# Synthetic anthropology and cultural evolution review

Reviewed 2026-09-04. This is an independently run AI review of the brief and sources, not external peer review or the opinion of a named scientist. Scope: claims connecting the Romanian meteorological calculation to ritual history. No participant beliefs, ritual effects or Romanian ritual prevalence are measured by the calculator.

## Verified source dossier

### QJE article

[Espín-Sánchez, Gil-Guirado and Ryan, *Praying for Rain*, QJE 141(3), DOI 10.1093/qje/qjag026](https://academic.oup.com/qje/article/141/3/2363/8676046).

- Table II: rain-ritual coding covers 1,208 groups; 39.2% coded positive. Hazard estimates cover 1,278 groups. The prose describes an atlas of 1,290 groups, while several table variables have 1,291 observations. Keep denominators attached to the specific analysis.
- Table III: 1,195 observations; increasing hazard associated with +14 percentage points, standard error 3.8 points, controlling for continent in column 1. The non-increasing comparison mean is 30%.
- Section V.B.3 classifies a fitted hazard by its derivative at the local 99th percentile of spell duration. This does not mean the daily probability rises throughout a dry spell. The authors report that 34% change from decreasing at p95 to increasing at p99.
- Section IV.B.1 codes no clear evidence of rainmaking as zero. This is a documentation measure with possible false negatives.
- Figure V's caption uses precipitation >5 cm for a separate comparison of seasonal historical rainfall reporting. Do not copy that threshold into hazard estimation.

### QJE supplement

[Publisher supplement](https://academic.oup.com/qje/article/141/3/2363/8676046#supplementary-data), opened through the article's Online Appendix link.

- Appendix B.5, PDF page 21, printed page 83: a failure requires daily precipitation exceeding 0.5 cm, hence strictly >5 mm. Monday rain followed by Thursday rain gives interval t=3. A calculator counting completed dry days uses a different displayed index and must define it explicitly.
- Appendix D.1, Table D11, PDF page 32, printed page 94, explicitly reports rounded raw means 44% for increasing and 30% for non-increasing hazard groups. These are the exact source of the 30/44 comparison. Subgroup counts are not printed there. The adjusted D11 coefficient is 0.16; main Table III gives 0.14. Do not silently merge them.
- Appendix D12, PDF page 34, printed page 96, contains placeholder text in the retrieved supplement. Its numerical robustness table was not verifiable.

### Romanian ethnographic documentation

[Repertoriu național de patrimoniu cultural imaterial, vol. II, hosted by INP/cIMeC](https://cimec.ro/wp-content/uploads/2021/09/Repertoriu-Patrimoniu-Cultural-Imaterial-vol-IIA-romana-si-franceza_0.pdf), chapter 1, *Cunoștințe despre om și univers*, *Astronomie – meteorologie*, printed page 24, notes 112–113. Downloaded and text-extracted locally because the web reader rejected the 22.95 MB PDF.

The official scholarly compilation records Caloian and Paparuda as rain-invocation customs. Note 112 describes Caloian after several weeks of drought and attributes its account to G. Dem. Teodorescu through Traian Gherman. Note 113 points to Cuceu 1988, pages 77–88, for regional names and to its corpus of 716 ritual texts, pages 181–243. It describes procession, rhythmic movement, ritual wetting and singing, with examples from Ibănești Pădure, Mureș, and Micșuneștii Mari, Ilfov.

This is institutional documentation of ethnographic testimony, not a contemporary field survey. The 716 are texts, not independent rituals, communities or current practitioners. Caloian and Paparuda must not be equated with present-day Orthodox processions or assumed to share a single uninterrupted history.

### Two dated event examples

| Event | Evidence and permitted use |
|---|---|
| Suceava, 9 August 2015 | [Basilica, published 11 August 2015](https://basilica.ro/procesiune-pentru-ploaie-la-suceava/), describes a procession with the relics of Saint John the New following prolonged drought. Event date and publication metadata verified in the fetched first-party HTML; also preserved in [Doxologia](https://doxologia.ro/procesiune-pentru-ploaie-la-suceava). |
| Parohia Prozanic, Călărași, Sunday before 22 July 2024 | [Basilica, published 22 July 2024](https://basilica.ro/procesiune-si-rugaciune-pentru-ploaie-intr-o-parohie-din-calarasi/), reports the parish's procession and prayers in a nearby stubble field, citing the parish press release. The implied Sunday is 21 July; the article itself writes only Sunday. Prefer publication date if avoiding an inferred event date. |

These are records that events occurred. Neither supplies a denominator of all processions or a matched meteorological test. No claim about subsequent rainfall is needed.

## Material overclaim risks and resolutions

1. **Transferring a global cultural association to Romanian stations.** A station map cannot establish where rituals are common. Label every location as a measurement site and keep event examples outside statistical comparisons.
2. **Treating a longer deadline as evidence of increasing daily probability.** Cumulative probability can increase with the deadline even if daily conditional probability declines. Show the daily and cumulative quantities separately on a common eligible cohort.
3. **Inferring intention or belief from timing.** Historical weather frequencies cannot reveal whether a priest consulted a forecast, sought to persuade people or believed a ritual worked. Such claims require separate evidence. The cultural-selection hypothesis does not by itself demonstrate deliberate deception.
4. **Collapsing distinct practices and periods.** Archival folklore, current parish events and a global ethnographic coding exercise have different units. Retain source dates and source types; use historical attestation language for the repertoire.
5. **Calling the calculator a Romanian replication of the social mechanism.** It estimates an empirical meteorological analogue from selected NOAA stations. Testing the social hypothesis would additionally require systematically collected ritual dates, locations, ascertainment rules and comparison events. Never call station coverage a representative sample of Romanian people.

Acceptance recommendation: publish the meteorological result with the limits above; treat Romanian ritual material as context. Inspect the final public copy for these boundaries before release. This review does not certify the implementation or observed outputs.