My research is dedicated to understanding and addressing the countless complex systems and institutions whose operations serve to bolster the status quo and propagate inequality. My research integrates multiple levels of analysis and diverse methodologies to understand how individuals are shaped by the systems they inhabit, and, in turn, how individuals’ behaviors contribute to these systems. This work aims to assess these dynamic reciprocal relations through a multidisciplinary lens, to the ultimate goal of addressing critical social issues like anti-minority bias and climate change. The following is a more in-depth dive into my research interests and current projects.

How do system-level processes impact individual-level cognition and behavior? 

Examining one aspect of the dynamic relations between individuals and systems, I explore the ways in which system-level processes impact individuals’ perceptions and representations in diverse contexts. One line of work in this thread tests whether economic scarcity directly impacts anti-minority stereotyping (Berkebile-Weinberg et al., 2022, JESP). In this work, we found that for non-Black American participants, the perception of economic scarcity increased endorsement of negative Black American stereotypes. In Study 1, experimentally induced scarcity heightened reported knowledge of stereotypes of Black Americans as low in socioeconomic status and as threatening, relative to control. In subsequent studies, perceived scarcity increased the stereotypicality of participants’ visualizations of a Black male face, as assessed using a reverse correlation procedure and judged by independent raters. Together, these studies shed light on the role of stereotyping as a means to justify racial discrimination and maintain power structures in response to system-level threat. I am currently following up on this work by testing the societal-level effects of systemic instability on anti-minority dehumanization. In this package we used large-scale natural language processing (NLP) to assess shifts in collective representations during the COVID-19 recession. Using a word embeddings approach with a large U.S. news media text corpus (roughly 1.3 billion words across 2.2 million articles), we found that the conceptual similarity (operationalized as cosine similarity) between terms describing humans and terms for Black people decreased relative to the conceptual similarity between human terms and terms for White people, as a function of COVID-19 progression. The second study in this line revealed that this process plays out in explicit self-reports in a diverse context as well – White Dutch participants reported that Moroccan Dutch people were less capable of uniquely human attributes compared to White Dutch people. Crucially, this effect was strongest for those higher in social dominance orientation, suggesting that system-level motives play a role in these individual-level representations. This work reveals that real-world systemic instability elicits the denial of racial and ethnic minorities’ uniquely human attributes, and supports the larger theoretical assertion that system-level motives impact individual-level cognition, to the aim of bolstering the status quo.

How are collective sentiments represented and reproduced? 

This investigation into the dynamic nature of collective representations inspired me to pursue work on the complex ways in which attitudes and sentiments are formed and propagate on the system level. To extend our understanding of these processes in a new direction, my recent work assesses the extent to which online search algorithms both learn and shape people’s perceptions and policy attitudes in the context of climate change (Berkebile-Weinberg et al., under review). This project examines the extent to which collective climate change concern manifests in internet search algorithm output, and subsequently whether such manifestations impact individuals’ climate-related beliefs downstream. The first study of this package tests this process with the ubiquitous online search engine Google Images. In a sample of 58 countries, we found that nation-level climate change concern predicted the emotion and need for climate action conveyed by climate change Google Image search results within a nation, as rated by a naive sample. In a follow-up experiment, we found that participants exposed to climate change search results from countries with high collective climate concern reported higher climate change concern, climate policy support, and willingness to engage in climate action, compared to those who viewed search results from countries with low collective climate concern. Together, this work provides insight into how these search algorithms learn and represent collective values, and how such representations can further propagate and influence policy attitudes downstream. This research highlights the central role that AI plays in informing the public about crucial social issues and provides an important access point for system-level interventions aimed at effectively informing the public about climate change. More broadly, this work exemplifies how dynamic systems between individuals and the technologies in our environments lead to emergent, collective phenomena.

How does institutional bias develop and propagate? 

My theoretical work bridges my empirical research with a diverse array of scholarship across numerous fields of study (e.g., psychology, AI, human computer interaction, cognitive science). This work offers a novel account of systemic bias as a fundamentally collective phenomenon that emerges from the distribution of cognitive effort between individuals and their environments, over time (Berkebile-Weinberg & Vlasceanu, under review). Although prejudice is understood to operate at both the individual and system level, little is known about how prejudice transcends from the individual to dictate the behaviors of institutions and systems. The existing work in this space fails to explain how such a process propagates over time, let alone how such patterns thrive despite cultural shifts toward egalitarian values. My approach applies the theory of distributed cognition (Hutchins, 1995) to understand systemic bias as the emergent product of the interactions of individuals with each other and with technologies in their environments, rather than a simple aggregate of individual cognition. This theoretical analysis shows that individuals in systems interact with each other, and with specific social and material tools in their environments that serve to transform the cognitive processes involved in discrimination, to the end of propagating institutional bias. This work proposes distributed cognition as a new foundation for the propagation of institutional bias, and outlines several concrete avenues for future research.