Salons & Methods

Salon Speakers

Informal events based on the concept of 16th C Italian and French Salons – a gathering to increase knowledge though conversation – hosted by topic experts. Come along to ask your burning questions, sharpen your knowledge, or simply enjoy lively discussion!

Ask Me Anything

Saturday, March 2

Industry Salon

Saturday, March 2

Communicating Science: Best Practices

Sunday, March 3

Where do scientific ideas come from anyway?

Sunday, March 3

Student Salons

Mastering Data Visualization: Exporting and Graphing Results with Prism

Saturday, March 2

Victoria Hart-Derrick

Participants will explore the benefits of utilizing Prism as a powerful tool for graphing and visualizing results, enabling them to effectively communicate their findings. From exporting data seamlessly to choosing the right graph types, this salon will provide hands-on experience and practical tips for optimizing the presentation of your research outcomes.

Navigating Qualitative Data – Benefits, Challenges, and Considerations

Sunday, March 3

Amy Gregory

Join us to explore the richness of qualitative data, and gain a deeper understanding of its unique advantages in uncovering nuanced insights. Participants will engage in discussions on the challenges inherent in qualitative analysis, such as subjectivity and potential biases, and strategies to address them.

Methods

Saturday, March 2nd

Sean Dae Houlihan,MIT, Dartmouth

Bayesian Theory of Mind Models

Moderator: TBA

Abstract: This workshop introduces a probabilistic approach to building models of people’s intuitive theories of emotion. We frame human emotion understanding as approximately rational inference over a causally-structured mental model of other minds. We then see how probabilistic programs can be used to formalize, test, and learn scientific theories of emotion understanding.

Stephanie Marita Carpenter, Arizona State University

Measuring Real-Time Emotions and their Relationship to Behavior

Moderator: Nicole Roberts, Arizona State University


Abstract: To better understand the complexity of emotion dynamics, it is critical to assess emotions both in traditional lab-based settings and out “in the wild,” in real-time, real-world contexts. This talk will explore the use and integration of multiple methods to assess and answer scientific questions related to emotion dynamics through the measurement of emotional processes in real-life settings. Specifically, we will explore the promise of simultaneously utilizing wearable technologies, ecological momentary assessment (EMA), and field experiments to address how emotions vary over time, as well as how to develop just-in-time adaptive interventions that modulate affective reactions in real-time, real-world settings to impact behavior change.

Robert Hawkins, University of Wisconsin – Madison

Using Causal Interventions to Probe Commonsense Affective Understanding in Neural Language Models



Moderator: Joao Guassi Moreira, University of California, Los Angeles

Abstract: Recent advances in machine learning have produced large language models (LLMs) like ChatGPT which increasingly display human-like behavior on commonsense tasks involving emotions, mental states, and social situations. However, these models are black boxes, limiting their relevance for psychology; it is unclear whether the representations of affective and social schemas underlying LLM task performance resemble those used by humans. In this talk, I’ll provide an overview of causal probing techniques adapted from psychology and neuroscience, which can be used to “peer inside” these models and gain some insight into their underlying computational processes. By surgically manipulating the activations at different layers of these models, we are able to identify distinct sub-circuits implicated in processing different kinds of information implicated in social reasoning. I will discuss current best practices for social scientists interested in critically evaluating claims about affective representations in neural networks, or interested in using mechanistic analyses of neural networks to inspire new hypotheses about affective cognition.

Sunday, March 3rd

Sophie Wohltjen, University of Wisconsin – Madison

Capturing and Assessing Dyadic Interactions

Moderator: Elizabeth da Silva, Indiana University-Purdue University Columbus

Abstract: Social psychological research benefits from the use of paradigms that situate group behaviors within their natural ecological niche—social interaction. However, conducting research on social interaction can be difficult, requiring different design considerations than more traditional empirical methods. In this workshop, I discuss and advise on things I would have liked to know when designing my first dyadic interaction study—specifically, 1) when and why dyadic interaction is useful in social research, 2) how to achieve experimental control in naturalistic paradigms, and 3) what to consider when analyzing data from two participants instead of one. In this workshop I will also point to softwares that I have found useful in my analysis of dyadic data, and I will include recent, published examples of excellent dyadic research from a range of scientists.

Thank you to our Sponsors & Exhibitors