Working papers, conference proceedings, and published articles can be found in the paper archive below.
I describe a model of strategic, decentralized and asynchronous communication in policy-making networks. Two central focuses of the model are the actors' awareness of who other actors will talk to in the future and the sequential ordering of actors' communications. I derive conditions for truthful ``cheap-talk'' communication within sequential communication networks and show that (1) the ordering of individuals within the network can matter above and beyond individuals' policy preferences and degree of decision-making authority, (2) sequential communication throughout can engender credible communication in situations in which private, dyadic communication will not, and (3) sequential communication can sometimes undermine credible communication, so that exclusion of one or more ``extreme'' (or extremely powerful) individuals from the communication network can be (Pareto) optimal. Finally, the analysis and results suggest that it is theoretically impossible to cleanly hive off homophily from the study of strategic information transmission in networks.
"Media events" such as political debates generate conditions of shared
attention as many users simultaneously tune in with the dual screens
of broadcast and social media to view and participate. Are collective
patterns of user behavior under conditions of shared attention
distinct from other "bursts" of activity like breaking news events?
Using data from a population of approximately 200,000
politically-active Twitter users, we compare features of their
behavior during eight major events during the 2012 U.S. presidential
election to examine (1) the impact of "media events" have on patterns
of social media use compared to "typical" time and (2) whether changes
during media events are attributable to changes in behavior across the
entire population or an artifact of changes in elite users’ behavior.
Our findings suggest that while this population became more active
during media events, this additional activity reflects concentrated
attention to a handful of users, hashtags, and tweets. Our work is the
first study on distinguishing patterns of large-scale social behavior
under condition of uncertainty and shared attention, suggesting new
ways of mining information from social media to support collective
sensemaking following major events.
Scholars of lobbying have been limited in their ability to measure organizational lobbying agendas, positions and coalitions. They are either forced to rely on time-consuming interviews or overly-broad Lobbying Disclosure Act-mandated issue codes. We propose a new approach. We use a hierarchical agglomerative clustering (HAC) algorithm to group bills within LDA issue codes based on their similarity, calculated using latent sematic analysis of a corpus we constructed from summary text provided by the Congressional Research Service (CRS). This technique allows us to disaggregate within the existing categories and label individual bills with higher resolution than was previously possible. We then use the clustering groups to label bills within a weighted affiliation network based on the volume of lobbying by a given industry on a given bill. The topic labels of bills within the network provide more detailed insight into the specific policies or provisions different industries have supported, and might be likely to support in the future. As a test, we apply this approach to lobbying on immigration legislation during the 109th – 112th Congresses
Individuals who are exposed to conversations about politics are more politically active. Analytical biases make it difficult to show evidence of a causal relationship between discussion and participation. It is also uncertain how long the relationship between discussion and participation lasts. Here both questions are addressed with panel data collected from college undergraduates who were randomly assigned to their dormitories. Random assignment to social context and measurement of behavior over time allows for more precise measurement of the relationship between discussion and participation. The data show that discussion is associated with higher levels of participation, immediately and years into the future. This relationship is more consistent over time in the case of conversations between roommates compared to conversations within the wider context of the dormitory. The initial increase in participation associated with discussion is a mechanism underlying the long-run relationship between discussion and participation. These results highlight the importance of accounting for social influences on political participation.
The Supreme Court’s importance as an institution stems largely from the precedents it generates. Classifying precedents as influential or authoritative requires a counterfactual model of precedent not yet seen in the literature. If a precedent influences the decision- making of future courts, then that precedent must alter how future courts would have ruled if the precedent did not exist. If behavior is unaltered, then citations to precedent are merely a legitimating device, and have no genuine authority in the legal system. Addressing this problem first requires a precise theoretical definition of precedent, a consideration of vari- ous ways of measuring precedential influence, and a summation of the tradeoffs associated with various approaches, provided in this paper. I argue that measuring precedent using citation counts measures not influence, but notoriety, and I discuss how efforts to measure precedent may end with the discovery that if truly influential precedents exist, their effect may be impossible to identify.
Recent debates in the US Congress over major policy issues, such as the US debt ceiling, the use of the filibuster in the Senate, and health care reform, have witnessed the emergence of small groups of legislators -- given names like “The Gang of Six” in popular press -- working to craft a bill that (they may expect) covers the middle-ground between opposing factions. Given the usual expectations that, 1) committee members are not preference outliers, and 2) committees have better policy expertise than the average chamber member, what purpose do these small groups serve? The argument here is that these gangs represent a focal point for accusations of ideological compromise and potential blame (if the product does not proceed to a floor vote). As partisanship in Congress has increased, the cost of compromise has increased, which may make the otherwise jurisdictionally-appropriate committee members less inclined to allow bills that would appeal to moderate voters to progress. Allowing other legislators to so visibly drive the work on moderate bills deflects the extremists from accusations of being “soft.” At the same time the heightened attention on the gang raises the reputational costs of failure. We should expect, then, to see gang membership to be comprised of more ideologically moderate members, who have served longer or who have won their seats by a wide margin (making them better able to absorb the reputational cost of failure). We review the (small) number of cases of emergent gangs to examine the model's comportment with observed behavior.
Congress has been increasingly criticized as a broken, gridlocked, polarized, inef- fective institution. In this paper we seek to explore the consequences of polarization and whether legislators take steps to alleviate them. We hypothesize that participa- tion in the voluntary, bipartisan, caucus system provides opportunities for legislators to build cross-partisan relationships and pro��t from shared information, which can alle- viate some of the negative e��ects of polarization. We operationalize polarization using dyadic covoting and show that legislators are more likely to covote if they share more caucus connections, controlling for a variety of factors that predict voting. The data in this analysis spans 9 congresses (1993-2010) and includes multiple connections between legislators.
This paper proposes that direct service networks providing food security and homeless services can operate most effectively when agencies that transcend service clusters are identified as primary points for new policy initiatives and funds disbursement. The analysis recognizes the correlation between food security, specifically the use of food pantries, and first time homelessness. First time homelessness is part of a reinforcing dynamic system that can severely limit the real opportunities individuals have for growth and self-determination.
This paper examines food security and housing security 501(c)3 classified agencies in Albany, New York. The analysis is prefaced by an extensive literature review of Nobel Laureate Amartya Sen's "Capability Approach" which examines development through a lens defined by the actual opportunities individuals have to define and drive their own political, social, and economic destinies. The paper examines the environment that constrains person centered development, focusing on the impact of hunger and homelessness on an individual’s opportunities within society.
The analysis includes a review of principal-agent theory which supports the theory that 501(c)3 agencies are particularly suited to basic needs, person-centered services, especially in the fields of food security and homeless services. The efficacy and flexibility of 501(c3) direct service agencies will be directly correlated to the collaboration, information sharing, and referral practices of the individual agencies.
The network data will be gathered through the distribution of surveys to relevant agencies in the Albany area that provide direct service to food insecure and homeless individuals. Due to the size of the sample set the results will be dichotomized to form a complete network map across all relations. The network data will be run through the UCINET 6 routines to determine clique and cluster densities within and across service areas. The data will also be submitted to the BLOCKS program within StOCNET to form a stochastic block-model that identifies agencies that transcend service areas to form "umbrella agencies" that are capable of organizing and initiating programs across service areas to prevent first time homelessness by empowering individuals.
The identification of "umbrella" agencies for funding and policy initiatives from is important to polity and funding organizations that seek to prevent the trauma and lifetime implications of first time homelessness on individuals, communities, and the nation at large. This specification is especially important given the shift to austerity in public funding domestically and abroad.
The fact that voters often fall far short of democratic ideals raises important questions about how stable, representative government occurs. One potential an- swer is that errors in individual attitudes cancel out in the aggregate, thereby mak- ing public opinion appear rational while allowing for its effective communication; another comes from the study of groups and juries. In this paper we address both of these possibilities, exploring the consequences of social networks for processes of aggregation. If social networks are organized in ways that facilitate the exchange of informed opinions, it is possible that they lead to aggregate opinions that are even more informed and rational than simple aggregation processes (e.g., averag- ing) might imply. We explore these questions through a computational model of the Condorcet Jury Theorem, advancing the literature by manipulating social links and the nature of the exchanges between actors.
While the causes of attitudes toward immigration have been studied extensively, what the literature is still grappling with is how much elite influence exists over public opinion. Studies exploring the effects of elite framing have largely used laboratory experiments or have focused on exogenous national events. What is lacking in the literature is a real-time measure of public sentiment that would allow researchers, in response to an elite framing attempt, to establish a “before, during, and after.” Twitter allows us to measure public sentiment in response to a major event before political elites are able to frame the discussion.
While ultimately we aim to explore how much elites influence public sentiment, this paper focuses on an important first step. This paper addresses whether or not existing theories about public opinion on immigration that have been developed using traditional public opinion surveys can explain what we observe in Twitter. That is, do feelings of racial threat and economic competition explain variation in Twitter sentiment? We find that existing theories developed using traditional public opinion surveys do successfully explain negative sentiment on Twitter. This paper has important implications for using Twitter as a tool to measure public attitude and helps further our understanding of the dynamics of opinion on immigration.
We describe and investigate a model of strategic sequential decision-making in networked policymaking environments with three agents. Our primary interest is the effect of network structure on sequential policymaking and information aggregation. Our model and results illus- trate how individual policy decisions of varying weight (in terms of a decision-maker’s unilateral effect on policy outcomes) can enable information aggregation in decentralized environments. In our environment the incentive compatibility conditions for information aggregation are not invariant to network isomorphisms: individuals’ positions in the network matter. We derive ex- act conditions for every acyclic network of 3 or fewer agents and illustrate the counterintuitive nature of comparative statics with respect to both network structure and individual agents’ policy preferences and discretionary authority.
This paper compares congressional networks within the 112th House of Representatives in order to examine the similarities and differences between social media networks and job-related networks within organizations. An understanding of the dynamics between social media networks and job-related networks becomes more and more important as social media networks gain importance in the day-to-day activities of political leaders. The paper compares Twitter and cosponsorship networks to analyze and compare coalitions that form within the body. The results indicate that while congressional cosponsorship networks are structured by committee, state of representation, and ideology, congressional Twitter networks show little evidence of social media relationships grounded in committee or state. Congressional Twitter networks tend to be structured by ideology and the leadership hierarchy of the parties. Both data sets can be useful, but not sufficient, for ideological prediction of the members’ voting patterns.
Increasingly amicus curiae briefs are led by large coalitions which include as many as a hundred dierent organizations on the same brief. While previous studies examine which interest groups le together, no study has yet examined the coalitional activities of state attorneys general (SAGs). SAGs, the representatives of the states at the Court operate in a dierent institutional context than other amici which makes it impossible to generalize from other amici. Moreover, sitting at the intersection of the state and federal levels of government SAGs are an attractive source of information for the Court; indeed SAGs are the second most inuential class of amici at the Court trailing only the federal solicitor general. Since previous research nds SAGs are more successful in larger coalitions, it is critical to understand how SAG amicus curiae brief coalitions form and which actors take central roles in their network. Drawing on the interest group literature network, as well as previous work on SAGs and elite attorneys, I argue SAGs coalitional activity is inuenced by a conuence of political and administrative factors. In this manuscript, I employ descriptive social network analysis and exponential random graph models to provide the rst systematic analysis of the SAG amicus curie brief network. I nd the shape of the SAG network cannot be reliably predicted by either political or administrative factors. However, I nd ties formed between SAGs are governed by a combination of institutional and resource based explanations.
This paper reconsiders Senate action on the 1957 Civil Rights Act, employing a network model of legislative behavior. The Congress possessed a long history of deep divisions over issues related to civil rights for African Americans. Indeed, the 1957 Act became the first major civil rights legislation adopted by the Congress since the end of Southern Reconstruction more than 80 years earlier. Why and how a civil rights bill managed to pass the Senate provides insight regarding both the dynamics of the American party system, as well as the processes by which legislative progress can be achieved in the face of seemingly intractable polarization. We argue that the key to legislative success, as well as to the influence of individual legislators, is directly related to the centrality of particular senators within particular voting blocks, located within networks of relationships that make legislative progress possible. Hence the influence of individual members is not simply due to their ability to cast pivotal votes, but also to their strategic capacity to construct networks of support for legislative initiatives, as well as to their centrality relative to networks of interest and communication within the chamber. Our argument is based on the analysis of several crucial votes during Senate consideration of this landmark civil rights legislation.