TITLE:
By Chance or Design? On the Locations of Controversial Billboards in Urban Ohio during the 2012 Election
AUTHORS:
R. Christopher Weaver
KEYWORDS:
Politics; Race; Voting; Simulation; Probability Applications
JOURNAL NAME:
Applied Mathematics,
Vol.3 No.12A,
December
31,
2012
ABSTRACT:
The 2012 American
Presidential Election was preceded by widespread efforts to promote stronger
voting requirements in states across the nation. In one case, a private,
anonymous foundation purchased advertising spots on billboards in urban Ohio to
communicate that voter fraud is a felony punishable by fines and imprisonment.
This action drew criticism from civil rights groups, who argued that a majority
of the billboards were located in minority census tracts, and that the
advertisements utilized intimidating language and imagery. This article applies
probability theory to the debate in order to determine the likelihood that the
observed patterns of billboards in two cities—Cleveland and Columbus—could have
occurred by chance. First, simulation is employed to compare the observed
allocations of billboards to white and non-white census tracts to patterns
generated under Complete Spatial Randomness (CSR). Second, simulation draws
from probit regression models, in which the dependent variable is the presence
of a billboard within a census tract, are used to generate distributions of
first differences between1) the
expected value of the dependent variable given that the census tract is
majority white and2) the expected
value of the dependent variable given that the census tract is majority
non-white. The results suggest that the billboard locations in Cleveland were
not significantly different from patterns produced under CSR, and that the
probability of a given census tract containing a billboard is not significantly
different for white and non-white majority tracts. The opposite inferences are
made for Columbus.