The Nonlinear Effects of Political Advertising

Abstract

Political communication scholars who study the persuasive effects of political advertising often assume that the relationship between advertising volume and vote choice is linear. Even though the product advertising literature has long recognized that advertising has both “wear in” and “wear out” effects, few political communication scholars consider this possibility. This study explains why it is both normatively and empirically important to determine the precise relationship between advertising and vote choice. We then use data from two published studies to investigate inductively how persuasion effects vary with different television advertising levels. The findings suggest that political advertisements do indeed have both “wear-in” and “wear-out” effects. We conclude by investigating three ways to model advertising effects in a more parsimonious manner and settle on log transforming the measure of advertising as the preferred method.