The reading sequence reorganizes the original material without dropping a section-level knowledge chain. Page references below use the supplied chapter PDFs, whose pagination differs from editions mentioned in the 2026 syllabus.
1. Primary sources and page conventions
Stock and Watson, supplied Global Edition chapter extracts: Chapter 2, printed pp. 60–110 (51 PDF pages); Chapter 3, printed pp. 111–154 (44 PDF pages). Thus Chapter 2 printed page = PDF page + 59; Chapter 3 printed page = PDF page + 110.
Chapter 2 PDF（原课程文件，本站未提供） · Chapter 3 PDF（原课程文件，本站未提供） · Econ 141 syllabus（原课程文件，本站未提供）
The syllabus's first topic is probability/statistics review and emphasizes mathematical derivations. The chapter files, rather than a different edition's page numbers, govern this guide.
2. Supporting materials actually consulted
- Schellenberg Section 1（原课程文件，本站未提供）, pp. 1–2: probability functions. Section 2（原课程文件，本站未提供）, pp. 4–8: covariance ratios, LIE, LLN/CLT, sample versus population, and Monte Carlo. These are historical supplementary sections, not the current syllabus.
- Aden Chen econometrics notes（原课程文件，本站未提供）, pp. 7, 12–14, 16, 23: conditional prediction, continuous mapping, Slutsky, moments, and OLS asymptotics.
- Princeton ECO 539B bootstrap lecture（原课程文件，本站未提供）, pp. 3–5, 7, 12–13: empirical resampling, studentized intervals, and failure examples.
- Graham Econ 240a syllabus（原课程文件，本站未提供）, p. 5: graduate-course roadmap. It is not used as a proof source.
- Cattaneo, Jansson, and Nagasawa (2020), Bootstrap-Based Inference for Cube Root Asymptotics, local PDF pp. 1–2 (printed 2203–2204): abstract/introduction read; full theorem proofs not reviewed.
3. What was consolidated
Repeated descriptions of the same definition, duplicate numerical drills, and repeated normal-curve explanations were consolidated. All 13 main sections (2.1–2.6 and 3.1–3.7), all four appendices, and all three empirical-exercise pathways have a destination. Historical application tables are discussed through their inferential purpose rather than reproduced in full. Every numbered textbook exercise is not duplicated; its distinct conceptual chains are represented by the main text and curated exercises.
4. Coverage map
| Original location | Knowledge chain | Destination | Status |
|---|---|---|---|
| SW §2.1 · pp. 61–65 | Outcomes → events → random variables → PMF/CDF/PDF → event probabilities | 01 | covered |
| SW §2.2 · pp. 65–71 | Expectation → Bernoulli moments → variance/SD → affine changes → skewness/kurtosis/raw moments | 02, 14 | covered |
| SW §2.3 · pp. 72–81 | Joint → marginal → conditional → conditional moments → LIE → zero conditional mean | 03 | covered |
| SW §2.3 · pp. 77–81 | Independence → mean independence → covariance/correlation → weighted-sum algebra | 04 | covered |
| SW Table 2.4 · pp. 79–80 | Conditional earnings distributions → means/medians/spread → subgroup definitions | 02, 10 | covered |
| SW §2.4 · pp. 82–88 | Normal standardization → multivariate normal properties → chi-squared → t → F and limits | 05 | covered |
| SW stock-return box · pp. 85–86 | Extreme observations → normal-model inadequacy → tails/variable volatility | 02, 14 | covered |
| SW §2.5 · pp. 89–93 | Sampling design → i.i.d. → sampling distribution → exact mean moments → diversification | 04, 06 | covered |
| SW §2.6 · pp. 93–99 | Exact versus asymptotic → LLN/consistency → standardized CLT → approximation quality | 06 | covered |
| SW Exercises 2.19–2.27 · pp. 106–109 | Multiple conditioning → moment expansions → portfolio → nonlinear dependence → summation → correlated averages → best prediction | 02, 03, 04, 05, 14 | covered; recurring calculations consolidated |
| SW Appendix 2.1 · pp. 109–110 | Variance/covariance identities → expansion proofs → correlation inequality | 02, 04 | covered |
| SW E2.1 · p. 109 | Joint earnings table → conditional means → LIE → covariance/correlation | 15 | workflow and synthetic worked example; original dataset unavailable |
| SW §3.1 · pp. 112–117 | Estimator versus estimate → bias/consistency/efficiency → BLUE → least squares → selection bias | 07 | covered |
| SW §3.2 · pp. 117–126 | Null/alternative → known/unknown variance → sample variance/SE → t statistic → p-value → size/power → sidedness | 08, 09 | covered |
| SW §3.3 · pp. 126–128 | Test inversion → confidence set/interval → coverage → confidence levels | 09 | covered |
| SW §3.4 · pp. 128–130 | Independent means → variance of difference → SE → tests/intervals | 10 | covered |
| SW §3.5 and applications · pp. 130–133,136 | Randomized experiment → conditional-mean causal effect → treatment estimation → natural experiments; earnings gap/time comparison | 10 | covered |
| SW §3.6 · pp. 133–137 | Normal-sample t → pooled variance → equality-of-variance restriction → large-sample practice | 11 | covered with source correction |
| SW §3.7 · pp. 137–141 | Scatterplots → sample covariance/correlation → consistency → unit invariance → nonlinear relationships | 12 | covered |
| SW Exercises 3.2–3.21 · pp. 143–149 | Proportions → precision planning → power → paired comparisons → gap changes → variance unbiasedness → transformations → covariance consistency → pooled equality | 08, 09, 10, 11, 12, 14 | covered; repeated numerical drills consolidated |
| SW E3.1 · pp. 149–150 | CPS means → real earnings → education/year comparisons → gap changes | 15 | complete analysis recipe; raw dataset unavailable |
| SW E3.2 · pp. 151–152 | Endowment experiment → 50% benchmark → Bernoulli test → experience-group comparison → List (2003) | 13, 14, 15 | benchmark derived; raw dataset unavailable |
| SW Appendix 3.1 · p. 152 | CPS source/design → full-time definition → population and weighting caveats | 15 | covered |
| SW Appendix 3.2 · p. 153 | Least-squares mean: calculus and algebraic proofs | 07 | covered |
| SW Appendix 3.3 · p. 154 | Second sample moment minus squared mean → LLN → consistent variance | 08 | covered |
5. Corrections and clarifications preserved in the notes
- The extended-reading README's “no formal citations” statement is incorrect when applied to the complete Chapter 3: E3.2 explicitly cites List (2003). Chapter 2's stock-return box also mentions Taleb's book. The original README was left unchanged; this guide records the correction.
- Equation (3.23), printed p. 135, uses the first group's mean inside the second group's squared deviations in the supplied PDF. The pooled variance must use each group's own mean. This was checked visually against the PDF.
- Equal group sizes equate pooled and unpooled SE formulas but do not, by themselves, establish an exact Student t distribution.
- The exact n−1 sample-variance convention for a Bernoulli sample differs slightly from the usual n-divisor plug-in proportion SE. Both are stated explicitly.
- SD(a+bY)=|b|SD(Y); the absolute value is needed for negative b. Symmetry implies zero skewness when the moment exists, but the converse is not universal.
- Finite-variance LLN and finite-fourth-moment sample-variance arguments are sufficient conditions used in the chapter. They are not stated as the weakest possible assumptions.
- Simple random sampling from a finite population without replacement is not literally i.i.d.; the guide distinguishes the textbook's working model from this design issue.
- Convergence in probability can be to a random variable in general. The LLN's constant limit is the case used here, not a universal definition.
6. Evidence and limits
The original exercise microdatasets were not found in the inspected local files. Empirical workflows and explicitly synthetic examples are supplied; no new CPS or sports-card estimates are claimed. Efron is an attributed reading lead based on the Princeton notes, not a full-paper review. List's publication metadata was checked against its publisher; the experiment description comes from SW. The Cattaneo–Jansson–Nagasawa introduction was read locally. Advanced extensions are marked and are not presented as original chapter requirements.
Source hashes, build checks, numerical checks, and editorial decisions are stored separately in the course Agent Workspace. The reader works offline: typography, navigation, search, and simulation use local assets.