Episode 16: Takin it to the Spreadsheets, with Professor Michael McDonald

Election polling and voter turnout analysis are far more intricate than most people realize. From the challenges of non-response bias to the nuances of early voting patterns, understanding election forecasts requires a sophisticated approach beyond simple percentages and headlines. 

Listen to the episode on Spotify:

The evolution of voter turnout calculations 

For years, conventional wisdom held that voter turnout in the United States was steadily declining. However, this perception was based on flawed calculations that didn't account for important demographic changes. By developing the concept of the "voting-eligible population" (VEP), researchers have provided a more accurate picture of turnout trends. This metric excludes non-citizens and ineligible felons from the voting-age population, revealing that turnout rates have remained relatively stable over time, with recent elections even showing increases in participation. 

Challenges in polling methodology and interpretation 

Polling has become increasingly challenging in recent years due to factors like low response rates and difficulty predicting who will actually vote. Pollsters use complex weighting schemes to adjust their samples, but these methods can introduce their own biases. The shift towards early voting and mail-in ballots has made capturing the total electorate in pre-election polls harder. Understanding these challenges is crucial for interpreting poll results critically and avoiding overconfidence in their predictive power. 

The impact of demographic shifts on voting patterns 

The United States is becoming increasingly diverse, with the non-Hispanic white share of the electorate steadily decreasing. This demographic shift, educational trends, and generational differences are reshaping political coalitions. For example, college-educated voters have become more likely to support Democrats in recent elections, while non-college-educated white voters have trended Republican. These changes make it crucial for pollsters and analysts to constantly reassess their assumptions about the electorate's composition. 

Early voting analysis and its predictive power 

Analysis of early voting patterns has emerged as a powerful tool for predicting election outcomes. By examining the party registration of early voters and comparing it to previous election cycles, researchers can gain insights into voter enthusiasm and potential shifts in the electorate. This approach can provide a valuable check on traditional polling data, especially in states with significant early voting. 

The need for standardized election data collection 

The decentralized nature of U.S. elections has led to a patchwork of data collection practices across states and localities. This lack of standardization makes conducting comprehensive analyses of voting patterns and election administration difficult. Sometimes, even basic information about local elections takes a lot of work to obtain. Improving and standardizing election data collection could enhance the transparency and efficiency of the democratic process. 

Potential benefits of a national identity card for voter registration 

Experts across the political spectrum have proposed a national identity card that could serve as both identification and voter registration. Such a system could potentially solve many of the current issues with voter registration, including concerns about fraud and the challenges of maintaining accurate voter rolls. While politically controversial, this approach could streamline election administration and increase confidence in the voting process. 

Get the episode on Apple Podcasts:

16: Takin It to the Spreadsheets, with Professor Michael McDonald
Podcast Episode · Cross Tabs · 08/20/2024 · 1h 12m

Resources 

"The Nation's Data at Risk" (American Statistical Association) 

  • A report highlighting threats to national data collection and sharing, emphasizing the lack of standardized data collection and sharing across agencies. 

"Mad Poll Disease" (Washington Monthly)  

  • A series of posts by Mike Podhorzer critiquing overreactions to individual poll results and emphasizing the need for a more measured approach to interpreting polling data. 

Voter Turnout Gap in Surveys  

  • A study co-authored by Michael McDonald examines discrepancies between survey-reported turnout and actual turnout rates, highlighting the challenges in accurately measuring voter participation. 

Performance of American Elections Survey (MIT Election Lab)  

  • A survey by Charles Stewart from MIT was used in litigation involving Dominion and Fox News to assess changes in public confidence in elections. 

Election Lab at UF  

  • This is a website for Michael McDonald's research group, which provides information and data on elections, including early voting analysis and precinct-level data collection. 

FiveThirtyEight Polls  

  • A source of election analysis and polling aggregation, mentioned for its historical approach to discussing polling uncertainty and more recent return-to-point estimates. 

Current Population Survey Voter and Registration Supplement (U.S. Census Bureau) 

  • A government survey providing demographic data on voters and registration, used as a benchmark for understanding voter turnout and demographics. 

American National Election Study  

  • A long-running academic survey on voting behavior and political attitudes, which has experimented with different question formats to improve accuracy in self-reported voting behavior. 

Cooperative Election Survey  

  • Mentioned in the context of a controversial study on non-citizen voting, highlighting the challenges of using survey data to make claims about voter fraud. 

“Weekend Reading”  

  • Mike Podhorzer's newsletter discussing political trends and data analysis, including regional maps of voter behavior that go beyond traditional demographic breakdowns. 

US Election Project (Michael McDonald's Substack newsletter) 

  • Newsletter where McDonald posts analysis on early voting and other election-related topics, offering insights into voter turnout and electoral trends. 

 Our Guest 

Dr. Michael McDonald is a professor of political science at the University of Florida and a renowned expert on American elections. He is best known for producing the United States Elections Project, which provides the country's most authoritative voter turnout statistics.  Dr. McDonald's work on early voting patterns and election administration has been widely cited in academic research and media coverage. His expertise extends to redistricting, voting rights, and the analysis of precinct-level election data. With a career spanning decades, Dr. McDonald has become a go-to source for understanding the complexities of the American electoral system and voter behavior. 

Your Host 

Farrah Bostic is the founder and Head of Research & Strategy at The Difference Engine, a strategic insights consultancy. With over 20 years of experience turning audience insights into effective strategies for B2B and B2C companies, Farrah helps business leaders make big decisions across various industries. Learn more at thedifferenceengine.co and connect with Farrah on LinkedIn.

Get the latest directly in your inbox