HCL+PrioR mtg. Jimin Nam. Monday 10/13 1:30-2:30pm ET

[Antonio Alonso Arechar]

Good morning everyone,

Jimin Nam will present next week (10/13 1:30-2:30pm ET; see blurb below).
Lab meeting will be hybrid, with a zoom link (https://mit.zoom.us/j/91655448125) and in person in the conference room on the 15th floor of E94 (1579; 245 First Street).

Best regards,

Antonio

---

Title: LLMs as Scalable Tools for Interactive Consumer Behavior Experiments: Comparing Persuasion Strategy Effectiveness

Authors: Jimin Nam, Reed Orchinik, and David Rand

Abstract: Recent advances in large language models (LLMs) present an opportunity to address methodological limitations in traditional online experimental approaches. Here, we demonstrate the value of LLMs as experimental tools for consumer behavior research using a case study on persuasion. We present two studies in which LLMs engage participants in back-and-forth conversations with the goal of shifting consumption-related attitudes. Study 1 examines which strategies are naturally employed by the LLM, and the effectiveness of these strategies, when persuading participants about smart- phones (iPhone versus Android). Study 2 experimentally manipulates which persuasive strategy the LLM uses when persuading about meat consumption or online shopping (Amazon versus Walmart.com). Our results suggest that conversations with LLMs can effectively shift both self-reported attitudes and actual behavioral choices, with factual and informational appeals proving particularly effective across outcomes. Conversely, overtly manipulative strategies–including social consensus appeals and bias framing– were less effective. Moderation analysis reveals that pre-existing trust in AI technology significantly enhances persuasive effectiveness, while mediation analysis identifies perceived AI manipulation and expressed skepticism as key psychological mechanisms underlying consumer resistance. Beyond these substantive findings, our research the value of LLMs in experiments allowing for the efficient testing of multiple theoretical predictions within single experimental frameworks and bridges observational and experimental paradigms through automated content analysis.

[Drazen Prelec]

Hi all, 

fyi, I had shown slides from Sebastian’s lab in my talk at neuroecon last week.

https://www.mos.org/events/subspace/mapping-human-brain-eyewire-founder-dr-sebastian-seung

Drazen