Spring 2026 Colloquia
Monday, March 30, 2026
Designing and Managing Nonprobability Panels: A Practitioner’s Perspective
ABSTRACT: The Bovitz nonprobability panel, originally built for consumer market research, has gained significant traction among academics conducting population-based survey experiments. This talk provides a behind-the-scenes look at how the panel is designed and managed to meet researchers’ expectations around representativeness and respondent engagement, highlighting key design decisions, quality control practices, and the practical tradeoffs inherent in nonprobability sampling. The presentation will include key panel metrics along with empirical results from select studies.
Dr. Greg Bovitz, founder and CEO of Bovitz, Inc.
Monday, April 6, 2026
A Dynamic Dyadic Systems Approach for Analysis of Conversations: Reconsiderations and Possible Routes Forward
ABSTRACT: Conversations between people are where stressors are amplified and attenuated, conflicts are entrenched and resolved, and goals are advanced and thwarted. What happens in dyads' back-and-forth exchanges to produce such consequential and varied outcomes? Although numerous theories in communication and in social psychology address this question, empirical tests of these theories often operationalize conversational behavior using either discrete messages or overall features of the conversation. Dynamic systems theories and methods provide opportunities to examine the interdependency, self-stabilization, and self-organization processes that manifest in conversations over time. In this talk I review how our Dynamic Dyadic Systems (DDS; Solomon et al, 2023)) approach emerged, the hurdles we encountered, and how the temporal analysis frameworks might be extended.
Nilam Ram, Professor of Psychology and Communication, Stanford University
Monday, April 13, 2026
Data Donation : Social Media Research After APIs
ABSTRACT: This talk introduces data donation as an alternative methodology to API-based studies for social media research. Drawing on recent pilot studies conducted at NORC at the University of Chicago, I demonstrate how data donation can be operationalized at scale and integrated with survey data to link observed exposures and interactions with demographics, attitudes, and outcomes. I highlight successful research applications in public health, including a large-scale pilot measuring exposure to gambling content on TikTok, Instagram, and Facebook. The talk concludes by briefly addressing methodological challenges and emerging work across platforms such as Netflix and ChatGPT.
Soubhik Barari, Senior Research Methodologist, University of Chicago NORC
Monday, April 20, 2026
Dumpster Diving to Data Mining: Assessing & Promoting a Culture of Sustainability at Stanford
ABSTRACT: From what we purchase, eat, and throw away to how we travel and the civic engagement opportunities and careers we pursue, our behaviors impact the health of people and the planet. The behaviors we establish and engage in during our time at Stanford as students, staff, faculty, or even visitors, are no exception. Not only do our behaviors impact campus sustainability, the identities, values, beliefs, knowledge, and other factors that drive these behaviors also persist and impact sustainability outcomes far beyond the university’s boundaries. This talk tells the tale of two campus sustainability practitioners (a frustrated sustainability social scientist and her star student) as they attempt to better understand and proactively shape pro-sustainability behaviors on campus and the cultural context at Stanford that drives them. Follow their journey from standing knee-deep in garbage to assessing sustainability culture and literacy through annual student surveys. Early findings from the first surveys released in 2025 indicate that Stanford’s students largely display strong biospheric and altruistic values, pro-sustainability identities and beliefs, and sustainability knowledge, and suggest that the campus environment generally facilitates pro-sustainability behaviors. However, results also reveal a relative weakness in public-sphere environmental behaviors (e.g., participating as an active member in a local environmental group, signing a petition about an environmental issue) as well as an opportunity to leverage findings to more effectively foster sustainability culture and help close this apparent value-action gap. Along the way, the talk does not shy away from a candid reckoning with what sustainability practitioners have gotten wrong and offers concrete examples of how survey insights and findings from the social sciences are driving a new generation of approaches that are more targeted, behaviorally informed, and, hopefully, more effective.
Gretchen Engbring, Sustainability Social Scientist, Stanford University Office of Sustainability
Sebastian Pintea, Environmental Systems Engineering BS Candidate, Stanford University
Monday, April 27, 2026
Linking Individual and Systems Level Climate Solutions Through Social Norms and Collective Advocacy
ABSTRACT: In recent years, there has been growing criticism of research in the behavioral sciences, behavioral public policy, and environmental psychology for treating climate change as an individual rather than systemic problem, with individual rather than systemic solutions. In this talk, I argue that this criticism overlooks the interdependent relationship between individual and systems change in two ways. First, individuals are socially embedded and interdependent. This means that behavior change interventions often have both direct effects but also indirect spillover effects on others that can continue beyond the intervention. These spillover effects have implications for whether a circumscribed intervention spreads or stalls, and who should be targeted by interventions to initiate broader social change. Second, structural or systems change often comes about because of public pressure, advocacy and collective action. Yet, these types of political behaviors are uncommon in the context of climate change. Recent efforts to encourage these behaviors with psychologically-informed interventions find mixed results. I will give a brief overview of this broader research agenda, and then discuss a couple of working papers in these two areas. I would love feedback as we are preparing the papers for submission!
Sara Constantino, Assistant Professor, Stanford Doerr School of Sustainability
Monday, May 4, 2026
From Democratic Theory to Platform Policy: Six Years Inside Big Tech
ABSTRACT: What does Communication scholarship look like from inside a large technology company? Drawing on six years in Meta's policy organization, Andreas Katsanevas, PhD, returns to the department with a report from the field. The talk traces the arc from his dissertation on democratic theories of the information society into the day-to-day of tech policy at Meta. It examines how Communication research travels inside tech companies and reflects on the gap between academic and corporate tempos of operation. It closes by asking what Al chatbots-a new kind of conversational infrastructure-mean for the questions Communication scholars could take up next.
Andreas Katsanevas, Head of Governance Innovation, Meta
Monday, May 11, 2026
Do We Know What We Think? The Availability and Accuracy of Information on Public Opinion on Public Policy
ABSTRACT: The idea that public opinion on policy should play an important role in government policymaking is a central idea of the Founders of the American republic and in current theories of democracy. Naturally for this to occur elected officials must have a good understanding what is public opinion on public policy. The public must also know what public opinion is for the public to know whether policy is in fact aligned with it. A key means for both policymakers and the public to gain such information is through press reporting of survey data. But does this occur? The present study took five policy topics prominent in the recent news discourse for which there was a readily available high quality polling data. Fifteen prominent media outlets were selected. All news articles on the selected policy topics from those media outlets over a 16 month period were assessed (6,094 articles) as to whether there was any mention of public opinion data. Polling data was mentioned in only a small minority of cases. Could this be because Americans do not have interest in such data? A nationwide survey of a non-probability sample of 1,174 Americans found large majorities expressing interest in such data and a strong preference for media outlets to report on them. Do Americans nonetheless have a good understanding of public opinion on these policy topics? In the same survey respondents were asked what they believed public opinion was on 23 questions related to the same five topics explored above and for which there were clear and consistent patterns of public opinion across multiple polls. Respondents made accurate assessments of majority views barely more than half the time. Partisans accurately assessed the majority opinion of partisans of the opposing party less than half the time.
Steven Kull, Director of the Program for Public Consultation (PPC), University of Maryland
A Reanalysis and Meta-Analysis of All Studies Comparing Probability and Non-Probability Samples In Terms of Accuracy
ABSTRACT: While much research has established probability sample surveys as the ‘gold standard’, non-probability sampling has been lauded by some commercial firms as quick, easy, and as accurate as probability samples, if not more. And the use of non-probability samples has been standard practice in some academic disciplines, such as political science and psychology, accompanied by claims that such samples are "representative". This has contributed to a long-running debate regarding which of the two sample types is more accurate. The debate is further complicated by the fact that papers providing relevant evidence have used a diverse variety of metrics to measure accuracy – RMSE, average absolute error, etc. In collecting and examining papers on comparative probability and non-probability sample accuracy, we have discovered a range of errors committed in papers. For example, in some cases, results obtained from commissioned surveys were compared to benchmark surveys that used substantially different question wordings. And in other cases, important information about benchmark sources or field dates were misreported or omitted by authors – meaning that, in some papers, we were unable to verify whether certain errors may have been committed. We can conclude that at least some data points in the sample accuracy debate are not entirely reliable.
Lavi Sundar, PPRG Undergraduate Researcher, Stanford University
Incentivizing Participation: Developing an Evidence-Based Guide for Best Practices in Survey Incentives
ABSTRACT: In an era defined by challenges to data integrity and declining public trust, the pursuit of representative sample survey data has never been more critical. Falling response rates are thought to threaten the validity of public opinion research and its role in informed decision-making. While survey incentives are a common industry practice for boosting response rates, their application is far from standardized. Researchers face complex choices regarding an incentive's form (cash, voucher), timing (prepaid vs. promised), and amount, with little clear guidance on how to maximize participation while minimizing costs. This project, which is being developed as a section of a larger manual on optimal survey design, is part of the ongoing work of the newly-established Stanford Institute for Excellence in Survey Research (SIESR). This review of survey incentives directly supports SIESR's mission to advance survey methodology and disseminate findings not only to academics but also to media and policymakers. Synthesizing findings from a wide-ranging literature review of experimental research, this project moves beyond academic study to propose a decision-making framework for practitioners, fostering methodological innovation and helping to maximize confidence in survey data.
Josearmando Torres, Predoctoral Researcher, Stanford Institute for Excellence in Survey Research
Resty Fufunan, Predoctoral Researcher, Stanford Institute for Excellence in Survey Research
Monday, May 18, 2026
Media Bias in Framing Violence in Israel/Palestine: How Language Shapes Public Perceptions of Agency and Accountability
ABSTRACT: News media play a central role in shaping popular narratives and public understanding of intergroup conflict and violence. In particular, media reportage may use linguistic forms that can obscure harm and deflect attention from perpetrators. Combining large-scale text analysis with experimental methods, this study investigates the prevalence and consequences of such language in news coverage of violence in Israel/Palestine. Using a large language model, we analyze thousands of headlines from major outlets—including the Associated Press, The New York Times, The Guardian, and Al Jazeera—and identify differences in how outlets describe intergroup violence, with the NYT and The Guardian more frequently employing agency-obscuring language when describing Israeli perpetrated harm. To examine the consequences of these patterns, we conduct three experiments with U.S. respondents (total N ≈ 3,500), including one preregistered study. Across studies, exposure to agency-obscuring language reduces perceived perpetrator responsibility, inferred intent to harm, and support for holding perpetrators accountable. These findings suggest that linguistic choices in news media may serve to condone intergroup violence, functioning as a tool of discursive oppression.
Siwar Aslih, Assistant Professor of Social Psychology, University of California, Santa Cruz
Monday, June 1, 2026
How Social Media Creators Shape Mass Politics: A Field Experiment during the 2024 US Elections
ABSTRACT: Political apathy and skepticism of traditional authorities are increasingly common, but social media creators (SMCs) capture the public's attention. Yet whether these seemingly-frivolous actors shape political attitudes and behaviors remains largely unknown. Our pre-registered field experiment encouraged Americans aged 18-45 to start following five progressive-minded SMCs on Instagram, TikTok, or YouTube between August and December 2024. We varied recommendations to follow SMCs producing predominantly-political (P), predominantly-apolitical (PA), or entirely non-political (NP) content, and cross-randomized financial incentives to follow assigned SMCs. Beyond markedly increasing consumption of assigned SMs' content, biweekly quiz-based incentives increased overall social media use by 10% and made, participants more politically knowledgeable. These incentives to follow PP or PA SMCs led participants to adopt more liberal policy positions and grand narratives around election time, while PP SMCs more strongly shaped partisan evaluations and vote choice. PA SMCs were seen as more informative and trustworthy, generating larger effects per video concerning politics. Participants assigned to follow NP SMCs instead became more conservative, consistent with left-leaning participants using social media more when right-leaning content was ascendant. These effects exceed the impacts of traditional campaign outreach and partisan media, demonstrating the importance of SMCs as opinion leaders in the attention economy as well as trust-and volume-based mechanisms of political persuasion.
Eunji Kim, Assistant Professor of Political Science, Columbia University