Skip to content

Communication Science Futures will take place September 18th-20th, 2026 at the MSU Union on the campus of Michigan State University. Below you will find the tentative conference schedule. Note that specific times and some elements of the schedule are still being finalized and are subject to change.

Conference Location

MSU Union, 49 Abbot Rd, East Lansing, MI 48824

Friday, September 18

4:00pm

Doors Open & Check-In
5:00 - 5:50pm

Keynote Address
James W. Pennebaker
Keynote Speaker
James W. Pennebaker
The Accidental Text Analyst: Tracing Words from Trauma to Artificial Intelligence
In this talk, I'll provide a brief history of how my research and thinking gradually changed from physical symptoms and expressive writing to text analysis, function words, and AI. It has not been a straightforward path. Rather, I've been guided by serendipity, rumor, and luck. Once I began studying language, it became apparent that the words we use can provide a window into how the individual mind works. More exciting, by exploring the language of large groups of people, we are now seeing the interplay of the individual with the culture around them. The revolution of big data and AI is now reshaping how we think about our research questions, methods, theories, our profession, and the entire academic enterprise. In some ways it is threatening. In others, it is exhilarating. I conclude with some thoughts about the future of theory, methods, and measurement in the social sciences.
6:00 - 8:50pm

Opening Reception & Dinner
Join us for dinner and drinks following the keynote address.

Saturday, September 19

8:30 - 8:50am

Welcome & Introduction
Breakfast, coffee, and tea will be provided.
9:00 - 10:20am

Panel 1 — Futures in Research
Stage presentations selected from the submission pool, exploring forces, trends, and topics shaping the future of communication research.

Presentations:

How can communication scholars best study gaps? This talk addresses such question from a social support perspective. Support gaps are discrepancies between the social support people desire and what they receive. The talk maps out several key theoretical questions: Where do support desires come from? Is an exact match between desired and received support always optimal? How are support gaps created or mitigated within networks of relationships? And how do these gaps change over time? This talk calls for applying a broader range of approaches, including dyadic, longitudinal, and network methods, to address theoretical questions that have long puzzled scholars.

Communication research has poured energy into the outcomes of communication while paying far less attention to the process itself, and to what it means to understand and be understood. This talk presents a model of communication as the creation of understanding, grounded in an evolutionary account of humans as social animals, with brains that operate as energy-efficient prediction engines. The model provides a common foundation for extant as well as the next generation of communication theories, addressing issues such as uncertainty, expectation, deception, and intergroup interaction.

Generative AI can search, synthesize, compare, and recommend in a single exchange, absorbing stages of evaluative work that earlier tools never touched. This talk introduces cognitive offloading as a bridge between distributed cognition and communication's models of persuasion and learning, and maps a delegation continuum along which that work shifts from person to system. Early experiments point to a recurring trade-off, where reliable AI help can improve decisions while weakening memory and inflating confidence, a tension the next wave of persuasion and learning research will have to take seriously.

A large amount of experimental mediation analysis in the field is run through PROCESS Model 4. This talk argues that the model often works against the experiments it is meant to serve. Simulations built on a real design show that the direct (c′) path controls away the very variance an experiment is built to induce, so PROCESS returns significant indirect effects whether or not the data fit the theory. The recommendation is a serial causal-chain SEM, and a cleaner standard for testing mediation going forward.

AI writing and coding agents have moved from novelty to infrastructure in many research workflows, but the field has no shared way to judge what they do to how knowledge gets made. This talk works through three fault lines: productivity versus rigor, authorship and transparency, and capability versus dependency. The goal is to locate where the real tradeoffs sit and to start naming the norms the field will need as these tools reshape how research gets done.

The continuums of moral influence paradigm explains how media cues move moral judgment along a continuum that runs from disengagement to engagement. The studies in this program stay deliberately non-normative and break moral influence into basic perceptual and judgment processes that can be measured directly. This talk covers the strategy behind the line of research, the initial findings, and where a paradigm like this could take moral-influence research, from entertainment to political communication to cyberbullying.

10:30 - 11:50am

Breakout Session 1
Choose one of nine 80-minute, facilitator-led round-table discussions. Select a card to see the question and conversation plan when details are available.

Choose a discussion (9):

In 1985, Kellermann wrote persuasively that ‘It seems trite to write that memory mediates the effects on an individual of mass communication...’ yet, forty years later, communication theories lack a comprehensive framework for memory’s role in downstream communication effects. In this discussion we will explore the core areas of current media memory research, illustrate operationalization challenges with two case studies, and attempt to reconcile the issues central to understanding memory as a key player in communication science research.

Central question

What is the role of memory in communication effects?

Conversation starters

  1. Communication studies often use recall or recognition as proxies for attention or manipulation checks rather than treating memory as an active cognitive process. Which parallel threads of memory research across communication and psychology could be consolidated to isolate the field’s core conceptual concerns?
  2. How can we operationalize memory across communication subfields? When are recall and recognition appropriate, when should other paradigms be used, and what do those measurement choices mean for models of message and media processing?

Where the conversation is headed

A roadmap for integrating memory into communication research.

Suggested reading

  • Braun-Latour, K. A., & Zaltman, G. (2006). Memory change: An intimate measure of persuasion. Journal of Advertising Research, 46(1), 57–72.
  • Furman, O., Dorfman, N., Hasson, U., Davachi, L., & Dudai, Y. (2007). They saw a movie: Long-term memory for an extended audiovisual narrative. Learning & Memory, 14(6), 457–467.
  • Lepsius, M. (2026). The brain edits the past: Neural systems of memory reconstruction and self-stability. New Ideas in Psychology, 80, 101221.
  • Shehata, A., Thomas, F., Glogger, I., & Andersen, K. (2024). Belief maintenance as a media effect: A conceptualization and empirical approach. Human Communication Research, 50(1), 1–13.

Social support is important for well-being, yet receiving support is not always beneficial. Support can impose face threats, create unwanted obligations, or differ from what recipients desire. These complexities have drawn researchers’ attention to support gaps, or the discrepancies between the social support people desire and receive. Although research on support gaps has generated important insights, several key questions remain underexplored: what contributes to the development of support gaps, and why deficits and surpluses produce distinct outcomes. Clarifying these issues can advance support gap theorizing and provide a stronger foundation for future research and replication.

Central question

How do support gaps emerge through communication within and across relationships?

Conversation starters

  1. What individual, relational, situational, and cultural factors shape support desires? How should we theoretically distinguish desired support from expected support, and to what extent are desires formed prior to interaction versus reconstructed through communication itself?
  2. While support gap research often focuses on individual recipients’ perceptions, supportive communication is inherently interactive. How do support gaps emerge as dyadic, two-way communication processes involving both the provider’s behaviors and the recipient’s responses?
  3. What theoretical mechanisms explain why support deficits and surpluses produce distinct consequences for personal and relational well-being? Under what specific conditions is exact matching optimal, versus an optimal support surplus?
  4. How do support gaps evolve across conversation moves, coping episodes, and broader relational trajectories, and how are they created or mitigated across a person’s broader relational network?

Suggested reading

  • Crowley, J. L., & High, A. C. (2020). Validating the support gaps framework: Longitudinal effects and moderators of experiencing deficits and surpluses during supportive interactions. Communication Quarterly, 68(1), 29–53.
  • Pederson, J. R., High, A. C., & McLaren, R. M. (2020). Support gaps surrounding conversations about coping with relational transgressions. Western Journal of Communication, 84(2), 204–226.
  • Youngvorst, L. J., & Ruppel, E. K. (2025). Social support among emerging adult friends: Dyadic and longitudinal associations between support gaps and relational quality. Journal of Social and Personal Relationships, 42(7), 1629–1656.

AI is increasingly becoming part of everyday online communication, helping people write interpersonal messages. How should we understand the changes AI introduces, and can existing computer-mediated communication (CMC) theories explain them? This discussion explores how CMC theories and concepts can inform the study of AI-mediated communication (AIMC), where existing theories may need revision, and what new theoretical perspectives are needed to address questions unique to AIMC.

Central question

How can existing CMC theories and concepts inform our understanding of AIMC, and what new theoretical questions and concepts should AIMC research address that may not have been central to traditional CMC research?

Conversation starters

  1. Do we need to reinvent our theories every time a new technology emerges? Why or why not?
  2. What role does AI play in your research, now or in the future? Does it challenge any theories or assumptions you work with?
  3. If you had to choose, is AI’s overall impact on interpersonal communication more positive or negative? Why?
  4. Which interpersonal communication or CMC theories or concepts are useful for understanding AIMC, and which most need reconsideration? Why?
  5. What theoretical questions or concepts become more important in AIMC than they were in traditional CMC research?

Suggested reading

  • Hancock, J. T., Naaman, M., & Levy, K. (2020). AI-mediated communication: Definition, research agenda, and ethical considerations. Journal of Computer-Mediated Communication, 25(1), 89–100.
  • Liu, B., Liu, J., & Wei, L. (2026). Thinking, fast, and artificial: Processing fluency in AI-mediated relational maintenance planning and its impact on relationship perceptions. Journal of Computer-Mediated Communication, 31(4), zmag018.
  • Zhu, R., Markowitz, D. M., & Van Der Heide, B. (2026). Presenting the augmented self: Extending sender effects of the hyperpersonal model in artificial intelligence-mediated communication. Annals of the International Communication Association, wlag029.

Political communication is being remade by AI-generated content, algorithmic curation, and synthetic media, often faster than our theories, methods, or institutions can keep up. This session invites participants to think beyond present-day fixes toward the deeper question of how what constitutes political communication, along with related concepts such as legitimacy, trust, deliberation, and political speech, will evolve over the next decade. Expect open discussion, skepticism, and predictions about the future of the field.

Central question

As AI-mediated content, algorithmic curation, and synthetic media reshape how political information is produced and consumed, how will our conceptualization of “political communication” evolve, who (or what) will count as a communicator under that new conceptualization, and what questions will become central to what these developments mean for democratic processes?

Conversation starters

  1. The authorship question: When a political message is drafted by an AI, personalized by an algorithm, and delivered through a synthetic avatar, who is accountable for its content: the candidate, the platform, the AI developer, or the voter who engaged with it?
  2. Fragmentation vs. fusion: Are we heading toward further fragmentation of political (and even non-political) publics through hyper-personalized feeds and niche communities, or a strange re-fusion around a handful of AI-mediated gatekeepers such as chatbots and recommendation engines? What evidence points each way? Does the field need to reconsider Postman’s thesis that current trends further facilitate the potential that we “amuse ourselves to death”?
  3. Trust infrastructure: Traditional trust signals (bylines, institutional mastheads, “I’m a real journalist”) are eroding as synthetic content becomes indistinguishable from human-made content. What new trust infrastructures, technical, social, or regulatory, might replace them? What are the implications for epistemic trust?
  4. The global asymmetry: Most political communication research and platform policy is built around U.S. and Western contexts. How might the futures we imagine look different in electoral democracies with weaker institutions, different media systems, or non-English-dominant information ecosystems? Which systems are best positioned to weather what is coming, if any?
  5. The future of democracy: As rapidly accelerating technologies reshape what counts as political communication, how should research agendas evolve to consider the implications for the exercise of, and support for, democratic systems?

Where the conversation is headed

Recommendations for new research areas, such as comparative platform research on how different regulatory approaches (the EU AI Act, national deepfake laws) shape political communication on the ground; AI-mediated persuasion at the “micro-targeting 2.0” level of real-time, conversational persuasion via chatbots and voice assistants; and political communication in non-Western, non-English information ecosystems.

Suggested reading

  • Young, D. G., & Miller, J. M. (2023). Political communication. In L. Huddy et al. (Eds.), The Oxford handbook of political psychology (3rd ed., pp. 555–600). Oxford University Press.
  • Lawrence, R. G. (2025). Editor’s note, January 2025. Political Communication, 42(1), 1–5.

Online hate research often centers on hateful messages, including their content, detection, prevalence, and effects. This discussion asks what changes when we instead treat online hate as an ongoing communication process. We will consider how targets, bystanders, and perpetrators respond; how these interactions shape subsequent communication and relational outcomes; and how platform features reinforce, redirect, or interrupt these processes.

Central question

How does our understanding of online hate change when we study it as an unfolding communication process rather than primarily as hateful content?

Conversation starters

  1. If hateful content is only one part of the communication process, what else should online hate research examine?
  2. Who are the relevant actors and audiences in online hate, and how should we understand their roles?
  3. How does feedback to online hate, and from whom, influence subsequent behaviors?
  4. What social and relational outcomes can emerge from online hate as a communication process?
  5. What theories and methods are needed to study online hate as a dynamic communication process?

Where the conversation is headed

Identify key questions and develop a research agenda for studying how online hate unfolds through interaction, feedback, and relational outcomes.

Suggested reading

  • Walther, J. B. (2025). Making a case for a social processes approach to online hate. In J. B. Walther & R. E. Rice (Eds.), Social processes of online hate (pp. 9–36). Routledge.
  • Matthes, J., Koban, K., Bührer, S., Kirchmair, T., Weiss, P., Khaleghipour, M., Saumer, M., & Meerson, R. (2026). The state of evidence in digital hate research: An umbrella review. Communication Research, 53(6), 867–896.
  • Walther, J. B. (2026). The effects of social approval signals on the production of online hate: A theoretical explication. Communication Research, 53(6), 751–777.

Experiments are designed to induce variance into a process and to isolate the effects of hypothesized causes. In experimental mediation analysis, these purposes require asking whether variance induced by a manipulation is transmitted through a proposed mediator to an outcome as theory predicts. Through staged analysis of contrasting simulated experiments, participants will (a) examine how stimulus design, random assignment, measurement validity, and model specification shape mediation claims, (b) compare competing analytical approaches, and (c) develop practical standards that maintain the possibility of falsification.

Central question

How should researchers use mediation analysis in ways that account for experimental design, random assignment, and induced variance while maintaining the possibility of falsification?

Conversation starters

  1. In a simple randomized experiment: (a) What relationship should we expect between X and Y, if X has no effect on Y? (b) How is variance in Y partitioned into variance associated with experimental condition and residual variance, and what role does random assignment play in interpreting that partition? (c) What does a significant observed X-Y relationship mean in this context, and what inference can be made?
  2. Suppose a theory proposes that M causes Y. Researchers decide to test this causal relationship by manipulating a message feature X to induce different levels of M. (a) What role does measuring M play in testing this theory? (b) What does the strength of the effect of X on M imply for the model? (c) If the hypothesized X→M→Y causal chain is correct, what relationships should we observe among X, M, and Y? (d) What does the model assume about the residuals of M and Y, and what would nonindependence of those residuals imply?
  3. When a direct X-Y path is included in the mediation model: (a) What variance is removed from the estimation of the M-Y path? (b) What variation in M is then used to estimate the M-Y path? (c) How does this change the meaning of the M-Y path and the indirect effect?
  4. How does including or excluding the direct X-Y path affect the falsifiability of the hypothesized X→M→Y causal chain? What can the fit or misfit of a model that excludes the direct X-Y path tell us that a fully saturated mediation model cannot?
  5. Given the inferential purpose of an experiment, what standards should researchers follow when designing, analyzing, interpreting, and reporting experimental mediation studies so that the hypothesized causal chain remains falsifiable?

Where the conversation is headed

A preliminary set of practical standards for designing, analyzing, interpreting, and reporting experimental mediation studies in ways that maintain the possibility of falsification and identify boundary conditions, along with areas of disagreement and unresolved methodological questions.

Suggested reading

  • Grizzard, M., Carpenter, C. J., & Frazer, R. (2026; pre-print). Concerning the c′ Path: How PROCESS-based Mediation Undermines Experimental Methods. https://doi.org/10.31234/osf.io/39uqf_v1
  • O'Keefe, D. J. (2003). Message properties, mediating states, and manipulation checks: Claims, evidence, and data analysis in experimental persuasive message effects research. Communication Theory, 13(3), 251–274. https://doi.org/10.1111/j.1468-2885.2003.tb00292.x

AI is entering communication research faster than we can agree about it (or societies, journals, and universities can put out policy). Some see a threat to rigor and to training; others see the biggest expansion of what our field can study in decades. Both may be right. Some of what divides us is a value judgment; some of it is simply operational and solvable. This session tries to tell those apart and make progress on how we deal with AI in communication.

Central question

How do we make sure we are getting the science right and the training right?

Conversation starters

  1. Think of one time you used AI in your research and weren’t sure it was okay. What was the task, and what made you hesitate?
  2. What is the best paper our field currently cannot write? If a question were unaskable only because it was too big, what would you go after?
  3. Would you rather train in a lab that treats offloading writing as offloading thinking, or one that treats learning these tools as part of scientific training? Which lab would you rather hire from?
  4. Your student can produce three papers this year using AI, or one without. What should they do?
  5. For graduate students: What does AI make better, and what does it make worse?
  6. For post-PhD researchers: What does AI make better, and what does it make worse?

AI chatbots offer promising new possibilities for expanding access to mental health information, support, and interventions. This discussion will explore the potential of AI chatbots to support mental health, alongside ethical challenges such as IRB review, data monitoring, privacy, and unintended harms. Participants will collaboratively develop a practical list of recommended safeguards for responsible research and implementation.

Central question

How can communication researchers study AI chatbots for mental health in ways that are theoretically grounded, practically effective, and ethically responsible?

Conversation starters

  1. What communication theories can inform mental health chatbot research?
  2. How should researchers design and implement chatbot interactions, including expert input, prompt development, and language model selection?
  3. How should we evaluate whether a mental health chatbot actually works?
  4. What safeguards are necessary when participants interact with mental health chatbots, including IRB disclosure, data monitoring, and procedures for responding to unintended harms or adverse events?
  5. What does rigorous and transparent research practice look like when the technology is constantly changing?

Where the conversation is headed

Develop a practical research and design roadmap organized around four stages: ground the research, build the interaction, protect participants, and evaluate and report.

Earlier, we argued that Communication should embrace creating a state of understanding as the fundamental goal of communication. Understanding is conceptualized as an isomorphic meme manifesting as neurological alignment. We suggested that three premises referring to evolved capacities of humans can serve as foundational to explanations about human communication behavior: humans are fundamentally social in orientation, human mental processes are governed by a predisposition for efficiency, and predictive inference-making is a core feature of human mental activity.

Central question

If the discipline of Communication were to embrace Creating Understanding as the fundamental goal of communication, how would that affect undergraduate and graduate education, research agendas, and the discipline’s relationship to natural science?

Conversation starters

  1. If understanding is a state of neurological alignment, how should graduate curricula expand to prepare the next generation of researchers?
  2. Does an emphasis on creating understanding suggest new paradigms for research?
  3. Is there any warrant for claiming that Communication should be considered a natural science first and a social science second?

Where the conversation is headed

Disruption with the potential for transition.

Suggested reading

  • Chapters 2 and 3 of Creating Understanding.
12:00 - 1:20pm

Lunch & Poster Session 1

Posters (19):

Coming to Like: A Drift Diffusion Model of Audiences' Enjoyment and Appreciation
Joshua Baldwin Boston University
Internationalizing Communication Science Beyond WEIRD Samples: The Case for the Taiwan Communication Survey
Nick Bowman Syracuse University
Advancing the Theory of Stigma Management Communication
Jenny Crowley University of Tennessee
Transactive Memory Systems in Patient-provider Relationships
Emily T. Dawson Purdue University
Feeling Scattered: Development and Validation of the Attention Fragmentation Scale
Jacob T. Fisher Michigan State University
What Makes a Supportive Relationship? Testing Support Gaps Across Relationships
Junwen M. Hu University of California, Davis
Visual Attention and Cognitive Responses in Ego-Defensive Message Processing
Moonsun Jeon Michigan State University
Inoxity: An Open-Source iOS Platform for High-Throughput Data Collection
Rachael Kee University of California, Davis
Should I Tell My Advisor? Negotiating Generative AI, Scholarly Independence, and Responsibility in Doctoral Research
Victoria A. Kyriakopoulos Syracuse University
Haptics as a Modality of Communication: A Systematic Review
Alex Lover Michigan State University
Can Constructive Memory Help Us Study Digital Persuasion?
Ewa Maslowska University of Illinois Urbana-Champaign
Formalizing Dynamic Theories of Media Selection and Effects Processes: From Four Models to Recurrent Neural Networks
Rinseo Park Stanford University
Capturing and analyzing TikTok content in low-resource languages
Yvens Rumbold Michigan State University
Communicating Rural Consciousness
Xavier Scruggs Wake Forest University
Disentangling Entertainment Appraisals: Tracking Enjoyment, Appreciation, and Psychological Richness
Manushka Sondhi Michigan State University
Perceptions, Trust, and Participation: A Mediation Model of Political Engagement
Stephen Spates Michigan State University
Sensation Seeking Dictionary: Developing a Theory-Driven NLP Tool for Communication Research
Yan Wang Michigan State University
Asymmetric Reciprocity in Posthuman Intimacy: Romantic Chatbots in China
Keting Zhang University of Nottingham Ningbo China
Everywhere and Nowhere: A Scoping Review of Anonymity in Online Hate Research
Yidi Zhang UC Santa Barbara
1:30 - 2:50pm

Panel 2 — Futures in Theory
Stage presentations addressing issues and advancements in communication theory building, testing, and formalization.

Presentations:

Communication science keeps accumulating undead theories, frameworks that persist even after they have been repeatedly disconfirmed. This talk traces the problem to the hypothetico-deductive habit and offers three correctives: abduction in place of theory-first testing, risky predictions in place of rejecting the nil-null, and attention to variance in place of group averages. Truth-Default Theory is the worked example of where a healthier approach to theory building can lead.

Research on human-AI interaction has piled up far faster than the theory needed to organize it, leaving a literature long on findings and short on coherence. This talk lays out four desiderata for the next generation of HAII theory: predictions that are bold and falsifiable, generality beyond any one platform or moment, ideas that open new questions rather than just sorting old ones, and communication kept at the center of the account.

Translating theory into method requires stripping away nuance. When the same nuances are stripped away study after study, this can lead to bodies of research drifting from the frameworks they intend to test and refine. This talk develops the idea of “theoretical drift” using a systematic review of the Differential Susceptibility to Media Effects Model, demonstrating how tightening the link between method and theory could allow research to better drive theory development.

Inoculation is an attitude theory, yet a striking number of recent studies that claim to use it never measure attitudes, threat, or refutational preemption, its core components. This talk reviews five decades of inoculation research to document that conceptual drift and asks what it means for replication when studies share a label but not the construct. The larger question is what the future of inoculation research looks like if the theory's own building blocks keep going missing.

Communication theories often rest on metaphors, with inoculation theory borrowing from biological immunity as a familiar example. This talk introduces the Borgian Problem, drawn from a Borges story about a map so detailed it becomes useless, to ask whether elaborating a metaphor (adding ever more vaccine-like constructs and mechanisms) actually advances understanding of persuasion or just perfects the metaphor. The talk employs a thought experiment to discuss what should count as theoretical progress.

Marginalization, the experience of being pushed to the edges of a group, is accomplished largely through communication, yet the research on it sits in separate silos under labels like stigma communication, disconfirmation, and disenfranchising talk. This talk presents a theory of marginalizing messages that pulls those strands into one framework, with five assumptions and five propositions about how, why, and when a message marginalizes. The aim is a shared foundation that future work on belonging and exclusion can build on.

3:00 - 4:20pm

Breakout Session 2
Choose one of nine 80-minute, facilitator-led round-table discussions. Select a card to see the question and conversation plan when details are available.

Choose a discussion (9):

We aim to explore how scientific progress unfolds in communication science and how metaphors and other surrogate reasoning systems advance or constrain theoretical progress. We also use the Borgian Map problem as a thought experiment to refine our assumptions about what counts as theoretical progress.

Central question

To what extent are metaphors useful for theoretical progress in communication science?

Conversation starters

  1. What does scientific progress look like in communication science? What are we trying to progress in theorizing (e.g., predictive power, explanatory power, question generation, parsimony, etc.)? To what extent can metaphors be useful for theoretical progress?
  2. To what extent are metaphors useful for advancing our goals for scientific progress? What are common metaphors that have been used in theorizing communication, and do they generally tend toward Borgian Map explanations of communication phenomena? Are there empirical or theoretical conditions that necessitate using metaphors as a surrogate reasoning system? Are there communication phenomena where a Borgian Map may actually be epistemically beneficial?
  3. Aside from metaphors, what other surrogate systems in communication science might enable or constrain progress? What are common surrogate systems used in the discipline and its different subfields?
  4. If not metaphors, what other approaches might overcome or complement the utility of metaphors in theorizing communication?

Suggested reading

  • Borges, J. L. (1946/1998). On exactitude in science (A. Hurley, Trans.). In A. Hurley (Trans.), Collected fictions (p. 325). Penguin Books. https://kwarc.info/teaching/TDM/Borges.pdf

This session will focus on how published research has drifted from the known parts of the theory. What happens when researchers ignore the underpinnings of the theory and its working components and this work makes it into the field?

Central question

How do researchers correct the literature when the research has gone off course, ignoring the variables measured and the established working components of the theory?

Conversation starters

  1. What role do we have as communication scholars to hold ourselves accountable for the theoretical drift occurring?
  2. What is the best course of action to correct published literature that does not measure the established theoretical variables?
  3. How do we address theoretical propositions woven into the literature as proven findings rather than claims about what might occur?

This session will focus on the process of testing a new theory. We will (a) talk about theory evaluation and discuss strengths and weaknesses of the theory of marginalizing messages (TMM), (b) develop ideas for testing TMM and communication contexts to which it can be applied, and (c) think about useful visual representations for theory building. While the focus will be on TMM, this discussion will be useful to anyone interested in theory building, testing, and evaluation.

Central question

How can a theory of marginalizing messages be tested and used in different contexts?

Conversation starters

  1. What constructs are similar to marginalizing messages, where do they overlap, and what makes marginalizing messages unique?
  2. What would it look like to test this theory, and which existing measures could be useful?
  3. To which communication contexts could the theory be applied, and what adaptations would be needed?
  4. How could the propositions be represented in a figure? What makes for a useful visual in a theory piece?
  5. Using DeAndrea and Holbert’s table as a starting point, how could areas of weakness in the theory be strengthened?

Where the conversation is headed

Sketch a research agenda for a systematic program of testing, extending, and applying the theory of marginalizing messages.

Suggested reading

  • Dorrance Hall, E., & Wilson, S. R. (2021). Explicating the dimensions and types of marginalized family members. Journal of Social and Personal Relationships, 38(7), 2099–2120.
  • Dorrance Hall, E., & Gettings, P. (2020). ‘Who is this little girl they hired to work here?’: Women’s experiences of marginalizing communication in male-dominated workplaces. Communication Monographs, 87(4), 484–505.

Narratives have been defined at different levels based on researcher assumptions, disciplinary conventions, and experimental needs. But an overarching examination of what constitutes a narrative in communication science has been lacking. We have gathered some extant definitions with the hopes of discussing which are useful for communication science and how we can productively think about and use narratives in our research.

Central question

What is a useful definition and (perhaps) operationalization for narrative in communication science?

Conversation starters

  1. What constitutes the smallest meaningful and measurable unit of a narrative? What are its necessary and sufficient components?
  2. How do our definitions of narrative serve different hierarchical levels of our research questions, both at the moment-to-moment (local) and broader story-structure (global) levels? Do the different levels affect how we characterize these components?
  3. What would a systems-level characterization of a theory of narrative experience look like in terms of how we specify its components, relations, environment, and evolution? How could we characterize internal dependencies across these units (e.g., multilevel networks)?
  4. Where is most communication research situated currently? Can we organize narrative research into buckets at different timescales?

Where the conversation is headed

Highlight a gap in current research on narratives and generate productive directions for researchers considering narratives at all levels in their work. The group may also develop the discussion into a research article on narratives in communication science.

Suggested reading

  • A handout with definitions will be provided.

Media outlets are regularly reporting and often conjecturing about the role of AI in the workplace. This discussion group will consider the various forms of AI and their uses in organizational contexts. Participants are encouraged to share their research interests and foci of study.

Central question

In what ways can different forms of AI (predictive, perceptual, generative, decisional, optimization, organizational, and robotic AI; Pilny et al., 2025) enhance or disrupt organizational processes and the employee experience?

Conversation starters

  1. What are the advantages and disadvantages of organizations using AI to socialize newcomers?
  2. Will AI make employee performance appraisals more accurate, helpful, and acceptable to employees than their present performance appraisal systems?
  3. What are the potential uses for AI in facilitating successful organizational change?

Where the conversation is headed

Refined directions for research and possible research collaborations.

Suggested reading

  • Liu, Y., Ding, H., Yu, E., & Song, K. (2026). Navigating onboarding in the artificial intelligence (AI) era: How augmentation-based AI usage shapes newcomers’ organizational socialization. Current Psychology, 45(1073), 1–15.
  • Pilny, A., Endacott, C., & Treem, J. (2025). AI in the workplace. Wiley.
  • Xie, W., & Xiao, Q. (2026). “Your first colleague might be an AI”: Bridging functional value and affective trust for theorizing AI avatars in new-employee organizational socialization. International Journal of Human–Computer Interaction, 1–31.

“AI will make our children stupid,” reads one headline. “I’m a child—here’s why I’m friends with my AI chatbot,” reads another. Coverage of children and AI swings between these poles but leans heavily toward the first: children cast as vulnerable, exploited, or missing entirely as agents. This discussion considers what a deficit frame allows communication research to ask, what it makes harder to ask, and what a more expansive, child-centered research agenda could look like.

Central question

What does it mean for communication science to study children and AI primarily through a deficit lens? What questions, methods, and forms of knowledge does that lens make harder to see?

Conversation starters

  1. What do we lose, as a field, by treating risk, harm, or vulnerability as the default variable of interest in research on children and AI?
  2. What methods exist (or do not) for treating children as agents in AI research, rather than primarily as subjects of protection?
  3. Why has a deficit lens been so much easier for the field to build on than one centered on children’s agency?
  4. If you don’t study children directly, has your own field faced a similar pull toward framing a population as vulnerable rather than capable? What made that framing hard to move past?

Where the conversation is headed

A working sketch of a research agenda for studying children and AI, naming concrete questions, methods, or approaches for the field to pursue.

Suggested reading

  • Buckingham, D., & Strandgaard Jensen, H. (2012). Beyond “media panics”: Reconceptualising public debates about children and media. Journal of Children and Media, 6(4), 413–429.
  • Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication, 43(4), 51–58.

In the age of the replication crisis, what is the empirical status of communication theory? Do we have a problem with undead theories? If so, how do we deal with them and how can we do better? Proposed solutions for discussion include taking an abductive approach to theory building, greater attention to variability and distribution shapes, and risky tests in theory testing.

Central question

Is there an empirical crisis for communication theory, and if so, what can be done to mitigate risk and produce empirically adequate theories?

Conversation starters

  1. Do we have undead theories, or is the concern misplaced?
  2. Do theories even need empirical support, and does falsification work in practice?
  3. Care to name names?
  4. Which of the proposed solutions resonate with you?
  5. What solutions should be added to the list?

Generative AI complicates a basic question in doctoral education: what does it mean for research to remain one’s own? Doctoral students may use AI across brainstorming, writing, coding, analysis, and interpretation, but expectations for when that use should be disclosed remain uneven. This discussion will examine where the boundary between individual research practice and shared scholarly responsibility should sit, and how that boundary changes across advising and coauthorship relationships.

Central question

Where should we draw the line between a doctoral student’s private research process and GenAI use that should be disclosed to an advisor or collaborator?

Conversation starters

  1. What should actually determine whether GenAI use requires disclosure: the amount of AI involvement, the intellectual significance of the task, the consequences of error, authorship, or something else?
  2. What do we mean by ‘independent scholarship’ when researchers routinely rely on tools, collaborators, editors, software, and now generative AI? Is AI meaningfully different from those other forms of assistance, and if so, why?
  3. Does an advisor have a legitimate claim to know how a student uses GenAI on independently authored work, or does that cross into unnecessary oversight of the student’s research process?
  4. What changes when the advisor becomes a coauthor? Does authorship create a stronger right to know because responsibility for the integrity of the work is now shared?
  5. Can disclosure be considered genuinely voluntary when the person receiving that disclosure also evaluates the student’s competence, progress, and future opportunities?

Where the conversation is headed

Identify the main factors that should shape GenAI disclosure decisions, clarify where people draw the line between a student’s own research process and shared responsibility with an advisor or collaborator, and surface areas of agreement, differing norms, and questions doctoral programs still need to address.

Suggested reading

  • A one-page table memo, preliminary findings, and selected theory readings will be provided at the session.

In light of a recent push toward grounding communication theory, as well as exploring potential communication laws, we propose selective exposure as a potential candidate. Grounding selective exposure in resource management and limited cognitive capacity makes the resulting heuristic cues obviously adaptive strategies from an evolutionary perspective. Further, selective exposure becomes a metatheory of human interaction beyond the traditional scope of media selection and political communication.

Central question

Is selective exposure a law of communication?

Conversation starters

  1. Is the principle of selective exposure evident in every function of communication?
  2. What are the underlying mechanisms that motivate or restrict (guide) selective exposure?
  3. Is selective exposure evident in dozens of psychological and communication theories, just under different construct labels?
  4. What novel avenues of research are unlocked by this paradigm shift?

Where the conversation is headed

Selective exposure is not merely a media or political communication phenomenon. It is evident across the various functions of communication: relating, informing, influencing, and entertaining. We argue that across all these domains, selective exposure is explained by humans’ limited cognitive-emotional capacity.

Suggested reading

  • Fisher, J. T., Huskey, R., Keene, J. R., & Weber, R. (2018). The limited capacity model of motivated mediated message processing: Looking to the future. Annals of the International Communication Association, 42(4), 291–315.
  • Knobloch-Westerwick, S. (2015). The selective exposure self-and affect-management (SESAM) model: Applications in the realms of race, politics, and health. Communication Research, 42(7), 959–985.
  • Schmälzle, R., & Huskey, R. (2023). Skyhooks, cranes, and the construct dump: A comment on and extension of Boster (2023). Asian Communication Research, 20(2), 84–94.
4:30 - 4:50pm

Closing Remarks
5:00 - 7:50pm

Happy Hour
Saturday evening
Jolly Pumpkin Pizzeria & Brewery
218 Albert Ave., East Lansing

Once the closing remarks wrap, walk over with us to Jolly Pumpkin, about five minutes from the Union. Drinks, pizza, and a chance to keep the day's conversations going.

Sunday, September 20

8:30 - 8:50am

Welcome & Introduction
Breakfast, coffee, and tea will be provided.
9:00 - 10:20am

Panel 3 — Futures in Methods
Stage presentations highlighting new methods for collecting, analyzing, and reporting communication data.

Presentations:

Communication theories describe rich processes that unfold over time, but the methods used to test them often capture a single slice in isolation. This talk proposes high-throughput communication science, an agenda that pairs passively-sensed multimodal data, from phones and wearables to mobile EEG, with Marr's three levels of explanation to study reciprocal causal dynamics as they actually unfold. The goal is a way to triangulate across data sources that matches the complexity of the theory, and a roadmap for the next generation of communication measurement.

Media use rarely happens in a social vacuum, yet interpersonal and mass communication research have mostly stayed in separate lanes. This talk presents a multimodal hyperscanning paradigm that records two people's brain activity at once (using fNIRS) as they watch emotionally rich clips together, talk about them, and watch again. Pilot data suggest that loneliness shapes how neural alignment shifts after conversation, pointing toward a future where co-viewing can be studied as the unfolding, multi-person process it actually is.

People think about stories long after they end. This reflection depends on real-time story processing, yet the neuroscience of shared viewing relies predominantly on short clips. Using EEG data from 95 viewers, we extracted inter-subject correlations from a feature-length film. Shared responses tracked some engagement measures and distinguished between shot and scene types, while later reflection looked more idiosyncratic. This talk maps the utility and potential limits of shared viewing measures in naturalistic feature-length stimuli.

Players often play their own background music while gaming, and this talk uses time-locked psychophysiology (facial EMG, skin conductance, heart rate) to ask how individual player attitudes toward a song influences emotional responses to game moments that matter. Instead of averaging physiology across a whole task, the study zooms in on individual game successes, showing that positive music heightens the emotional response to winning while negative music dampens it. The talk will make a case for the ability to measure emotion at the level of discrete, meaningful events.

Why do some stories captivate audiences while others fall flat? This talk treats narratives as character networks, with characters as nodes and their interactions as edges, and asks how a story's structure shapes its success. Across more than 10,000 novels and 1,000 film screenplays, network features predict real-world popularity and audience ratings. The approach links how people learn from networks to the architecture of stories at scale, and points toward a more computational, design-oriented future for narrative research.

Can large language models replace human respondents when testing messages? Doing so could make message evaluations faster and cheaper, but how far to trust it is still unclear. This talk benchmarks LLM-simulated ratings against 1,045 human evaluators across six models and two prompting strategies, finding the rankings track human judgments closely enough to pre-screen large message pools. However, larger models and fancier prompts do not always help, sharpening the question of how simulation should be designed.

10:30 - 11:50am

Breakout Session 3 - Tool Demos
Seven quick pitches, followed by an open demo lab. Expand a card for details.

Explore the demos (7):

Researchers studying how people communicate (e.g., in negotiations, therapy, or conflict resolution) have no purpose-built tool. Today, academics and industry professionals use surveys, Zoom recordings, or spreadsheets when trying to draw insights from conversation data, if they can obtain such data at all. Attempts to study conversational dynamics often result in slow studies, messy data, and unscalable designs. Dyadic is a fully integrated platform for running conversation studies between humans or between humans and AI.

What you'll see

See how a researcher sets up, runs, and deploys human-human and human-AI conversation studies in Dyadic.

What you'll leave with

Understand how Dyadic supports online conversation studies and where it might fit future research.

Before the session

None.

Links and resources

Human-machine communication (HMC) research often relies on vignettes or recollections rather than live interaction, limiting ecological validity. We introduce DiSCoKit, an open-source toolkit enabling in-survey conversations with LLMs by linking Qualtrics, Microsoft Azure, and a conversation-logging database. Customizable for experimental manipulation, it supports genuine human-machine interaction studies. This demo covers the Qualtrics integration, underlying architecture, limitations/costs/technical requirements, and discussion of customization for diverse research needs.

What you'll see

See participant-side views of a live survey alongside the researcher-side backend, including system prompts, API hooks, Qualtrics random assignment, and examples of other multi-agent implementations.

What you'll leave with

Learn how controlled human-AI interactions can run inside a survey platform and what an institution needs to deploy the toolkit.

Before the session

No setup is needed to attend. Deploying DiSCoKit requires a JavaScript-enabled survey platform such as Qualtrics, Microsoft Azure, and a database for logging conversations.

Links and resources

Ostracism Online provides systematic control of social inclusion/exclusion (like its precursor, Cyberball). But social feedback comes in the forms of giving/receiving likes and dislikes in response to self-disclosures. This opens up new opportunities for examining communicative processes and digital interactions. We’ll (1) tour the open-source tool, (2) see how to edit various experimental parameters, and (3) workshop its potential use cases in interpersonal, CMC, and media research.

What you'll see

Tour the open-source tool, see how to change experimental parameters, and workshop communication-research use cases with the group.

What you'll leave with

Understand what Ostracism Online does, how its parameters can be manipulated, and how the paradigm could support new communication questions.

Before the session

A laptop with web access and a GitHub account may be helpful, but neither is required.

This live demonstration provides a start-to-finish walkthrough of Inoxity, an open-source platform for high-throughput communication science and configurable iOS research studies. Using an example study, attendees will see how researchers build and manage a study through the dashboard, locate supporting documentation, enroll participants through the iOS app, and access study data. The demonstration will conclude with a behind-the-scenes look at the control and study backends that support configuration, routing, and data storage.

What you'll see

Follow an example study from configuration in the researcher dashboard through participant enrollment in the iOS app, data access, and a behind-the-scenes look at the control and study backends.

What you'll leave with

Understand the complete Inoxity workflow and evaluate how the platform could be customized for a research protocol.

Before the session

None. Just come curious!

TikTok offers valuable opportunities to examine public opinion and emerging digital and social norms in communication, but collecting and analyzing data in low-resource languages can be challenging. Combining Zeeschuimer, large language models, and R, researchers can capture, preprocess, and analyze TikTok posts and comments. This workflow also supports reproducibility and replicability by documenting analytical steps, preserving code, and enabling researchers to apply the same procedures to new datasets and linguistic contexts.

What you'll see

Walk through collecting TikTok posts and comments, from query design and web-based capture to preprocessing in R and analysis.

What you'll leave with

Learn how to design a TikTok dataset, capture public data, apply the supplied R Markdown workflow, and begin an analysis.

Before the session

Anyone who wants to follow along should install Firefox and RStudio and create a dedicated TikTok account rather than use a personal account. A research account with a preferred generative-AI platform may also be useful. No prior R knowledge is required.

Links and resources

Large language models can speed up systematic reviews, but only if their decisions can be inspected, challenged, and reproduced. This hands-on demo uses a 500-plus-article review of technology-delivered haptics to show a human-in-the-loop workflow built in Cursor: retrieving full texts from Zotero, translating a codebook into structured prompts, processing articles in documented batches, exporting decisions for human review, and refining the workflow when human and model judgments disagree.

What you'll see

Follow one article from Zotero retrieval and PDF extraction through two-stage screening, structured model output, CSV export, human audit, and prompt refinement.

What you'll leave with

Leave with a concrete codebook-to-audit workflow and a practical way to pilot, validate, and document LLM-assisted screening.

Before the session

None. A laptop is optional. Cursor, Python, Zotero, and an API account are not required.

The proliferation of online-based paid research, coupled with advances in software development employing large language models and artificial intelligence, have made it increasingly difficult to verify that research participants are actually human. We introduce a system that uses standard webcams to capture and analyze, in real time, multiple physiological signals that serve as robust markers of genuine human presence.

What you'll see

See how a simple bot gets through common bot and attention checks, then how webcam-based psychophysiological capture can help verify that a participant is human.

What you'll leave with

Learn a new approach to preventing and detecting research-participant fraud.

Before the session

None.

Links and resources

12:00 - 1:20pm

Lunch & Poster Session 2

Posters (19):

Making Harm Harder to Ignore: Testing Moral Cues in Strategic Prosocial Persuasion
Tse-hsi Chien University of Florida
The Future of Worship?: Online Worship and Loneliness in Religious Americans, Mediated by Self-Disclosure
Brooklyn Monroe Chillemi Wake Forest University
AI Agency Framing and Task Interdependence
Soumyajit De University of California, Santa Barbara
Negativity and High Arousal Sustain Online News Reading
Xuanjun Gong Texas A&M University
Humans Behind the AI: Trust and Ethics in Mental Health Chatbots
Jinxu Li University of Minnesota
Frames We Can(not) See: How Deficit Thinking Shapes the Communication Science of AI and Children
Zhixin Li Syracuse University
Well-being Moderates Emotional Responses to Positive Daily Events and Positive News
Yifei Lu University of Pennsylvania
Modifying the Decision Structure of Screen-Time Self-Nudges: From Stop to Action
Elisa Ragone Purdue University
Do LLM-Agent Groups Exhibit a Hidden-Profile Effect?
Torsten Reimer Purdue University
Media Literacy, Media Self-Efficacy, and News Credibility in Algorithmically Mediated Social Media Environments: A Study of University Students
Kyrmyzy Turebayeva Michigan State University
Mapping the News-Finds-Me Construct: Evidence From a Nomological Network Analysis
Sylo de Vegvar University of California, Santa Barbara
Fight and Flight? How Psychological Reactance and Learned Helplessness Shape Privacy Decision-Making
Laurent Wang University of Texas at Austin
Intellectual humility and learning outcomes in online health-information-seeking
Xinyi Wang University of Pennsylvania
Understanding the Relationship between Technology Use and Well-Being in Older Adults
Joshua Weinzapfel Purdue University
The Tug-of-War between Fear of Missing Out (FoMO) and Disconnection Motivations on Mental Stress
Fan "Ellie" Yang Illinois State University
Understanding Human–AI Emotional Support: The Influence of Chatbot Design on Anthropomorphism and Mind Perception
Lucas J. Youngvorst Pennsylvania State University
Mapping Methods in Online Hate Research: A Review, Comparison, and Future Directions
Yidi Zhang UC Santa Barbara
Not Just Old Wine in New Bottles: Extending CMC Theory to AI-Mediated Communication
Rui (Cara) Zhu Michigan State University
Beyond Message Effects: The Impact of Choice Set Composition on Perceptions of Science and Public Opinion
Yijia (Erika) Zhu University of Wisconsin–Madison
1:30 - 4:50pm

Follow-Up Discussions & Unstructured Time
Time reserved for attendees to continue discussions, meet with collaborators, or explore campus and the surrounding area.