From Archives to Artifacts

A Forensic Analysis of Noisy Data

Bibliographic Data

ID13100232
AuthorsDwayne Woods (0000-0002-7033-794X, Purdue University West Lafayette, corresponding author)
Year2025
Volume11
Issue1
Pages2-37
Publication date2025-08-05
Open AccessYes
TypeARTICLE
VenueChinese Political Science Review (JOURNAL)
Journal identifiersISSN: 2365-4252 • E-ISSN: 2365-4244
PublisherSpringer Science+Business Media (PUBLISHER • DE)
DOI10.1007/s41111-025-00309-y
OpenAlexW4412993897
LanguageEN
Citations received1
References cited29

This paper presents a forensic analysis of a dataset used by Boix to argue that imperial legal emancipation contributed to the spread of Jewish national identity by establishing Zionist and Hebrew institutions. We demonstrate that, although the dataset is extensive, it is logically inconsistent, fragmented over time, and geographically incoherent. Despite claims of establishing cause-and-effect, the data’s structure prevents such conclusions. Using only the most organized variables in a simulation, we demonstrate that even sophisticated machine learning models cannot accurately find the pattern without creating false signals. The main point of this paper is simple: messy historical data that is layered, repetitive, and poorly organized cannot produce clear empirical results. This work adds to the growing field of quantitative history by providing both a critique and a practical guide for maintaining data quality in historical social science

Archaeology · Data science · Forensic science · Computational and Text Analysis Methods · Computer Science · Forensic and Genetic Research · Generative Adversarial Networks and Image Synthesis · History

Tools

Open DOIOpen Access