Anxiety is a pervasive challenge among undergraduate students, particularly during high-stress academic periods such as midterm examinations. This study examined the effects of two cognitive reappraisal techniques - Construal and Repurpose - on undergraduate anxiety levels. Additionally, the role of peer social influence in the transmission of emotion regulation strategies was investigated through the manipulation of the proportion of trained individuals within each group. Thirty-two Princeton University Undergraduates were assigned to reappraisal conditions (Construal or Repurpose) or a Control group, and further divided into High (3 trained) and Low (2 trained) proportion conditions based on the number of trained peers within each discussion group. Participants completed pre-, post-, and follow-up questionnaires assessing anxiety levels, predicted and observed grades, stress management ability, perceived grade importance, group connectedness, and discussion usefulness across three timepoints: before, immediately after, and two weeks following the group discussion session conducted over Zoom. Despite null ANOVA findings, descriptive trends indicated that Construal conditions demonstrated the most consistent reductions in anxiety across timepoints compared to Repurpose conditions. Findings suggest that intuitive, accessible reappraisal strategies, such as Construal may be more effectively transmitted through peer social influence and more beneficial in fostering group cohesion alongside emotion regulation adoption. Future research engaging with larger samples, longer follow-up periods, and more direct measures of social influence is needed to build on these preliminary findings.

Interdisciplinary teams are increasingly tasked with solving complex, ill-defined problems, yet the real-time processes underlying their success remain understudied behaviorally. This thesis develops and pilots a paradigm for observing interdisciplinary team process on a bounded wicked problem, combining a novel sustainability task (Project CLEARPATH), a discussion quality rubric, and a transcript-derived analytical pipeline. Seven epistemically diverse triads of Princeton undergraduates were assigned to one of four conditions crossing role assignment and discussion scaffolding. The rubric achieved good composite reliability, and the task elicited genuinely distinct epistemic approaches across teams. Exploratory analyses revealed that discussion quality only weakly predicted written output quality, that the two interventions appear to operate at different phases of the collaborative cycle, and that teams self-organized rapidly into qualitatively distinct interaction profiles. The paradigm is offered as reusable infrastructure for confirmatory research on interdisciplinary team process under uncertainty.

Much of the previous research on social activism has focused on large scale collective action and the traits of individual actors. Yet, emotions including hope and fear are understood to serve as important motivating forces on our behavior. In this investigation, I expand upon our understanding of hope and fear motivation to ask how the emotional qualities of the language used to present a social issue might impact an individual's desire to support that cause. Exposing participants to either hope-inspiring language, fear-inducing language, or a neutral control, changes in perceived message strength as well as several behavioral measures (including hypothetical action, time allocation, and token donation tasks) are assessed as indicators of social activist behavior. Between treatment groups, no statistically significant differences were found to indicate that one particular emotional frame (or even the absence of emotional framing) was more effective in increasing perceived issue importance or solution-oriented behavior. Given that participants were able to detect the relevant emotional frame in a manipulation verification check, I hypothesize that this difference is the result of short-term exposure and stimuli format — not necessarily a failure of emotional presentation to influence persuasion or solution-oriented behavior as a whole.

In recent years, the climate for political discourse in the United States has become increasingly heated, with the 2024 presidential election serving as an apparent inflection point. Concurrently, developments in machine learning have enabled large language models (LLMs) to become progressively more human-like, allowing them to take on the role of a social companion. Previous research has suggested that AI chatbots can be successful persuaders in political contexts due to their wealth of factual knowledge and their objective nature. The present study looks to 1) further evaluate AI chatbots as an effective agent in reducing polarized attitudes on political topics, 2) compare their effectiveness to other humans, and 3) to pursue an ideal political identity for the chatbot. College students from Princeton University (N = 42) participated in an 8-round debate on two of their strongest political viewpoints with an AI chatbot (powered by GPT-4o), in which they were informed they were either chatting with the AI or a neighboring human participant through an online message room. Chatbots were prompted to either personify an extreme or moderate opposition to that of the participant. Results from the debates revealed that the chatbot was successful in depolarizing attitudes across all conditions, though differences between conditions were all statistically insignificant. As a follow-up, this paper also includes a textual analysis of the dialogues produced, building on prior research that suggests linguistic styles between humans and AI tend to converge over the course of their conversations.