Hostname: page-component-7b7567c8c-z4zkx Total loading time: 0 Render date: 2026-11-16T01:04:45.001Z Has data issue: false hasContentIssue false

Political Scandal and Bias in Survey Responses

Published online by Cambridge University Press:  06 October 2014

Nicholas Goedert*
Affiliation:
Lafayette College
Rights & Permissions [Opens in a new window]

Abstract

This article provides evidence for bias in the polling of American political candidates who are accused of personal or financial scandal, wherein the support of the accused candidate is understated. Evidence for this phenomenon is found in the analysis of a dataset of district-level polls of US House elections during the 2002–2012 election cycles. This bias helps to explain several unanticipated outcomes in recent American legislative elections, in which scandal-tarred incumbents unexpectedly were reelected or defeated by surprisingly narrow margins. The article also finds evidence of a smaller bias, previously observed by practitioners, wherein support is overstated for incumbents who are not accused of scandal.

Information

Type
Features
Copyright
Copyright © American Political Science Association 2014 
Figure 0

Table 1Incumbent Party Underperformance: Scandal versus Non-Scandal

Figure 1

Figure 1Kernel Density of Incumbent-Party Underperformance

Figure 2

Table 2Regression Results of Model, Including Incumbency and Scandal

Figure 3

Table 3Regression Results of Model, Including Scandal (Matched Sample with Open Seats Excluded)

Supplementary material: File

Goedert supplementary material

Supplementary appendix

Download Goedert supplementary material(File)
File 227.1 KB