* Graduate student coauthor, #Undergraduate student coauthor
Under review
1. A Statistical Model of Bipartite Networks: Application to Cosponsorship in the United States Senate
Adeline Lo, Santiago Olivella & Kosuke Imai. Revise & Resubmit, American Journal of Political Science. arxiv link
Many networks in political and social research are bipartite, with edges connecting exclusively across two distinct types of nodes. A common example includes cosponsorship networks, in which legislators are connected indirectly through the bills they support. Yet most existing network models are designed for unipartite networks, where edges can arise between any pair of nodes. We show that using a unipartite network model to analyze bipartite networks, as often done in practice, can result in aggregation bias. To address this methodological problem, we develop a statistical model of bipartite networks by extending the popular mixed-membership stochastic blockmodel. Our model allows researchers to identify the groups of nodes, within each node type, that share common patterns of edge formation. The model also incorporates both node and dyad-level covariates as the predictors of the edge formation patterns. We develop an efficient computational algorithm for fitting the model, and apply it to cosponsorship data from the United States Senate. We show that senators tapped into communities defined by party lines and seniority when forming cosponsorships on bills, while the pattern of cosponsorships depends on the timing and substance of legislations. We also find evidence for norms of reciprocity, and uncover the substantial role played by policy expertise in the formation of cosponsorships between senators and legislation. An open-source software package is available for implementing the proposed methodology.
2022 APSA Political Networks Section Honorable Mention for Best Conference Paper Award.
Paired R package: NetMix.
2. Covariate Screening in High Dimensional Data: An Application to Congressional Text
Revise & Resubmit, Political Science Research & Methods.
High dimensional (HD) data, where the number of covariates or meaningful covariate interactions might exceed the number of observations, is increasingly used in social science prediction. An important question for researchers is how to select the most predictive covariates among all available ones. Common approaches use ad hoc rules to remove noise covariates, or select covariates through the criterion of statistical significance or by using machine learning techniques. These can suffer from lack of objectivity, choosing some but not all predictive covariates, and failing reasonable standards of consistency that are expected to hold in most HD social science data. The prediction literature is scarce in statistics used to directly evaluate covariate predictivity. I address these issues by proposing a variable screening step prior to traditional statistical modeling, whereby covariates are screened for their predictivity. I propose the influence (I) statistic to evaluate covariates in the screening stage, showing that the statistic is directly related to predictivity and can help screen out noisy covariates and discover meaningful interactions. I illustrate how this screening approach can removing noisy phrases from U.S. Congressional speeches and rank important ones to predict partisanship.
3. Can Praise from Peers Promote Empathy and Inclusive Behavior towards Racial or Ethnic Outgroups?
Adeline Lo, Jonathan Renshon & Lotem Bassan-Nygate*. Revise & Resubmit, Nature Human Behavior. Working paper here.
Outgroup bias is a well-documented and pernicious phenomenon, manifesting in negative attitudes and behavior towards outgroups. Empathy—taking the perspective and understanding the experiences of others—holds considerable promise for attenuating outgroup bias. Yet, engaging in empathy is costly and existing interventions to encourage it are expensive and difficult to scale. Through six pilots, we develop a non-invasive, low-cost, peer praise intervention that encourages empathetic behavior towards generalized “others” by stimulating positive emotions. This research tests the hypothesis that our peer praise intervention promotes empathetic behavior among white respondents in the U.S. towards black and Latino/a Americans, a context where racial/ethnic outgroup bias is particularly durable and pernicious. We (1) measure real choices to engage in empathy with outgroups (2) test whether effects of peer praise are durable using a panel design (3) explore downstream effects on attitudinal/behavioural support for historical civil rights and advocacy groups (UnidosUS, BLM).
PolMeth 2021 Annual Summer Meeting Best Faculty Poster Award.
4. When Hearts Meet Minds: Complementary Effects of Perspective-Getting and Information on Refugee Inclusion
Claire Adida, Adeline Lo, Melina Platas, Lauren Prather & Scott Williamson. Conditionally Accepted, Political Science Research & Methods. Working paper here.
Perspective-getting and correcting misconceptions are two common interventions to promote inclusion toward outgroups. However, each strategy has its limitations. Empirical work on the role of correcting information yields inconclusive results, and empathy-based interventions may reproduce the biases they are meant to alleviate. We clarify the strengths and weaknesses of each strategy, and offer a design to identify the conditions under which they are most effective. Using three studies on refugee inclusion with nearly 15,000 Americans over three years, we find that information and perspective-getting affect different outcomes. Perspective-getting affects warmth, policy preferences, and behavior, while information leads to factual updating only. We show that combining both interventions produces an additive effect on all outcomes, but neither strategy enhances the other. Bundling the two strategies helps guard against potential backfire effects of information, however. Our results underscore the promise and limits of information and perspective-getting for promoting inclusion, highlighting the benefits of integrating the two strategies.
5. How Increasing Refugee Visibility on TV News Causes Viewers to Support Refugees More, but Like Them Less
Adeline Lo and Oliver Lang*. Reject & Resubmit, Journal of Politics. Working paper available upon request.
As global refugee flows accelerate, so does local news coverage on the subject and the potential for major political consequences. In an analysis of all broadcasts of the most famous television news program in Germany 2013-2019, we first show observationally (through text and image analysis) that refugee coverage increases with immigration and is correlated in different ways with public opinion about refugees. Conditioning on these patterns, we then implement a nationally representative block randomized media experiment. We find that TV news coverage of refugees causes viewers to be more willing to donate money to pro-refugee organizations, but surprisingly to feel colder and more socially distant towards them. We discuss the far reaching consequences of this divergent pattern for the future of local politics in Germany and potentially around the world.
Work in progress
Political Methodology
Methods for Sequence Analysis
with Héctor Pifarré i Arolas & Keith Levin.
Causal Network Mediation in IR
with Yehzee Ryoo*, Keith Levin, and Alex Hayes.*
Intergroup Empathy
Co-Identity Peer Praise and Empathy
with Jonathan Renshon & Lotem Bassan-Nygate*.
The Costs of Empathy
with Jonathan Renshon & Lotem Bassan-Nygate*.
Dispositional and Activated Empathy
with Claire Adida, Melina Platas, Lauren Prather & Scott Williamson.