Aaron Shaw

Jessica Hullman, David Broska, Huaman Sun et al.

A growing literature uses large language models (LLMs) as synthetic participants to generate cost-effective and nearly instantaneous responses in social science experiments. However, there is limited guidance on when such simulations support valid inference about human behavior. We contrast two strategies for obtaining valid estimates of causal effects and clarify the assumptions under which each is suitable for exploratory versus confirmatory research. Heuristic approaches seek to establish that simulated and observed human behavior are interchangeable through prompt engineering, model fine-tuning, and other repair strategies designed to reduce LLM-induced inaccuracies. While useful for many exploratory tasks, heuristic approaches lack the formal statistical guarantees typically required for confirmatory research. In contrast, statistical calibration combines auxiliary human data with statistical adjustments to account for discrepancies between observed and simulated responses. Under explicit assumptions, statistical calibration preserves validity and provides more precise estimates of causal effects at lower cost than experiments that rely solely on human participants. Yet the potential of both approaches depends on how well LLMs approximate the relevant populations. We consider what opportunities are overlooked when researchers focus myopically on substituting LLMs for human participants in a study.

Joon Sung Park, Carolyn Q. Zou, Aaron Shaw et al.

The promise of human behavioral simulation--general-purpose computational agents that replicate human behavior across domains--could enable broad applications in policymaking and social science. We present a novel agent architecture that simulates the attitudes and behaviors of 1,052 real individuals--applying large language models to qualitative interviews about their lives, then measuring how well these agents replicate the attitudes and behaviors of the individuals that they represent. The generative agents replicate participants' responses on the General Social Survey 85% as accurately as participants replicate their own answers two weeks later, and perform comparably in predicting personality traits and outcomes in experimental replications. Our architecture reduces accuracy biases across racial and ideological groups compared to agents given demographic descriptions. This work provides a foundation for new tools that can help investigate individual and collective behavior.

Marlon Twyman, Brian C. Keegan, Aaron Shaw

Social movements use social computing systems to complement offline mobilizations, but prior literature has focused almost exclusively on movement actors' use of social media. In this paper, we analyze participation and attention to topics connected with the Black Lives Matter movement in the English language version of Wikipedia between 2014 and 2016. Our results point to the use of Wikipedia to (1) intensively document and connect historical and contemporary events, (2) collaboratively migrate activity to support coverage of new events, and (3) dynamically re-appraise pre-existing knowledge in the aftermath of new events. These findings reveal patterns of behavior that complement theories of collective memory and collective action and help explain how social computing systems can encode and retrieve knowledge about social movements as they unfold.

Haoqi Zhang, Andes Monroy-Hernandez, Aaron Shaw et al.

Many celebrate the Internet's ability to connect individuals and facilitate collective action toward a common goal. While numerous systems have been designed to support particular aspects of collective action, few systems support participatory, end-to-end collective action in which a crowd or community identifies opportunities, formulates goals, brainstorms ideas, develops plans, mobilizes, and takes action. To explore the possibilities and barriers in supporting such interactions, we have developed WeDo, a system aimed at promoting simple forms of participatory, end-to-end collective action. Pilot deployments of WeDo illustrate that sociotechnical systems can support automated transitions through different phases of end-to-end collective action, but that challenges, such as the elicitation of leadership and the accommodation of existing group norms, remain.