Director’s Cut Sessions

At a moment when survey research is being tested, transformed, and reimagined, the Director’s Cut Sessions at the AAPOR 81st Annual Conference take on the biggest questions facing our field. From AI’s disruptive promise and qualitative innovation to representation, dissemination, and the urgent challenge of rebuilding trust, these sessions bring together diverse leaders for candid, critical, and deeply practical conversations. Expect rigor, honest exchange, and space to wrestle with what comes next—for our methods, our data, and the public we serve.

See what each of the chairs have to say about what’s in store for their Director’s Cut Session.

Director’s Cut #1: Survey Methods and Data Science
Wednesday, May 13 | 9:00 am –12:15 pm

“In a rapidly evolving technological and socio-political landscape, our field is facing insecurity, but can also provide guidance. What excites us about the Director’s Cut Session on Survey Methods & Data Science is that it brings together a broad spectrum of perspectives and experiences with AI in our field and dedicates lots of space to honest, critical, and constructive exchange among them. We are also really excited because we are flipping the script in this session, giving the audience a chance to ‘be the panelists’ to speak of their own experiences in adapting to these new technologies. We hope that all who attend will leave the session with more insights and ideas about how survey researchers can navigate this new reality.”

Director’s Cut #1 Chairs

Trent Buskirk, Old Dominion University
Leah Von Der Heyde, GESIS – Leibniz Institute for the Social Sciences

Director’s Cut #1 Session Papers

Can I Prompt You for a Proper Estimate Please? Evaluating How Large Language Models Analyze Complex Survey Data
Trent D Buskirk, Old Dominion University

Why Survey Researchers Are Well Prepared for Generative AI: A Hot (Deck) Take
Josh Pasek, University of Michigan

Beyond Algorithmic Accuracy: Understanding Data Quality in Interactive Occupational Coding with the Coarse Framework
Olga Kononykhina, Ludwig-Maximilians Universität München

A Generative AI Approach for Integrating Synthetic Respondents with Probability-Based Human Panels for Social Science Applications
Leah Christian, NORC at the University of Chicago

Making Humans Count: How Using LLMs in Survey Research Affects Total Error
Dr. Leah von der Heyde, GESIS – Leibniz Institute for the Social Sciences

The Future of AI in Survey Research
Dr. Leah von der Heyde, GESIS – Leibniz Institute for the Social Sciences
Trent D Buskirk, Old Dominion University

Director’s Cut #2: Nonprobability and Probability Sampling
Thursday, May 14 | 9:30 am – 12:45 pm

“Survey research faces a moment of reckoning. High-profile polling misses, disagreements among surveys that seem like they should align, and growing public skepticism have raised persistent questions about which surveys people should trust. At the same time, demand for survey data has not waned. Businesses increasingly rely on customer surveys, and academic disciplines continue to turn to online samples for fast, affordable data collection. Meanwhile, artificial intelligence adds both promise and disruption—fueling speculation about whether AI could one day replace surveys altogether, while also complicating data quality through increasingly sophisticated bots and click-farm activity.”

Director’s Cut #2 Chairs

Paul Scanlon, National Center for Health Statistics
Jon Krosnick, Stanford University

Director’s Cut #2 Session Papers

Improved Mass Imputation in Probability Samples Via Adjustment of Imputation Models Based on Non-Probability Samples for Selection Bias
Brady T. West, University of Michigan

The Past, Present, and Future of Probability and Non-Probability Samples for Surveys
Jon Krosnick, Stanford University

Building and Maintaining an Online Panel
Greg Bovitz, Forthright

Recruiting, Maintaining, Effectively Utilizing Probability Panels for Survey Research: Challenges and Opportunities
Cameron McPhee, SSRS

Evaluating Commercial Synthetic Data for Market Research: Accuracy, Cost, and Speed vs. Traditional Methods
Curtiss Cobb, Meta

Opt-In Online Panels: Questions to Ask
Gary Langer, Langer Research Associates

Integrating Probability and Non-Probability Survey Panels into the Development and Evaluation of Surveys and Questionnaires
Paul Scanlon, National Center for Health Statistics

Director’s Cut #3: Large Language Models and Qualitative Methods
Thursday, May 14 | 2:15 – 5:30 pm

“AI introduces a fundamental shift in how we conduct and analyze qualitative research. This session explores the intersection of traditional qualitative rigor and the transformative scale of AI. We are moving beyond the hype to examine how LLMs can actually strengthen the depth, diversity, and integrity of qualitative insights.”

Director’s Cut #3 Chairs

Heather Ridolfo, U.S. Energy Information Administration
Emily Geisen, Associate Chair

Director’s Cut #3 Session Papers

The Power of Amplification: How AI and LLMs Can Elevate Qualitative Research
Emily Geisen, Qualtrics International Inc.

Bridging Human and LLM Text Annotations for Statistically Valid Inferences
Kristina Gligoric, Johns Hopkins University

Augmenting the Analyst: An Evaluation of an AI-Assisted Methodology for Qualitative Research Analysis
Andrew Stavisky, U.S. Government Accountability Office

Can AI Conduct Expert Review of Survey Questions? Lessons from Using Large Language Models in Questionnaire Evaluation
Zoe Slowinski, NORC at the University of Chicago

Integrating AI-Moderated In-Depth Interviews with Survey Data: Evidence from Five Countries
Charles Lau, Gallup

Evaluating AI-Assisted Qualitative Analysis in Qualitative Research
Mary Meyn, Fluent Research

Determining Minimum Sample Size to Support Thematic Analysis: A Monte Carlo Approach
Carol Sue Haney, Qualtrics International

Quality & Integrity of Qualitative Research: Critical Thinking in the Age of LLMs
Margaret Roller, Roller Research

Director’s Cut #4: Representation and Dissemination: Leveraging Small Domain Estimation and Cross-Cultural Research
Friday, May 15 | 8:30 – 11:45 am

“Across the survey lifecycle, representation is both a technical challenge and a responsibility. This Director’s Cut session brings together researchers who are redefining how we generate and share statistics that meaningfully reflect the people and communities we study. From small domain estimation and cross‑cultural methods to mixed‑methods innovation and thoughtful approaches to respondent protections, our speakers will explore what it takes to produce data that is granular, trustworthy, and truly grounded in lived experience. As sample sizes shrink, response rates decline, and expectations for relevance grow, this session invites us to think critically about how we design, estimate, and disseminate information that resonates with the populations it aims to represent.”

Director’s Cut #4 Chairs

Morgan Earp, Conference Chair
Mandy Sha, Chief Research Scientist

Director’s Cut #4 Session Papers

Representation of Rural America: Perspectives on Survey Design, Identity, and Estimation
Kristen Olson, University of Nebraska-Lincoln

Representing Diverse Communities Throughout the Survey Life Cycle: Understanding California Through the California Health Interview Survey
Todd Hughes, UCLA Center for Health Policy Research

Leveraging Mixed Methods Approaches and Community-Based Recruitment Strategies to Improve Estimations of Public Opinion Among Historically Marginalized Populations
Shakari Byerly, PhD, EVITARUS

Balancing Respondent Protections and Data Needs
Michael Schober, The New School

Small Domain Estimation: A New Tool to Provide More Granular and Timely Estimates by Subgroup
Lauren Rossen, National Center for Health Statistics

Methodological Approaches for Representing Small Populations in National Public Opinion Surveys
Mark Hugo Lopez, Pew Research Center

Director’s Cut #5: Polling, Official Statistics & Public Trust: Reinforcing the Public Value and Accessibility of Official Data
Friday, May 15 | 2:15 – 5:30 pm

“We’re excited about this session because it goes straight at one of the biggest issues facing our field right now: trust. Polling and official statistics are part of the country’s decision-making infrastructure, but that only works if people believe the numbers, understand what they mean, and can actually use them. Bringing together federal, state, nonprofit, and private perspectives lets us look at the full ecosystem and get specific about what strengthens credibility and what gets in the way.”

Director’s Cut #5 Chairs

Jenny Hunter Childs, U.S. Census Bureau
Stephanie Coffey, U.S. Census Bureau

Director’s Cut #5 Session Papers

We Have to Make the Case — A Discussion of American Public Opinion about the U.S. Federal Statistical System
Jordon Peugh, SSRS

How Private Polling Assists, Complements, and Builds Up the Credibility of Governmental Health Surveys
Ashley Kirzinger, KFF

The Case for National Measures of Loneliness, Social Support, and Social Isolation
Edna Ishayik, National Conference on Citizenship

Public Trust in Official Statistics: Evidence from the Household Trends and Outlook Pulse Survey
Jennifer Childs, U.S. Census Bureau

The National Survey of Children’s Health: Informing Policy, Program, Legislation, Practice, and Research
Reem Ghandour, Health Resources and Services Administration

The Importance of Federal Data in Advancing State and Local Policy Improvements
Monica Saucedo, California Budget & Policy Center

Where to Go from Here to Increase Credibility and Public Trust in Federal Statistics
Nick Hart, Data Foundation