United States campaign finance is an area commonly alleged to be rife with fraud – and if so, fraudulent transactions or unusual errors should show up as anomalies in the data. In 2007, Dr. Cho and Dr. Gaines evaluated the extent to which actual first-digit frequencies of campaign finance transactions fit to Benford’s distribution in Breaking the (Benford) Law: Statistical Fraud Detection in Campaign Finance, where a significantly poor fit would indicate a deviation from natural randomness. The previous study found an increasingly poor fit from 1994 to 2004. I reproduced this analysis using the same type of campaign finance data from 2006 to 2020. I found that though visually most of the election cycle data appeared to fit quite well to Benford’s distribution, according to the two of the goodness of fit tests, all significantly deviated from Benford’s Law, a confirmation of patterns Dr. Cho and Dr. Gaines found in more recent data. However, according to the Euclidean distance test, most of the campaign finance data fits almost as well as the original data Benford used to demonstrate applications of Newcomb’s distribution.
Authors
First Name
Last Name
Elisa
Chen
File Count: 1
Leave a comment
Submission Details
Conference URC
Event Interdisciplinary Science and Engineering (ISE)