Evaluating federated voluntary associations' membership data: An application of Benford's Law

In: Social science quarterly, Band 102, Heft 4, S. 1590-1601

Abstract

AbstractObjectiveLarge, federated voluntary associations of the late 19th and early 20th centuries produced detailed reports of national meetings that included comprehensive data on state‐level membership. Though used in previous scholarship, the veracity of these data has never been evaluated.MethodsWe test 16 associations' state‐level membership data, generally between 1880 and 1920, using Benford's Law, a mathematical principle used to discern whether data are potentially problematic—through error, fraud, or other reasons—and require further investigation.ResultsOur initial analyses reveal that these data deviate statistically from Benford's predictions. Further investigation, however, finds that these differences are not substantively significant; an aggregate measure of all associations shows close conformity to the Newcomb–Benford distribution.ConclusionThe data presented in voluntary associations' annual reports are likely trustworthy. These findings have important implications for future use of these data, other association data, and the role of associations in promoting civil society.