Chapter 40 — Exercises

These exercises are weighted toward doing the audit rather than describing it. Roughly two-thirds of them ask you to open your own dossier and change something in it. That is deliberate: a chapter about assembling and auditing a document cannot be practiced by writing about assembly and auditing.

Items marked † are the harder ones. Several will make you uncomfortable, which is a feature. No answers are provided — for most of these the answer is in your document, and for the rest there is no single correct response, only a defensible one.

Work with your actual dossier open.

Group 1 — Assembly (§40.1–§40.2)

A. In one sentence, and without using the words "summary" or "notes," say what your dossier is. Then check your sentence against §40.1: does it capture all three load-bearing properties?

B. Count your entries. If you have fewer than five or more than ten, write one sentence explaining why, and one sentence on whether you will actually maintain what you have.

C. For every entry, mark each of the twelve fields complete / partial / empty. Do not fill anything in yet. Produce the completion map first.

D. Tally your map by field rather than by compound. Which field number is empty most often across your entries? Compare against §40.2's prediction that most readers find 9, 10, and 12 empty.

E. † Read your Field 1 entries and answer the last sub-question for each: is the same name reliably the same molecule? If you answered yes for anything not obtained through a regulated pharmacy channel, defend the answer or change it.

F. Find an entry where Field 3 (Mechanism) is longer than Field 5 (Evidence). Rewrite Field 3 down to one plain sentence plus target. Then note how much of what you deleted was doing persuasive work rather than informative work.

G. Write the Field 4 line §40.2 says does the most work — does the route it is actually used by match the route it was studied by? — for every entry, including the ones where the answer is obviously yes.

H. Fill in every empty Field 9. For any compound without human trial data, the entry may not read "no known side effects." Use the §40.2 formulation or write your own, but it must distinguish no harms from no reporting pathway.

I. Fill in every empty Field 10, keeping the five sub-questions separate: regulatory, sport, how it is sold, access and cost, preparation quality. If you write one sentence covering all five, you have committed the field's failure mode.

J. † Fill in every empty Field 12. Specify the study — population, endpoint, comparator, duration, rough size — and whether one is underway. Do this for your ❌ entries too, not only the hopeful ones.

K. Pick the entry where writing Field 12 was hardest and say why in two sentences. Difficulty here is almost always diagnostic of something specific.

Group 2 — The internal audit (§40.3)

L. Run the population check across every Field 6 row. Count the rows that name a population and an endpoint, and the rows that do not. Report both numbers before fixing anything.

M. Rewrite three unpopulated rows into complete claims. Note how much longer each becomes, and whether the rating you originally assigned still applies once the claim is narrowed.

N. Run the date check. Date every undated row. Where you cannot remember when you made a rating, date it to when you read the relevant chapter and mark it approximate.

O. † Run the same-evidence test. Find two entries with comparable evidence bases and different ratings. Write the pair down, then decide which of §40.3's three situations you are in — justified by design, population, or endpoint; justified by something you failed to record; or unjustifiable — and fix accordingly. If you cannot find a pair on the first pass, look among entries with no human evidence, where non-evidence reasons operate most freely.

P. † Run the Field 12 check and apply the §40.3 test to every entry: could a reasonable person disagree about whether the named readout had arrived? List every Field 12 that fails and sharpen it.

Q. Take one of your ❌ entries and mark which kind it is: evidence absent, or evidence present and negative. If you cannot tell from your own Field 5, that is the finding — write down what you would need to look up.

R. † Identify one entry where your rating is defensible but you suspect you reached it by a method that does not generalize — you were right for the wrong reasons. Explain how you can tell the difference from the inside.

Group 3 — Your own error (§40.4)

S. List every compound where your rating differs from Chapter 37's, with the direction of each difference. Present it as a column, not as a set of individual comparisons.

T. † Look at the column, not the rows. Write the single sentence §40.4 asks for, naming your direction: generous toward what you hoped would work, harsh toward what sounded like marketing, or genuinely scattered. Put it at the bottom of your dossier.

U. For each disagreement in your column, note whether the compound was one you had a prior stake in. Does the pattern hold?

V. Someone tells you that being systematically harsh is "at least the safe direction to be wrong in." Write a two-paragraph reply using rating rule 4 and the specific cost described in §40.4.

W. † Take the compound you are most invested in and argue the opposite rating from the one you assigned, in writing, using only your own Field 5. Note where the argument becomes hard, and whether the hard part is evidential or emotional.

Group 4 — Maintenance and drift (§40.5–§40.6)

X. Date your dossier and save a frozen copy of this version, named with today's date. This is exercise-shaped but it is the real thing: do it, do not simulate it.

Y. Write the maintenance line from §40.5 at the top of your document, in your own words.

Z. List the events that would trigger a review of each of your top three entries. Be specific enough that you would recognize the event if it happened.

AA. † Recall a piece of news you encountered about one of your compounds while reading this book — a study, a headline, a friend's report, a video. Check it against the relevant Field 12. Did it satisfy the named readout? Did your confidence move anyway? Write down the gap between those two answers.

AB. Explain the selection mechanism in §40.6 to someone who has not read the chapter, in under 120 words, without using the word "bias."

AC. Name three things that look like evidence and cannot upgrade a rating, and for each say what it legitimately is evidence of. Then identify one entry you should delete, delete it, and record in your version history that you did and why.

Group 5 — Use and transfer (§40.7–§40.10)

AD. Convert your three most important Field 11 verdicts into questions a clinician could answer. Apply the §40.7 test: is the answer in the room?

AE. Take a claim you have made to another person in the last year about a health-related product. Identify which of the three sentences you actually meant — "I don't know," "nobody knows," or "it was tested and it doesn't work" — and whether the person could have told which.

AF. Choose something from outside this book entirely: a supplement, a diet, a consumer device, a wellness intervention. Run all twelve fields on it, aiming for one page. Mark which fields are structurally empty rather than merely unfilled, and say why that distinction matters.

AG. † Using the same non-peptide claim, write one complete Field 6 row you could defend today — population, endpoint, rating, date, and if ❌, which kind — and one Field 12 entry that passes the §40.3 reasonable-person test.

AH. §40.9 argues that a well-specified open question is a result rather than a failure. Take your least resolved entry, write the open question out in full, then write the vague version of the same question. Say what the precise version lets you do that the vague one does not.

AI. † Find a claim about society rather than about a molecule — something about what a drug is doing to a culture, a market, or a population. Attempt Field 5 on it. Where exactly does the field break, and what would you have to substitute for a trial arm, an endpoint, and a comparator? Keep your answer; Chapters 41–44 will test it.

AJ. Write the sentence that goes at the very top of your dossier: what you would need to learn to change your mind about the entry you care about most. One sentence. It stays there permanently.