops’ work becomes your claim, and write down why no role is allowed to take it | |
| Step 5 | Run the round | |
| Step 6 | Log every loop in your AI Research Ledger, and verify at least one output with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.39 Ch. 39 — Conflicting Agents and Human Escalation
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Write your escalation rules before the next round starts | |
| Step 2 | Add the stopping conditions that are not about content | |
| Step 3 | Take one load-bearing pair of outputs your roles have returned, whether they agree or disagree, and diagnose it on the record: real disagreement,… | |
| Step 4 | Resolve it yourself with a check the loops cannot run: recompute the number, open the source, or compare against a figure that was never in any prompt | |
| Step 5 | Take one place your roles agreed on something that mattered and test whether the agreement was independent | |
| Step 6 | Log the pair you examined, your diagnosis, and the check that settled it in your AI Research Ledger, with the override on record when you made one,… | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.40 Ch. 40 — Final Research and AI-Management Portfolio
Use this as a self-check while you work. It is also the bar the same work meets later, once your project carries it. Each row: 0 missing, 1 attempted but incomplete, generic, or unverified, 2 complete, specific to your own project, and verified where a check applies. 14 points in all.
| # | Criterion | 0–2 |
|---|---|---|
| Step 1 | Finish the research artifact itself: your paper, note, or poster, with each claim stated at its boundary and your uncertainty in the same sentence as… | |
| Step 2 | Assemble your AI-management portfolio around the team you directed in this studio: the decomposition of loops with their wiring, one conflict or… | |
| Step 3 | Write your stopping rule as a sentence you could read aloud, phrased around what you have verified rather than what your tools approved | |
| Step 4 | Rehearse the defense in three movements: your claim and its boundary, the central choice you made and the road you did not take, and the verification… | |
| Step 5 | Read the whole artifact once more against your ledger and cut anything you cannot trace to a check | |
| Step 6 | Close the ledger with the last entry, and verify your headline claim one final time with a named method from the Verification Guide | |
| + | Craft and verification record: AI use logged in your AI Research Ledger, claims stated with their uncertainty, and each key claim verified with a named method |
B.6.41 Which lab goes with which lesson
Each lesson’s companion notebook belongs to the book. The table below maps every lesson to the classroom lab that carries it in the companion course, and says what that lab does.
| Lesson | Course lab | What the lab does |
|---|---|---|
| Ch. 1 | nb01 — Studio 1: Frame the inquiry — from curiosity to a research problem | You will watch one curiosity sharpen the instant it touches real survey columns, then walk your own curiosity from your nb01 map down the funnel, draft two candidate problems, and defend the one you keep. |
| Ch. 2 | nb02 — Studio 2: Govern the work — your rules and your question | In the lab you run a small seeded simulation showing why how confident a source sounds tells you nothing about whether it is real, then take one topic you care about, ask an AI for scholarly sources, and verify each by retrieving it yourself. |
| Ch. 3 | nb02 — Studio 2: Govern the work — your rules and your question | In the “safe to delegate, verify first, or never delegate” practice you sort five real research tasks into the three buckets and defend each call, and at the Human-Only Checkpoint you close the AI and name, for your own project, two decisions you will never hand over. |
| Ch. 4 | nb02 — Studio 2: Govern the work — your rules and your question | There you run the protocol live: you catch a fabricated citation by retrieving it yourself, then catch a confident summary that overstates a number the data never held. Both are the Inspect and Verify steps under real pressure. |
| Ch. 5 | nb02 — Studio 2: Govern the work — your rules and your question | There you sort real research tasks by how safely they delegate, catch a live citation by retrieving it yourself, and fill your first AI Research Ledger row: the record that turns “I checked it” into something you can show. |
| Ch. 6 | nb02 — Studio 2: Govern the work — your rules and your question | There you walk your own curiosity down to a question, watch a naive keyword sorter get fooled on a deck of eight questions, and place your own question on the compass in the shape you will reuse every time you declare a design. |
| Ch. 7 | nb02 — Studio 2: Govern the work — your rules and your question | — |
| Ch. 8 | nb03 — Studio 3: Ground it in verified evidence | In the lab you watch a pile of fifteen AI citations collapse to the few that survive retrieval, build an evidence map in code, and run the retrieval-verification loop on your own claim. |
| Ch. 9 | nb03 — Studio 3: Ground it in verified evidence | In the lab you run the retrieval-verification loop on your own claim and watch a simulation collapse a tidy pile of fifteen AI citations down to the handful that survive retrieval, then build an evidence map whose one empty square is your gap. |
| Ch. 10 | nb04 — Studio 4: Declare and diagnose provisionally | There you build a small simulated world, write its four MIDA parts, and watch what happens to your answer when a data strategy quietly stops matching the inquiry. |
| Ch. 11 | nb04 — Studio 4: Declare and diagnose provisionally | There you write your own design’s four parts and watch the pile of answers move when you change one of them. |
| Ch. 12 | nb04 — Studio 4: Declare and diagnose provisionally | In the lab you declare the notebook’s mentoring design in a few lines of Python, run it thousands of times to read its bias, variance, and power with your own eyes, then watch one redesign rescue a fair-but-powerless study while a confounded twin stays wrong at any sample size. |
| Ch. 13 | nb04 — Studio 4: Declare and diagnose provisionally | There you run the re-identification check on your own planned columns and find out which ones you can drop. |
| Ch. 14 | nb05 — Studio 5: Develop the pathway — the route hub | In nb05’s route hub you meet this route as one of five cards: what a description without intervention can estimate, and the claim it never licenses. If it is YOUR route, you apply its full lesson to your project and defend the choice against one assigned contrast. |
| Ch. 15 | nb05 — Studio 5: Develop the pathway — the route hub | In nb05’s route hub you weigh this route’s card: a causal effect without assignment demands an identification argument, and a causal question with weak data stays causal — currently unidentified, never relabeled. Your-route students carry the argument into their pathway declaration. |
| Ch. 16 | nb05 — Studio 5: Develop the pathway — the route hub | In nb05’s route hub you meet the card that surprises most people: random assignment used to MEASURE, not to estimate an effect. The jigsaw advocate for this route explains why assignment does not automatically make an inquiry causal. |
| Ch. 17 | nb05 — Studio 5: Develop the pathway — the route hub | In nb05’s route hub you meet prediction as its own answer objective: generalizing to unseen cases, licensed by held-out honesty rather than identification. Your-route students commit to the declared-protocol discipline their later weeks will verify. |
| Ch. 18 | nb05 — Studio 5: Develop the pathway — the route hub | In nb05’s route hub you weigh the assigned-intervention card: what randomization licenses, what attrition can still break, and what the route costs to run. Your-route students carry the design into their declaration. |
| Ch. 19 | nb05 — Studio 5: Develop the pathway — the route hub | In nb05 you add the hybrid overlay only if your design has distinct stages, naming the route and claim licence at each stage. If your design has one stage, this lesson stays a reference. |
| Ch. 20 | nb06 — Studio 6: Govern data and measurement | In nb06 you document how your data actually reached you — who produced them, how, when, under what terms — and recheck your permission determination against the file that arrived, not the file you planned for. |
| Ch. 21 | nb06 — Studio 6: Govern data and measurement | In nb06 you build the concept → construct → indicator chain for your own project, check reliability on items (never by splitting respondents), argue validity as interpretation-and-use, and version your Contract where the settled measure differs from the provisional one. |
| Ch. 22 | nb07 — Studio 7: Produce a reproducible first analysis | In nb07 you translate your declared analysis into an executable pipeline, verify every delegated line against a known answer, and produce a first result with its uncertainty. The same week’s second lecture attacks that result from a clean restart. |
| Ch. 23 | nb07 — Studio 7: Produce a reproducible first analysis | In nb07 you demand the execution record, restart and run all from a fresh state, rederive key numbers independently, and check every claim against a specific output — the difference between a result and a story about one. |
| Ch. 24 | nb08 — Studio 8: Stress-test and adjudicate | In nb08 you pre-list a robustness grid before seeing any outcome, run commensurable panels, and learn why a same-sign specification spread is a direction plus a range, never an uncertainty interval. |
| Ch. 25 | nb08 — Studio 8: Stress-test and adjudicate | In nb08 you run licensed negative tests and read each against what an honest check could detect — a permutation check against its justified null, a negative control against its own estimate and uncertainty — never against exact zero. |
| Ch. 26 | nb08 — Studio 8: Stress-test and adjudicate | In nb08 you commission an adversarial AI review of your own analysis and adjudicate every flag against a real data check, cataloguing the confident wrong flag that proves confidence is not evidence. |
| Ch. 27 | nb08 — Studio 8: Stress-test and adjudicate | In nb08 you close the stress-test by hunting false confidence: correlated reviewers agreeing for the same wrong reason, checks that cannot fail, and the complete-case contrast narrated as an effect among stayers. |
| Ch. 28 | nb09 — Studio 9: Write, bound, and disclose | In nb09 you convert audited results into claims that carry their kind and their reach, and catch the quiet verb upgrades that turn an association into a cause under compression. |
| Ch. 29 | nb09 — Studio 9: Write, bound, and disclose | In nb09 you build the claim-evidence table for your project: every claim a row naming its evidence, its verification, and its boundary. A claim with no row does not ship. |
| Ch. 30 | nb09 — Studio 9: Write, bound, and disclose | In nb09 you draft your AI disclosure statement FROM your ledger, never from memory, and check it against what the ledger actually records — including the delegations you rejected. |
| Ch. 31 | nb09 — Studio 9: Write, bound, and disclose | In nb09 you write the bounded research note itself — the core artifact every later genre adapts — with uncertainty in the same eye-span as the claim. |
| Ch. 32 | nb10 — Studio 10: Adapt and defend — the venue contract and the artifact | In nb10 you write your venue contract and adapt the bounded note into the artifact that satisfies it, labelled as the preliminary edition it honestly is. |
| Ch. 33 | nb10 — Studio 10: Adapt and defend — the venue contract and the artifact | In nb10 you run the criticism gallery on your own artifact: what the adaptation may lose, what it may never lose, and whether every number still traces to a cell. |
| Ch. 34 | nb10 — Studio 10: Adapt and defend — the venue contract and the artifact | In nb10 you build the layered pitch under the one rule that governs it: compression never expands the claim. |
| Ch. 35 | nb10 — Studio 10: Adapt and defend — the venue contract and the artifact | In nb10 you rehearse the hardest fair questions with spoken uncertainty and the defend-or-concede call. |
| Ch. 36 | nb15 — Studio 11: Reproduce and package | In nb15 you cold-run your own work from a clean environment, reproduce the uncertainty as well as the point estimate, and label an author’s own rerun honestly: a solo proxy, never independent verification. |
| Ch. 37 | nb15 — Studio 11: Reproduce and package | In nb15 you assemble the reusable package — README, data or an access recipe, code, environment record, licence — and it is those reproduced numbers that lock onto the poster. |
| Ch. 38 | nb16 — Studio 12: Release and direct the next cycle | In nb16 you wire worker and critic roles for the final push and learn why four agreeing agents can share one blind spot. |
| Ch. 39 | nb16 — Studio 12: Release and direct the next cycle | In nb16 you resolve a live conflict between agents by escalating to the only reviewer who can settle it: a human rerunning the check. |
| Ch. 40 | nb16 — Studio 12: Release and direct the next cycle | In nb16 you run the release audit on your own package, decide release or withhold pending a named repair, and write the next-study agenda as directions, never findings. |
Revisit-only labs (no new lesson; the week revisits earlier chapters): nb11 — Poster production and peer review; nb12 — Presentation preparation: the pitch, the hard questions, the invitation; nb13 — The public test: dress rehearsal and the Expo; nb14 — Async module: reflection on what the public test returned.