A Computational Analysis of Women’s Suffrage in Historical Newspapers

A computational analysis was performed on over 600,000 digitized newspaper pages from California and New York to measure coverage of the women's suffrage movement from 1890 to 1920. The study demonstrates that raw counts of search terms are misleading due to variations in archive size, necessitating normalization to reveal accurate trends. It explores how different modeling choices, such as strict versus broad keyword queries, can alter specific findings like the peak year of coverage, even while general patterns remain robust. The analysis also highlights how dataset composition, including the dominance of a single newspaper or the presence of non-English text, can significantly skew results. Ultimately, the research serves as a case study on the challenges of measurement and interpretation when working with historical digital archives.