How to take this check
- Do it in one sitting.
- For the multiple-choice and short-answer sections, close every AI tool and tab. This checks your internalized model, not the model’s.
- For each multiple-choice item, pick an answer before you reveal the explanation. Guessing and then reading the answer is not the same as knowing.
- For short-answer items, write your response on paper or in a text file before you reveal the model answer. Compare honestly.
- For the applied section, open your research-brief-log-v1.md, your source-list.md, the open-before-cite checklist, and your synthesis plan. Q9 and Q10 are open-workstation on purpose.
- If you miss a question, the feedback names the lesson(s) to revisit.
Multiple choice CORE
Six questions. Concept recall and diagnosis. One point each.
Show explanation
Answer: C. The four moves are scope, retrieve, triangulate, synthesize, with the director’s check laid over the top. Review Lesson 4.1 Block 2 if missed.
Show explanation
Answer: C. Five articles quoting one release are not five sources — they are one source in five costumes. Triangulation counts independent sources, and a careful brief would cite the press release directly. Review Lesson 4.3 Block 2 if missed.
Show explanation
Answer: C. Searching finds invented sources; opening catches broken URLs; searching the quote in the paper catches fake quotes. A real source whose argument is misstated survives all three cheap checks — only a careful read catches it. Review Lesson 4.3 Block 3 if missed.
Show explanation
Answer: C. The comparison is bounded, the answer depends on sources outside the model, and a defendable answer is worth producing. (A) is a primary document short enough to read yourself. (B) is a summary task, not research. (D) is high-stakes medical and should not be answered by an agent alone. Review Lesson 4.1 Block 4 if missed.
Show explanation
Answer: B. The false voice is the synthesis equivalent of averaging away contested findings — the brief sounds authoritative and has quietly lied about the state of the evidence. The fix is to name the disagreement where the claim appears. Review Lesson 4.4 Block 2 if missed.
Show explanation
Answer: D. The rule is uniform. Any exception (“only the long ones,” “only the academic ones”) is exactly where fabrication slips through — short sources and news URLs are among the most commonly hallucinated kinds. Review Module 4 README and Lesson 4.3 Block 4 if missed.
Short answer CORE
Two questions, three to four sentences each. Up to 1.5 points each. Write your answer before revealing the model answer.
Q7. A classmate says: “Opening every source before citing it is too slow. I spot-check a few and trust the agent on the rest.” Name three specific problems with this stance, referencing Module 4’s failure modes.
Model answer.
- Students routinely pick the sample in a way that misses the fabricated entry. Humans spot-check the citations that look sketchy and trust the ones that look clean — which is exactly how a well-formed fabrication lands unexamined.
- Fabrication patterns 2 and 3 survive casual checks. A real-source-fake-quote or real-source-misrepresented-finding citation has a correct-looking top line; only opening and reading the specific passage catches the problem. A “spot check” that just confirms the paper exists does not touch either failure mode.
- The habit only works if it is uniform. Exceptions metastasize, because “I’ll skip the easy ones” becomes “I’ll skip the ones I’m pretty sure about,” becomes the status quo — and the one fabricated citation you ship is the one everyone remembers the brief by.
Time-cost framing (bonus, not required): the audit takes a short, bounded amount of time and saves the credibility of the whole brief. “Too slow” is the wrong denominator.
Scoring: Full credit (1.5) for naming at least three of {sampling misses fabricated entries, fabrication patterns 2/3 survive casual checks, habit only works if uniform, cost of one shipped fabrication is the brief’s credibility}. Partial credit (0.75) for two reasons without the time-cost framing. No credit for “you just should” without specific failure modes named. Review Lesson 4.3 Block 3 and Block 4 if missed.
Q8. A student drafted a synthesis brief. The uncertainty in the topic appears only as a hedged paragraph at the end: “Of course, there are limitations to this analysis, including X, Y, Z.” The body argues confidently throughout. What is the specific failure mode, why is it bad, and how does Module 4 recommend fixing it?
Model answer.
- Failure mode. The hedged-postscript pattern — a variant of false voice. Uncertainty is quarantined at the end of the document instead of held near the claim that depends on it.
- Why it is bad. Readers remember the confident body and the postscript vanishes; the brief misleads even an honest reader, because the body says one thing and the postscript technically covers a different one. The brief has outsourced the truth-telling to a paragraph the reader will skim or forget.
- The Module 4 fix. Move each piece of uncertainty to the paragraph where the relevant claim appears, using the confidence-tier vocabulary (low / medium / high) or explicit disagreement-naming (e.g., “Thompson (2024) reports X; Garcia & Lee (2024) report no effect”). The confident sentence and the uncertainty about it live together.
Scoring: Full credit (1.5) for naming the hedged-postscript pattern, the misleading-confident-body problem, AND the hold-uncertainty-near-the-claim fix. Partial credit (0.75) for naming the failure mode without the fix. No credit for “it’s fine” or “just add more hedges everywhere.” Review Lesson 4.4 Block 4 if missed.
Applied CORE
Two questions, half a page each. Open-workstation — keep your frozen log, source list, audit checklist, and synthesis plan in front of you. Up to 3 points each.
Below is a short passage from a synthesis brief, with three citations. You have a short, time-pressured window to decide: ship as-is, revise, or return to research. If revise, name the specific revisions. If return to research, explain what is missing.
Passage (excerpt from a brief on the four-day school week, for a homeschool co-op board):
Research consistently finds that the four-day school week produces academic outcomes equal to or better than the five-day calendar [1]. Districts that have adopted it report improved teacher retention and student attendance [2]. The RAND Corporation has concluded that the model is a cost-effective modernization of the school calendar for rural districts [3].
Citations:
[1] Thompson, J. (2024). “The Four-Day School Week: A Meta-Analysis.” Journal of Educational Policy, 44(2), 115–138.
[2] Morris, K. (2023). Modern Calendars, Modern Learners. Harvard Educational Review.
[3] RAND Corporation (2023). Rural School Calendar Models. rand.org/pubs/research_reports/RR-A1875.html.
Show model answer
Model answer. Decision: return to research. The passage fails multiple Module 4 tests.
- Claim 1 is false-voice. “Research consistently finds... equal to or better” smooths what every real meta-analysis would report as heterogeneity. The word “consistently” is the tell. Even if Thompson (2024) exists and is real, the paraphrase overclaims what a meta-analysis would conclude.
- Citation [2] is structurally suspicious. It does not identify what kind of source Morris (2023) is — book? journal article? — and the publication name (Harvard Educational Review) mixes book and journal conventions in a way that should be verified. The verb “report” covers an unacknowledged triangulation gap (whose report?).
- Citation [3] needs the URL opened. The paraphrase “cost-effective modernization” is the kind of soft summary language that often survives despite the actual RAND document using different language. Open the URL, read the abstract and conclusion, and either rewrite with properly-attributed claims or drop the claim.
The correct move is to run the five-step open-before-cite audit on the passage, open each URL, and rewrite with properly-attributed claims. Drop any claim that cannot be supported.
Scoring: Full credit (3) for flagging at least two of the three failures (false-voice on claim 1, structural problems with citation [2], unopened-URL problem on [3]) and naming return to research or a revision specific enough to fix them. Partial credit (1.5) for flagging one failure without a clear path forward. No credit for “ship — the citations look correct” — the look-correct signal is precisely the failure mode being tested. Also credit students who revise instead of return-to-research IF their revision list is specific enough to catch the three failure patterns named above. Remediation: re-run the fabricated-citation drill, then re-do Entry 2 of your capstone log with a fresh, deliberately-skeptical audit.
Pick a real research question you would actually want answered — one you did not use for your Module 4 capstone, so this is fresh work. In the space below, produce a scoping brief with the five required sections:
- Question (ends with a question mark; names specifics — who, what, where, when).
- Reader and Decision (a specific person and the decision they will act on).
- Answer Shape (one paragraph sketching what the finished output will look like).
- Out of Scope (at least two items that are genuinely adjacent to the question).
- Constraints (time budget and source-type constraints, concrete).
Target 250–400 words for the scoping brief. Then, in one additional paragraph, predict: which of the three research-output shapes — scoping brief, fact-check memo, synthesis brief — fits this question, and why? And: what single source would you open first, and why that one?
Show model answer
Scoring: Full credit (3) requires:
- All five scoping-brief sections present and coherent.
- The question ends with a question mark and names specifics (who, what, where, when).
- The reader is named by person, and the decision they will act on is explicit.
- The out-of-scope list has at least two genuinely adjacent items — not obviously-unrelated filler.
- The output-shape paragraph correctly matches the question to one of the three shapes with a defensible reason (yes/no claim → fact-check memo; bounded-but-multi-claim question → synthesis brief; still-defining-the-work → scoping brief).
- The first-source prediction names a specific source with a specific reason (not “I would Google it”).
Partial credit (1.5) for missing one of the five sections or a vague output-shape justification. No credit for a scoping brief generic enough to apply to any topic — the question explicitly requires specificity, because the point of scoping is that it is specific. Remediation: redo Lesson 4.2 Activity with a new topic and compare your second scoping brief to your first.
Parent / instructor scoring summary
Total: 15 points across 10 questions.
- Multiple choice (Q1–Q6): 1 point each — 6 points.
- Short answer (Q7–Q8): up to 1.5 points each — 3 points.
- Applied (Q9–Q10): up to 3 points each — 6 points.
Passing bar: 11.5 of 15 or better, with at least one applied question at full credit. A miss on Q9 sends the student back to Lesson 4.3 and the fabricated-citation drill; a miss on Q10 sends them back to Lesson 4.2 before Module 5.
Weighting suggestions for parents issuing credit:
- Multiple choice (Q1–6): 40% of Module 4 score.
- Short answer (Q7–8): 20%.
- Applied (Q9–10): 40%. Q9 and Q10 are the load-bearing items — they demonstrate applied judgment on a real passage and real planning, not recall.
Evidence to file in the student’s credit portfolio for Module 4:
- This completed check (all ten answers written out).
- The frozen /capstone/research-brief-log-v1.md — three entries, module summary, open-before-cite rule at the bottom.
- The source-list.md with every cited source marked opened-and-confirmed.
- The synthesis-plan.md — the six-part plan the student built the synthesis brief on.
- The fabricated-citation drill results in /resources/module-04/fabricated-citation-drill-results.md.
- A short (2–3 sentence) instructor note on Q9 and Q10 — which failure mode the student identified in Q9, and which output shape they matched the question to in Q10.
Transcript language. If a parent is assembling a transcript, the transcript line for Module 4 can accurately say: “Directed AI research agents to produce three structured research outputs — a scoping brief, a fact-check memo, and a synthesis brief — on a real topic, with every cited source opened and confirmed before shipping.”
Remediation if missed:
- Q1: Re-read Lesson 4.1 Block 2 and review the Anatomy of a research task activity.
- Q2: Re-read Lesson 4.3 Block 2 (triangulation) and run the fabricated-citation drill a second time.
- Q3: Re-read Lesson 4.3 Block 3 (four fabrication patterns). Print the four patterns and keep them taped near your workstation for the rest of the course.
- Q4: Re-read Lesson 4.1 Block 4 (when a research agent is the right tool) and re-run the anatomy-of-a-research-task activity.
- Q5: Re-read Lesson 4.4 Block 2 (three synthesis failure modes). Re-read your own Module 4 synthesis brief with the three failure modes in mind and mark any sentence that fits.
- Q6: Re-read the Module 4 README’s safety-norm section and Lesson 4.3 Block 4 (the open-before-cite checklist).
- Q7: Re-read Lesson 4.3 Block 3 and Block 4. The habit this question tests is the single load-bearing habit of the module.
- Q8: Re-read Lesson 4.4 Block 4 (holding uncertainty in the text). Rewrite one paragraph of your own synthesis brief to pull the uncertainty out of the postscript and into the body.
- Q9: The source-integrity audit is the load-bearing skill of Module 4. Re-do the fabricated-citation drill, then re-do Entry 2 of your capstone log with a fresh, deliberately-skeptical audit.
- Q10: Scoping is the skill Module 4 builds toward. If this is weak, redo Lesson 4.2 Activity with a new topic and compare your second scoping brief to your first.
If the student passes at 11.5 / 15 or above with at least one applied question at full credit, Module 4 is complete and Module 5 can begin. Below that bar, target remediation to the specific lesson(s) listed above before moving on.
Next up
Module 5 — Email & calendar agents.
The directing muscle transfers again. Module 5 picks up the same four-move loop — this time read, categorize, draft, review — on work that sees your inbox and your calendar: triage, drafting, meeting prep, and the safe patterns for letting an agent touch sensitive data. The operational norm here is two lines long: drafts, not sends and least access for the task.
Module 5 opens when the Module 4 portfolio is complete — this check, the frozen research-brief log, the source list, and the synthesis plan.