{
  "citation": "Goodman-Bacon, A. (2021). Difference-in-Differences with Variation in Treatment Timing. Journal of Econometrics 225(2): 254-277. Cell read from the NBER Working Paper 25018 (revised September 2018).",
  "deviation_evidence": null,
  "doi": "10.3386/w25018",
  "excerpt": "DD Coefficient = -3.08 (s.e. = 1.27)",
  "independent": null,
  "item_set": null,
  "kind": "paper",
  "known_deviation": null,
  "licence": "NBER working paper, copyright the author; cited, not reproduced.",
  "locator": "NBER Working Paper 25018, Figure 5 (event-study and difference-in-differences estimates of the effect of no-fault divorce on female suicide: replication of Stevenson and Wolfers 2006), the printed DD coefficient",
  "metric": null,
  "n_items": null,
  "page": "35",
  "read_copy_sha256": "4f877146cba18f2d19781e2e5ea0be9275d698ef1295f166f7d1778bfc1e17fc",
  "retrieved_at": "2026-10-15",
  "se": 1.27,
  "se_as_published": "1.27",
  "source_id": "goodman_bacon_2021.fig5_dd.paper",
  "tolerance_reason": null,
  "units": "suicides per million women",
  "url": "https://www.nber.org/system/files/working_papers/w25018/w25018.pdf",
  "value": -3.08,
  "value_as_published": "-3.08",
  "verified_by": "primary_source",
  "verified_note": "Read on 2026-10-15 from the NBER Working Paper 25018 PDF (48 pages, created 2018-09-06) with pdftotext -layout and from the page image rendered at 80 dpi: Figure 5 is on PDF page 37 (printed page 35). Its notes say the specification has no other controls, does not weight by population, and that standard errors are robust to heteroskedasticity. The text of section 4 (PDF page 22) states the same coefficient with 's.e. = 1.13'; the figure, whose note states the SE convention, is the cell used, and the working paper's Table 2 column (1) also prints [1.27]. The text's 1.13 matches no standard convention on this file (classical 1.11, HC2 1.27, HC3 1.30, clustered by state 2.49; B-join review). Which SE the Journal of Econometrics version prints is unverified. The published Journal of Econometrics version could not be read without a subscription. Data facts checked before any estimate: bacon_example.dta has 1,617 state-year rows (49 states, 1964-1996) with no missing outcome or reform indicator."
}