“Democracy needs polls”
(Author Torbjörn Sjöström chairman Gallup Nordic)
I originally didn’t plan to attend AAPOR this year and hadn’t submitted a paper. However, I was ultimately asked to come because the core focus of the conference aligned so closely with the issues I addressed in my recent open letter, “Defending Polling Integrity | Open Letter to WAPOR & AAPOR“, Looking back, I am very glad I went.
From a scientific perspective, AAPOR is incredibly valuable, bringing together the finest minds in our field. This year was no exception, though the current geopolitical climate meant there were fewer international guests and presenters than in previous years.
Due to this heavy scientific focus, many presentations can be a bit too narrow and detailed to be of immediate practical use for clients or those outside the academic circle of our industry. However, there are some standout exceptions, which I have highlighted via the links below.
First among them is Jon Krosnick from Stanford. He completely debunks the claim that non-probability panels work, demonstrates that weighting bad data actually makes the results worse, and proves that non-response is not nearly the fatal flaw it is often made out to be. As he powerfully summarized it: “Don’t lie, don’t repeat history’s mistakes” a direct challenge to those who advocate for non-scientific methods without being fully transparent about their limitations.
Gary Langer was equally definitive in his presentation. He set the stage by explaining that non-probability panels inherently fail to meet AAPOR’s fundamental criteria, as their claims of achieving true representativeness are fundamentally false. From there, he shifted focus to the operational pitfalls, highlighting severe errors and high-profile retractions.
Most notable this year was a major blunder by YouGov regarding churchgoers in the Gen Z generation, which resulted in a full retraction and left the industry with far more questions than answers about how deep the issue truly runs. Langer detailed how easily these panels can be compromised, pointing to data manipulation driven by both fraudulent human respondents and automated bots. While his approach was more field-oriented and practical than Jon’s, his ultimate takeaway was perfectly aligned: if you want trustworthy, credible data, you must stay away from non-scientific, non-probability polls.
AAPOR 2026: Gary Langer, Opt-In Online Panels: Questions to Ask
About AAPOR
The theme for the 81st Annual AAPOR Conference in Los Angeles is:
“An LA Love Story of Data, Innovation, and the Quest for Truth”
At its core, the 2026 theme centers on reaffirming the critical role of survey research in society and rebuilding public trust in data. Recognizing an era of shifting public perceptions and rapidly evolving methodologies, the conference focuses heavily on bridging the gap between scientific polling and the broader public, ensuring that representative data remains a trusted pillar for informed decision-making.
To ground this “Quest for Truth,” the conference structure is built around core pillars addressing the field’s most immediate methodological shifts, including:
- The Sampling Debate: Navigating the tensions and trade-offs between the traditional “gold standard” of probability sampling and the cost/speed efficiencies of nonprobability methods.
- Data Science & Survey Integration: Exploring how machine learning, automation, and predictive modeling fit into standard total survey error frameworks to detect bias and improve data quality.
- The Rise of AI: Assessing how Large Language Models (LLMs) can complement traditional qualitative research (like text classification and sentiment analysis) without losing vital human oversight and contextual nuance.
- Representation & Dissemination: Making data more meaningful and localized to better resonate with the actual populations being represented, particularly as overall response rates decline.
About AAPOR: The American Association for Public Opinion Research (AAPOR), founded in 1947, is the leading professional organization dedicated to the methodology and ethics of public opinion and survey research in the United States and globally. www.aapor.org
Its membership spans academia, media, government, the non-profit sector, and private industry, united by a core conviction: that scientifically valid public opinion research is essential to a healthy democracy.
