ISERP Working Papers - Summer Issue
The following initiative showcases the working papers of Columbia's faculty in the Social Sciences departments. The papers have not been peer reviewed. To add your voice to the discussion, please submit your paper to iserp-communication@columbia.edu, or upload it here. Thank you to the faculty who have submitted a working paper for our first bulletin, we are hoping this initiative will tighten the ties of our research community.
To read and download any of these working papers, click on the paper, or visit our new site: ISERP's Working Papers. This site is UNI-protected, and only Columbia and Barnard affiliates have access to the full text.
To receive subsequent bulletins, please sign up here.
Please see below our summer issue of the Working Papers Bulletin. To view the spring issue of our bulletin, please click here.
Topics covered in this bulletin:
- Political Messages and Behavior
- Information and Choice
- Effective Science Communication
- Education
- Violence, Politics, and Crime
- Civilian Behavior in Political Conflict
Political Messages and Behavior
Michael Schwam-Baird, Political Science, Jonathan S. Krasno, Political Science, Binghamton University, Costas Panagopoulos, Political Science, Northeastern University, Donald P. Green, Political Science
We report the results of a field experiment that examines the effect of political information on the decision to contribute. In advance of the November 2014 election, postcards with information about the major party candidates for Ohio governor and secretary of state were mailed to nearly 40,000 randomly selected likely donors in Ohio. Seven types of mailings provided factual information regarding campaign fundraising and endorsements, or a simple election reminder. Notably, the messages did not include encouragements to donate or partisan cues. The experimental results show that likely donors respond to electoral threats as well as electoral opportunities under different conditions. Donors are more likely to give to the stronger candidate when they receive a simple election reminder with no fundraising information. But when presented with information about which candidates are ahead and behind in total fundraising, donors give more to the candidate who is behind while donations to the candidate with more money are unaffected. The results suggest that donors respond to objective information about fundraising weakness in order to help their preferred candidate. Information about the relative size of individual donations and ideological endorsements has no effect on donations.
keywords: information, Field experiment, Political information, Decision to contribute, Electoral threats, Fundraising
BISG: When inferring race or ethnicity, does it matter that people often live near their relatives?
Philip Greengard, Statistics, Andrew Gelman, Political Science and Statistics
Bayesian Improved Surname Geocoding (BISG) is a ubiquitous tool for predicting race and ethnicity using an individual’s geolocation and surname. BISG assumes that in the United States population, surname and geolocation are independent given a particular race or ethnicity. This assumption appears to contradict conventional wisdom including that people often live near their relatives (with the same surname and race). We demonstrate that this independence assumption results in systematic biases for minority subpopulations and we introduce a simple alternative to BISG. Our raking-based prediction algorithm offers a significant improvement over BISG and we validate our algorithm on states’ voter registration lists that contain self-identified race/ethnicity. The proposed improvement and the inaccuracies of BISG generalize to applications in election law, health care, finance, tech, law enforcement and many other fields.
keywords: Bayesian Improved Surname Geocoding (BISG), Race and ethnicity prediction, Geolocation, Minority populations, Raking-based prediction algorithm
Narrative Entertainment Changes Minds: A Meta-Analysis of Recent Experiments
Bardia Rahmani, Political Science, Dylan W. Groves, Political Science, Beatrice Montano, Political Science, Donald P. Green, Political Science
Narrative entertainment has attracted increasing attention from social scientists and policy makers. One strand of research seeks to understand whether the entertainment that audiences consume has incidental effects on their beliefs, attitudes, and behaviors; another, whether purpose-built narrative media campaigns hold promise as a means of addressing a wide array of social, economic, and political problems. Building on previous literature reviews, we present results from a meta-analysis of 81 experiments that assess the persuasive effects of narrative entertainment, including a recent wave of studies in low- and middle-income countries. The results suggest that narrative entertainment, on average, is quite influential, with persuasive effects that remain apparent weeks to months after initial exposure. A smaller literature reports head-to-head tests of the relative effectiveness of narrative versus non-narrative messages; although inconclusive, the evidence suggests that narrative messages may be only slightly more persuasive as non-narratives messages. If true, this finding would imply that the main advantage of narratives may be their ability to attract and engage large audiences. We conclude by calling attention to gaps in the literature and proposing avenues for further research.
keywords: Narrative entertainment, Meta-analysis, Beliefs, attitudes, and behaviors, Media campaigns, Audience engagement
Lemonade from Lemons: Information Design and Adverse Selection
Navin Kartik, Economics, Weijie Zhong, Business, Stanford University
A seller posts a price for a single object. The seller’s and buyer’s values may be interdependent. We characterize the set of payoff vectors across all information structures. Simple feasibility and individual-rationality constraints identify the payoff set. The buyer can obtain the entire surplus; often, other mechanisms cannot enlarge the payoff set. We also study payoffs when the buyer is more informed than the seller, and when the buyer is fully informed. All three payoff sets coincide (only) in notable special cases—in particular, when there is complete breakdown in a “lemons market” with an uninformed seller and fully-informed buyer.
keywords: Price posting, Interdependent values, Payoff vectors, Information structures, Seller and buyer relationship
A Better Test of Choice Overload
Mark Dean, Economics, Dilip Ravindran, Economics, Humboldt University Berlin, Jorg Stoye, Economics, Cornell University
Choice overload – by which larger choice sets are detrimental to a chooser’s wellbeing – is potentially of great importance to the design of economic policy. Yet the current evidence on its prevalence is inconclusive. We argue that existing tests are likely to be underpowered and hence that choice overload may occur more often than the literature suggests. We propose more powerful tests based on richer data and characterization theorems for the Random Utility Model. These new approaches come with significant econometric challenges, which we show how to address. We apply our tests to new experimental data and find strong evidence of choice overload that would likely be missed using current approaches.
keywords: Choice overload, Wellbeing, Economic policy, Random Utility Model, Experimental data
Effective Science Communication
The ladder of abstraction in statistical graphics
Andrew Gelman, Political Science and Statistics
Graphical forms such as scatterplots, line plots, and histograms are so familiar that it can be easy to forget how abstract they are. As a result, we often produce graphs that are difficult to follow. We propose a strategy for graphical communication by climbing a ladder of abstraction (a term from linguistics that we borrow from Hayakawa, 1939), starting with simple plots of special cases and then at each step embedding a graph into a more general framework. We demonstrate with two examples, first graphing a set of equations related to a modeled trajectory and then graphing data from an analysis of income and voting.
keywords: Graphical forms, Abstraction, Scatterplots, Line plots, Histograms, Voting, Communication, statistical information
PLUS or Minus? The Effect of Graduate School Loans on Access, Attainment, and Prices
Sandra E. Black, Economics and SIPA, Lesley J. Turner, Economics, Vanderbilt University, Jeffrey Denning, Economics, Brigham Young University
In 2006, the federal government effectively uncapped student borrowing for graduate programs with the introduction of the Graduate PLUS loan program. Access to additional federal loans increased graduate students’ borrowing and shifted the composition of their loans from private to federal debt. However, the increase in borrowing limits did not improve access to existing programs overall or for underrepresented groups. Nor did access to additional loan aid result in significant increase in constrained students’ persistence or degree receipt. We document that among programs in which a larger share of graduate students had exhausted their annual federal loan eligibility before the policy change—and thus were more exposed to the expansion in access to credit—federal borrowing and prices increased.
keywords: Graduate PLUS loan program, Student borrowing, Federal debt, Access to credit, Persistence, Student loans, Education
Wouter Dessein, Business, Alex Frankel, Booth School of Business, University of Chicago, Navin Kartik, Economics
The Covid-19 pandemic has accelerated the trend of many colleges moving to test-optional, and in some cases test-blind, admissions policies. A frequent claim is that by not seeing standardized test scores, a college is able to admit a student body that it prefers, such as one with more diversity. But how can observing less information allow a college to improve its decisions? We argue that test-optional policies may be driven by social pressure on colleges’ admission decisions. We propose a model of college admissions in which a college disagrees with society on which students should be admitted. We show how the college can use a test-optional policy to reduce its “disagreement cost” with society, regardless of whether this results in a preferred student pool. We discuss which students either benefit from or are harmed by a test-optional policy. In an application, we study how a ban on using race in admissions may result in more colleges going test optional or test blind.
keywords: Covid-19 pandemic, Education, College admissions, Standardized test scores, Diversity, Social pressure, Test-blind policies, Test-optional policies, Consideration of race in admissions
Gender After Genocide: How Violence Shapes Long-Term Political Representation
Nikhar Gaikwad, Political Science, Erin Lin, Political Science, Ohio State University, Noah Zucker, Niehaus Center for Globalization & Governance
What are the legacies of violence on women’s political representation? We examine the long-term effects of a watershed conflict of the twentieth century: the Khmer Rouge genocide, during which 50–70% of Cambodia’s working-age men were killed. Using original data on mass killings and economic and political conditions in Cambodian communes, we find that genocide exposure is positively associated with women’s economic advancement and present-day indicators of women’s representation in local-level elected office. We conduct in-depth, ethnographic interviews with genocide survivors to explore the mechanisms by which violence spurred women into elected office. A crucial finding emerges: In areas that suffered the genocide’s worst killings, widows obtained economic autonomy, providing a template for the economic advancement of women in traditional households with surviving men. The shift in norms regarding the sexual division of labor and its transmission through intra-communal and intergenerational pathways allowed women to adopt larger public roles over time.
keywords: Gender violence, Legacies of violence, Women’s political representation, Genocide, Economic advancement, Gender roles
The Fall of Violence and the Reconfiguration of Urban Neighborhoods
Gerard Torrats-Espinosa, Sociology, Patrick Sharkey, Sociology, Princeton University
Over the past few decades, American cities have undergone dramatic change driven in large part by two major trends: the fall of violence, and the rise of urban inequality. Despite the attention given to each of these trends, there has been little research designed to assess how they are related to each other. This study is the first one to generate causal evidence on the impact of violent crime on economic residential segregation. We document the impact of the crime drop on economic segregation in 500 cities in the United States between 1990 and 2010, using temporal shocks to city crime rates to identify causal effects. We find a strong causal effect of violence on the segregation of poor households, but we find no impacts on the segregation of affluent households. Our findings indicate that while the crime decline has not overturned the trend toward rising economic segregation, it has slowed its pace. In cities where crime declined more substantially, the segregation of poor households has grown more slowly or not at all. Additional exploratory results suggest that the changes in the segregation of low-income households are partially driven by how the decline in violence changed the demographic composition of low-income neighborhoods. Our findings reveal that the fall in violence increased the concentration of white and college-educated residents in the neighborhoods that were at the bottom of the income distribution in 1990. We also find suggestive evidence that falling violence led to the migration of poor households out of low-income neighborhoods, a finding that could reflect a process of gentrification and displacement of the poor. Although the rise of urban inequality has continued even as violence has fallen, the crime decline has had its greatest impact on concentrated poverty, which has long been thought of as one of the most problematic and harmful dimensions of urban inequality.
keywords: Violence, American cities, Urban inequality, Economic residential segregation, Crime drop
Civilian Behavior in Political Conflict
The Politics of Allyship: Multiethnic Coalitions and Mass Attitudes Towards Protest
Devorah Manekin, International Relations, Hebrew University of Jerusalem, Tamar Mitts, SIPA, Data Science Institute and the Saltzman Institute, Yael Zeira, Political Science, Maxwell School of Citizenship & Public Affairs, Syracuse University
Recent work finds nonviolent resistance by ethnic minorities is perceived as more violent and requiring more policing, reducing its impact and effectiveness. We ask whether allies–committed participants from advantaged groups–can mitigate these barriers, and at what cost to movements’ core constituencies. We draw on the case of Black Lives Matter, which, in 2020, attracted unprecedented white participation. Employing a national survey experiment, we find sizeable white presence at racial justice protests increases protest approval, reduces perceptions of violence, and lowers support for policing among majorities, while not causing backlash among minority audiences. We complement our experimental results with analysis of tens of thousands of images shared on social media during the 2020 BLM protests, finding a significant association between the presence of whites in images and user engagement, amplification, and positive sentiment. These findings suggest that allyship can be a powerful tool for promoting sociopolitical change amidst deep structural inequality.
keywords: Nonviolent resistance, Ethnic minorities, Black Lives Matter, Sociopolitical change, Political sentiment, Policing, Racial Justice, White participation in BLM protests
Information Operations, Territorial Control, and Perceptions of Power During Conflict
Tamar Mitts, SIPA, Data Science Institute and the Saltzman Institute, Asfandyar Mir, United States Institute of Peace, Austin Wright, Harris School of Public Policy, University of Chicago
A large literature on civilian behavior during conflict points to territorial control as an important factor driving citizens’ conduct in war. According to the conventional argument, civilians’ effort to maximize survival leads them to politically align with the actor that controls territory. We argue that this account needs to be adapted to the digital era, where perceptions of power and control can transcend territory as a result of information operations on social media platforms. When conflict actors use digital media to influence public views about their military strength, they can induce favorable civilian behavior – even in areas that they do not physically control. We test this argument in the context of the Taliban’s takeover of Afghanistan in 2021. Drawing on rich data on the group’s online influence efforts, as well as on geolocated data on Afghan citizens’ activity on Twitter, we find, in line with the canonical literature, that civilian behavior varies with territorial control. But patterns of citizens’ resistance and compliance with the Taliban are strongly moderated by the degree of exposure to the group’s influence efforts: individuals that had greater exposure to the Taliban’s propaganda on social media started complying earlier, and stopped resisting faster, even in areas that the group did not yet physically control. These findings point to insurgents’ online information operations as an important, yet under-appreciated, channel that can shape civilian behavior during war.
keywords: Violence, Terrorism, Civilian behavior, Territorial control, Digital platforms, Information operations
To read and download any of these working papers, click on the paper, or visit our site, ISERP's Working Papers. This site is UNI-protected, and only Columbia and Barnard affiliates have access to the full text.
To receive subsequent bulletins, please sign up here.