Week 7 — Capstone build upcoming¶
August 3 – August 7, 2026 · Chapters 7.1 – 7.5 · Lab 7 · Mentor Session 7 · Project deliverables
Turn a hunch into a falsifiable, pre-registered empirical design with data in hand.
This is the week where the textbook stops being something you read and becomes something you use. By Friday at 5 pm your team has a registered pre-analysis plan, a cleaned analysis dataset that survives reproducibility checks, and a one-page identification memo that names the threats and the responses.
Daily schedule¶
| Block | What |
|---|---|
| 9:00 – 12:00 | Read Ch 7.1 — Idea-Generation Workshop (so-what filter; cast → dataset map; feasibility/novelty rubric). Work nb7.1. |
| 1:00 – 4:00 | Team kickoff — assign roles (PAP author, data lead, code lead, writing lead). |
| 4:00 – 5:00 | Office hours |
| Block | What |
|---|---|
| 9:00 – 12:00 | Read Ch 7.2 — Data Acquisition in Practice (WRDS, EDGAR, FRED, HMDA, 13F, PatentsView). Work nb7.2 (multi-source pull harness). |
| 1:00 – 4:00 | Lab 7 (Part A) — your team's pull script runs end-to-end, with cache + audit log. |
| 4:00 – 5:00 | Office hours |
| Block | What |
|---|---|
| 9:00 – 12:00 | Read Ch 7.3 — Pre-Analysis Plan (Olken 2015 short form). Work nb7.3 (power calc + BH-FDR + PAP emitter). |
| 2:00 – 3:30 | Mentor Session 7 — "Pre-registering before peeking" |
| 4:00 – 5:00 | Office hours |
| Block | What |
|---|---|
| 9:00 – 12:00 | Read Ch 7.4 — Building the Analysis Dataset (PERMNO/GVKEY/CIK crosswalks, survivorship/look-ahead, winsorizing). Work nb7.4. |
| 1:00 – 4:00 | Lab 7 (Part B) — assemble the analysis dataset with validate= merges and the merge_asof look-ahead guard. |
| 4:00 – 5:00 | Office hours |
| Block | What |
|---|---|
| 9:00 – 12:00 | Read Ch 7.5 — The Identification Memo. Write yours. |
| 1:00 – 3:00 | PAP file + identification memo submitted; first-look regressions remain frozen until the PAP is in. |
| 3:00 – 4:00 | Demo + Week-8 preview |
| 5:00 pm | Deliverable cutoff |
What to read¶
- Ch 7.1 — Idea-Generation Workshop
- Ch 7.2 — Data Acquisition in Practice
- Ch 7.3 — The Pre-Analysis Plan
- Ch 7.4 — Building the Analysis Dataset
- Ch 7.5 — The Identification Memo
Notebooks: notebooks/week-07/.
Mentor reading: Mentor 7 — "Pre-registering before peeking".
What to submit (due Fri Aug 7, 5:00 pm ET) — team deliverables¶
- Pull script + data card in
projects/<team>/data/(PS 7.2 style). - Pre-Analysis Plan — registered, time-stamped, in
projects/<team>/pap.md. - Merged analysis dataset — passes
pytest -q projects/<team>/tests/. - Identification memo + threats table in
projects/<team>/identification-memo.md. - Per-student journal entry —
journal/week-07.md. - Pull requests — one per team member, against the team branch.
PAP discipline
Once your PAP is filed Friday at 5 pm, the primary specification is frozen. Anything that deviates in Week 8 must be reported in a deviations log in the final paper. This is not bureaucracy — it is the difference between a finding and a fishing expedition.
Data we touch this week¶
Whatever your team's charter calls for. The most common combinations:
- CRSP + Compustat — corporate-finance and asset-pricing teams.
- HMDA + FFIEC Call Reports — fair-lending and bank-regulation teams.
- USPTO PatentsView — innovation teams.
- EDGAR 10-K/10-Q + Loughran–McDonald — text-as-data teams.
- FRED — every team, for macro controls.
Next week preview¶
Week 8 — Symposium →. Specification curve, robustness battery, poster, defense — and Wednesday, August 12, at the GMU ASSIP Symposium.