Harvard Data Science Review Podcast Episode: Digesting 2024 Election Polls: How the Media Reports and Decodes the Numbers
Release Date: October 29, 2024

Introduction

In this insightful episode of the Harvard Data Science Review Podcast, hosts Liberty Vittert and Shaoli Meng delve deep into the complexities surrounding the 2024 US Presidential election polls. Joined by esteemed guests Leland Vittert, anchor of On Balance with Leland Vittert, and Colby Hall, founder of Mediaite, the discussion navigates the intricate landscape of poll accuracy, media reporting, demographic influences, and the evolving role of data science in shaping public perception.

Understanding Poll Accuracy and Data Reliability

Overwhelming Abundance of Data
The episode kicks off with Liberty expressing the current inundation of data available for analysis:
"There is so much data now. You know, it's overwhelming. There's so many models, there's so many polls. There's so, I mean, there's just so much information and so much detailed information" [00:01].

Assessing Poll Quality
Leland Vittert emphasizes the importance of distinguishing between good and bad data, highlighting that not all polls are created equal:
"The polls are only as good as the models... the polling averages... give you a better idea, just sort of the direction of things over time rather than just focusing on this snapshot of this one poll at this one moment." [01:22].

Mediaite’s Approach to Polling Data
Colby Hall echoes the sentiment, stressing the value of aggregate data over individual polls to filter out noise:
"Trending is way more interesting data for us than the latest outlier poll. That may end up being meaningless." [02:24].

Media Reporting Techniques and Challenges

Selective Reporting and Bias
The conversation shifts to how media outlets may skew data to fit their narratives:
"It's very easy to find data now that cheerleads one side or cheerleads the other side." [03:04] - Leland Vittert.

Historical Lessons and Ethical Reporting
Reflecting on past elections, Colby highlights instances where non-traditional pollsters influenced public perception, pointing out the necessity for methodological evolution:
"Polling methodology has been rendered not obsolete, but almost outdated." [04:31].

Shaoli Meng raises concerns about potential biases in data collection, leading to fundamental questions about polling paradigms:
"A political scientist wrote about how the polling paradigm needs to be shifted... what if there are something fundamentally biased?" [06:11].

Demographic Influences and Voting Trends

Critical Voter Demographics
Leland identifies key demographics that could influence the 2024 election results, such as young single white women and young Latino males in pivotal states:
"Groups like single white females, young African American males... could sway the results in very unexpected ways." [07:01].

Shy Voters and Realignment
The discussion explores the concept of "shy voters"—voters who may not openly support a candidate but do so in secret—and the ongoing realignment of the electorate since 2016:
"If we underestimate how far that trend has gone, that will skew things pretty significantly." [10:43].

Lessons from Past Elections

Transparency and Viewer Understanding
Leland advocates for greater transparency in reporting, suggesting that explaining real-time numbers builds public trust:
"The more transparency you can give, the better... if somebody decides because so many people have voted, I'm not going to go vote, shame on you." [18:49].

Media’s Role in Election Narratives
Colby discusses the ethical considerations in reporting early votes and the media's responsibility not to influence voter behavior:
"Total number of votes is totally fair game... but there's a question whether outlets should report that... it could have an impact." [20:17].

The Role of AI in Journalism

Challenges with Generative AI
Shaoli introduces the topic of generative AI, questioning its impact on journalism:
"Generative AIs can help... sometimes they're hallucinating. How are you navigating that world?" [29:18].

Navigating Misinformation
Leland emphasizes the importance of verifying the authenticity of information in the age of AI-generated content:
"Is this real? Is this photograph or is this video even real?" [30:33].

AI’s Limitations and Human Value
Colby expresses skepticism about AI replacing human nuance in reporting:
"AI is something that we're... still unique to the human brain. So the plethora of content makes good content even more valuable." [34:08].

Liberty shares a personal anecdote highlighting both the strengths and flaws of AI-generated content, underscoring the irreplaceable value of human insight:
"I had written an op ed... some lines were really good... some were really cheesy and stupid." [34:39].

Ethical Considerations in Reporting Early Votes

Transparency vs. Influence
The hosts and guests debate the ethical implications of reporting early voting data, weighing the benefits of transparency against the potential to influence voter behavior:
"If Harris is up big in early voting... a lot of people would be upset... that's an interesting ethical question." [21:01].

Decision-Making in Media Outlets
Leland and Colby discuss how media organizations handle the dissemination of early voting information, balancing transparency with ethical reporting:
"It's a political decision of what numbers to release... The media's job is to report." [22:35].

Final Thoughts and Conclusions

The Evolving Media Landscape
Colby reflects on the shift towards partisan media and the challenges it poses for unbiased reporting:
"50% of Americans are going to be gobsmacked by the result because they've been tuning into outlets that have led them to believe that the impossible actually happened." [28:04].

Human Element in Data-Driven Reporting
Both guests agree on the enduring value of human perspective in an increasingly data-saturated media environment, emphasizing that thoughtful analysis and unique viewpoints remain irreplaceable:
"Having a perspective that's been thought about and kicked around... that's still unique to the human brain." [34:08].

Closing Remarks
Shaoli Meng concludes by highlighting the importance of a high voter turnout with minimal drama, quoting a Minnesota Secretary of State:
"Ensure there will be high turnout but low drama." [40:24].

Key Takeaways

1. Poll Accuracy: The reliability of polls depends heavily on the underlying models and the aggregation of data over time rather than isolated snapshots.

2. Media Influence: Media outlets often shape narratives by selectively reporting data, which can both inform and mislead the public.

3. Demographic Shifts: Critical voter demographics, such as young single white women and young Latino males, play pivotal roles in determining election outcomes.

4. AI in Journalism: While AI tools can aid in data processing, the human element remains essential for nuanced and credible reporting.

5. Ethical Reporting: Balancing transparency with the potential influence on voter behavior is a complex ethical challenge for media organizations.

Notable Quotes

• Leland Vittert:
  "The polling is going to be wrong. So the groups I would look at, single white females, young African American males are gonna be really important." [07:01]

• Colby Hall:
  "Using the parlance of some of the younger staff, when every story is lit, no stories are lit, meaning if everything is a bombshell and this is the game changer, this is the big thing, then nothing is." [31:55]

• Liberty Vittert:
  "Ensure there will be high turnout but low drama." [40:24]

This episode offers a comprehensive exploration of the multifaceted relationship between data science, media reporting, and electoral outcomes. By dissecting past election challenges and contemplating future trends, the conversation provides valuable insights for anyone looking to understand the evolving dynamics of political polling and media influence in contemporary elections.