Polling in Practice: Accuracy of National and State Polls in the 2016 Election

<a href="https://embed.stanford.edu/iframe/?url=https%3A%2F%2Fpurl.stanford.edu%2Fhp050gh2646" class="su-underline">Show Content</a>

Abstract/Contents

Abstract: This study focuses on the impact of mode and sampling procedures on polling accuracy during the period of the 2016 presidential election. Utilizing all national and state level polls released during the final 7 days of the election, polls are classified based on the mode and sampling procedure utilized. Accuracy is then analyzed using the average absolute error of the top two candidates (Donald Trump and Hillary Clinton). The study finds that polling methodologies are utilized at very different rates, with polling in the final weeks of the 2016 election being dominated by Internet / River sampling methodologies (45% of polls). Within the sample of polls, RDD / Live Interviewer polls were the most accurate at both the state and national level. The least accurate polls were those that utilized online / river sampling methods. Significance testing was also conducted to measure the significance of the differences in average absolute errors between RDD / Live interview polls and the least accurate, Internet / River Sampling. Both a stratified Fisher-Pitman permutation test and a test using the untransformed absolute difference from the election outcome find strong support for the hypothesized difference between RDD and river surveys.

Type of resource	text
Date created	March 16, 2018

Author	McLean, Amanda
Primary advisor	Krosnick, Jon A.
Degree granting institution	Stanford University, Department of Communication

Subject	survey methodologies
Subject	pre-election polling
Subject	mode
Subject	sampling
Subject	Stanford University
Subject	Department of Communication
Subject	journalism
Subject	2016 presidential elections United States
Genre	Thesis

Location	https://purl.stanford.edu/hp050gh2646

Use and reproduction: User agrees that, where applicable, content will not be used to identify or to otherwise infringe the privacy or confidentiality rights of individuals. Content distributed via the Stanford Digital Repository may be subject to additional license and use restrictions applied by the depositor.
License: This work is licensed under a Creative Commons Attribution Non Commercial No Derivatives 3.0 Unported license (CC BY-NC-ND).

Preferred Citation: McLean, Amanda. (2018). Polling in Practice: Accuracy of National and State Polls in the 2016 Election. Stanford Digital Repository. Available at: https://purl.stanford.edu/hp050gh2646

Masters Theses in Media Studies, Department of Communication, Stanford University

Loading usage metrics...