My work focuses on how institutions, public policy, and the physical environment shape preferences and behavior related to climate change. I use a combination of experiments, public opinion data, and formal theory to answer questions such as: When do people believe in climate change? When are they willing to support climate change mitigation policies? How do we govern emerging technologies (e.g., geoengineering) in the context of climate change? I am also more broadly interested in political behavior, and how public opinion is shaped by disasters.
I am an assistant professor in the department of political science at the University of Connecticut. Previously, I was a postdoctoral research associate at Princeton University in the Center for the Study of Democratic Politics. I earned my PhD in the department of Political Science at Stony Brook University in August of 2020.
Postdoctoral Research Associate, 2020 - 2021
PhD in Political Science, August 2020
Stony Brook University
BA in Political Science & Psychology, 2016
University of Portland
Climate Change Literacy in Africa
Nicholas Philip Simpson, Talbot M. Andrews, Matthias Kronke, Christopher Lennard, Romaric C. Odoulami, Birgitt Ouweneel, Anna Steynor, & Christopher H. Trisos
Nature Climate Change, Accepted
Climate change literacy is defined as being both aware of climate change and its anthropogenic causes, and thus underpins informed mitigation and adaptation responses. However, climate change literacy rates and their predictors remain poorly understood across the Global South. Here, we analysed Africa’s largest, representative public opinion survey to quantify change literacy rates. Climate change literacy ranges from 23–66% of the population across 33 countries, with larger variation at subnational scales (e.g., 5–71% among states in Nigeria). Strong positive predictors of climate change literacy are education and mobility, but poverty decreases climate change literacy and women are less likely to be climate change literate than men. Perceived drought experiences and historical trends in precipitation are also important predictors. These results highlight where services can target specific regions and demographics to increase climate change literacy and help ensure adaptation responses are informed by better understandings of future climate change.
Preferences for Prevention: People Assume Expensive Problems Have Expensive Solutions
Talbot M. Andrews & John Barry Ryan
Risk Analysis, 2021
People support inefficient spending on preventing disasters, and these preferences are translated into inefficient policies as elected officials try to appeal to their constituents. Here, we find preferences for prevention spending are biased by the “cost conflation” mechanism, where people assume expensive problems have expensive solutions. In this paper, we present a formal model of collective action, and illustrate how cost conflation causes people to deviate from the equilibria. We test for these hypothesized deviations using an incentivized experiment. The experimental subjects engage in cost conflation – they believe the costs of disaster prevention are positively related to the costs of disaster damages, even when explicitly told otherwise. As a result, they fail to prevent smaller disasters and pay too much to prevent large disasters when cheap solutions exist. Furthermore, we provide evidence that overemphasizing disaster damages undermines successful disaster prevention because people view these disasters as too big to solve.
Is a Rational Politics of Disaster Possible? Making Useful Decisions for Others in an Experimental Disaster Game
Talbot M. Andrews, Andrew W. Delton, & Reuben Kline
Political Behavior, 2021
Disaster responses are political. But can citizens make useful disaster decisions? Potential obstacles are that such decisions are complex, involve public goods, and often affect other people. Theories of political decision-making disagree on whether these problems can be overcome. We used experimental economic games that simulate disaster to test whether people are willing and able to prevent disasters for others. Groups of players face a complex task in which options that might help vary in their riskiness. Importantly, although all options are reasonable, which option is most useful depends on the experimental condition. We find that players will pay to help, can identify which option is most useful across experimental conditions, and will pay to learn how best to help. Thus, players were able to make useful and costly decisions to prevent others from experiencing disaster. This suggests that, in at least some situations, citizens may be able to make good disaster decisions.
From Moral Hazard to Risk-Response Feedback
Joseph Jebari, Olufemi Taiwo, Talbot M. Andrews, Valentina Aquila, Brian Beckage, Mariia Belaia, Maggie Clifford, Jay Fuhrman, David P. Keller, Katharine J. Mach, David R. Morrow, Kaitlin T. Raimi, Daniele Visioni, Simon Nicholson, & Christopher H. Trisos
Climate Risk Management, 2021
The Intergovernmental Panel on Climate Change assessments (IPCC) Special Report on 1.5°C of global warming is clear. Nearly all pathways that hold global warming well below 2°C involve carbon removal (IPCC, 2015). In addition, solar geoengineering is being considered as a potential tool to offset warming, especially to limit temperature until negative emissions technologies are sufficiently matured (MacMartin et al., 2018). Despite this, there’s been a reluctance to embrace carbon removal and solar geoengineering, partly due to the perception that these technologies represent what is widely (though, we argue here, problematically) termed a “moral hazard”: that geoengineering will prevent people from developing the will to change their personal consumption and push for changes in infrastructure (Robock et al., 2010), erode political will for emissions cuts (Keith, 2007), or otherwise stimulate increased carbon emissions at the social-system level of analysis (Bunzl, 2008). We argue that debates over “moral hazard” in response to carbon removal and geoengineering are unhelpful and misleading. We also propose an alternative framework for dealing with the tradeoffs that motivate the appeal to “moral hazard,” which we call “risk-response feedback.”
Who feels the impacts of climate change?
Talbot M. Andrews & Oleg Smirnov
Global Environmental Change, 2020
Feeling affected by climate change related disasters has the potential to mobilize belief in climate change, concern about the issue, and support for mitigation policies – even when accounting for the effects of physically living through a disaster. In this study we use a two-wave survey design where respondents in the United States were interviewed before and after Hurricane Florence to better understand who feels affected by such disasters. First, we find that being worried about climate change increases the feeling of being affected by the hurricane among those who regularly discuss climate change. Second, we find that those who are high in perspective taking are more likely to feel affected. However, those who are high in empathic concern, but feel obligated to help victims of disasters, are less likely to feel affected. This suggests that hurricanes may cause a collapse of compassion, where those who are especially sensitive to the suffering of others down-regulate their emotional response to costly disasters.
Rebel Recruitment and Retention in Civil Conflict
Katherine Sawyer & Talbot M. Andrews
International Interactions, 2020
While the conflict literature has examined the use of forced recruitment in conflict, the question remains why groups would choose to do so when forced recruits require expensive coercion (Eck 2014) and time intensive socialization processes (Gates 2017). The prevailing wisdom in the literature is that forced recruitment is a tactic of the weak; yet empirically, we often observe relatively strong rebel groups employing forced recruitment. In this paper, we argue that credible threats of punishment for desertion are a prerequisite to successful coercive recruitment. Thus, stronger rebels, those that are able to credibly threaten punishment, are more likely to engage in forced recruitment than are weaker rebels. Forced recruitment is not a tactic of last resort but a human rights abuse frequently exploited by already advantaged rebel groups. We find strong support for our argument quantitatively and qualitatively using cross-national data on rebel recruitment practices (Cohen 2016) and case illustrations of the contras in Nicaragua and the Farabundo Martí National Liberation Front (FMLN) in El Salvador. The results speak to the growing literature emphasizing the importance of integrating individual and group level processes both theoretically and empirically.
When Trust Matters: The Case of Gun Control
John Barry Ryan, Talbot M. Andrews, Tracy Goodwin, & Yanna Krupnikov
Political Behavior, 2020
Declining trust in government is often cited as the cause of declining support for policies that require ideological sacrifices. At the same time, whether trust in government affects attitudes in a particular issue area is conditional on the political context and can vary over time. We argue and show that when political parties polarize on an issue, then individuals who do not trust the government fear the “slippery slope”. Trust in government affects public policy attitudes when individuals believe small ideological costs now could be the beginning of a process that leads to large ideological costs later. We test the argument on the case of gun control using two datasets. We first show that trust in government affects conservatives’ gun control attitudes as polarization increases over the issue. We then use a continuum of gun control policies to demonstrate that fear of the slippery slope is the mechanism explaining why trust matters.
High Risk-High Reward Investments to Mitigate Climate Change
Talbot M. Andrews, Andrew W. Delton, & Reuben Kline
Nature Climate Change, 2018
Some technologies, such as solar or wind power, create certain but relatively small reductions in greenhouse gas emissions. Others, such as carbon sequestration devices, have larger potential upsides, but a greater possibility of failure. Here we show using economic games that people will invest in high-risk high-reward technologies when more certain options will not be sufficient. Groups of players had to contribute enough to avoid a simulated climate change disaster. Players could defect, make a certain contribution or make a risky contribution with a high potential gain. Across four studies using both laboratory (n= 296 and n= 297) and online (n= 501 and n= 499) samples, we found that more players made riskier contributions when necessary targets could not be met otherwise, regardless of the magnitude of potential losses. These results suggest that individuals are willing to invest in risky technology when it is necessary to mitigate climate change.
Replication data available here
Cue-based estimates of reproductive value explain women’s body attractiveness
Talbot M. Andrews, Aaron W. Lukaszewski, Zachary L. Simmons, & April Bleske-Recheck
Evolution and Human Behavior, 2016
Women’s body attractiveness is influenced by specific anthropometric cues, including body mass index (BMI), waist-to-hip ratio (WHR), waist-to-stature ratio (WSR), and shoulder-to-waist ratio (SWR). Despite the existence of multiple functional hypotheses to explain these preferences, it remains unclear which cue-based inferences are most influential in regulating evaluations of women’s body attractiveness. We argue that (i) the common link to the morphological cues that influence women’s body attractiveness is that they all reliably indicate high reproductive value (as defined by youth and low parity); and (ii) ancestrally, selection pressures related to tracking between-women differences in reproductive value would have been among the strongest acting on adaptations for body evaluation. An empirical study then tested the resulting prediction that cue-based estimates of reproductive value function as powerful regulators of women’s body attractiveness judgments. Subjects viewed standardized photos of women in swimsuits (with heads obscured), and were assigned to either estimate components of their reproductive value (age or number of offspring) or rate their attractiveness. Structural equation modeling revealed that a latent variable capturing estimated reproductive value was almost perfectly correlated with a latent variable capturing body attractiveness. Moreover, unique associations of women’s BMI, WHR, and WSR with their body attractiveness were entirely mediated via estimated reproductive value. These findings provide strong support for the longstanding hypothesis that women’s body attractiveness is primarily explained by cue-based estimates of reproductive value – expected future utility as a vehicle of offspring production.
Two-wave panel survey dataset on who feels affected by Hurricane Florence
Talbot M. Andrews & Oleg Smirnov
Data in Brief, 2020
Feeling affected by climate change related natural disasters is an important predictor of engaging in climate change mitigation behavior. We therefore collected data to identify who felt affected by Hurricane Florence, which made landfall in the United States on September 14th, 2018. In the months before Hurricane Florence, we collected survey responses from a nationally representative sample of United States citizens. We measured their attitudes towards climate change, emotional predispositions, and demographic information. Then, in the days after the hurricane, we re-contacted respondents to identify whether or not they felt personally affected by Hurricane Florence. These data can be used first to identify variables associated with climate change attitudes, and second to identify the traits that predispose individuals to feel affected by climate change related disasters.
The Political Complexity of Attack and Defense
Talbot M. Andrews, Leonie Huddy, Reuben Kline, H. Hannah Nam, and Katherine Sawyer
Commentary in Behavioral and Brain Sciences, 2019
De Dreu and Gross’s distinction between attack and defense is complicated in real-world conflicts because competing leaders construe their position as one of defense, and power imbalances place status quo challengers in a defensive position. Their account of defense as vigilant avoidance is incomplete because it avoids a reference to anger which transforms anxious avoidance into collective and unified action.
Changing Climate in the Lab
Talbot M. Andrews
Article in The Experimental Political Scientist: Newsletter of the APSA Experimental Research Section, 2019
Graduate Women’s Writing Groups
Talbot M. Andrews
Blog Post on Behavioral and Social Sciences at Nature Research, 2019
The Connection Between Foraging Birds and Fighting Climate Change
Talbot M. Andrews, Andrew W. Delton, & Reuben Kline
Blog Post on Behavioral and Social Sciences at Nature Research, 2018
Beyond Market Behavior: Evolved Cognition and Folk Political Economic Beliefs
Talbot M. Andrews & Andrew W. Delton
Commentary in Behavioral and Brain Sciences, 2018
Mobilizing Mitigation: Overcoming the Social Dilemma of Climate Change
Committee: Reuben Kline (Co-Chair), John Barry Ryan (Co-Chair), Andrew W. Delton, Yanna Krupnikov, & Jennifer Wolak
In this series of papers I use a variety of methods to answer important questions related to climate change mitigation. Why do we cooperate (and fail to cooperate) with each other when overcoming the social dilemmas inherent in climate change mitigation? How do natural disasters shape mitigation policy preferences? How do citizens and environmentalists trade off between environmental conservation and climate change mitigation?
Chapter 1: The Paradox of Climate Concern: Declining Mitigation Support after Natural Disasters
Can people respond to political shocks with appropriate policy preferences? Political shocks capture the attention of the public and encourage citizens to consider how to respond to the event. However, they also activate competing considerations which may lead to counterintuitive and counterproductive policy preferences. Here I focus on the case of exposure to natural disasters and support for climate change mitigation policies. Leveraging Hurricane Sandy and Hurricane Matthew as natural experiments, I find that exposure to hurricanes has a limited effect on belief in climate change and general climate change mitigation. Unfortunately, exposure to these disasters significantly decreases support for a key mitigation technology: nuclear power. While personal experiences may help people update their beliefs, the path from personal experience to policy preferences is complex and undermined by other considerations activated by disasters.
Paper presented at the 2020 Southern Political Science Association Conference
Paper to be presented at the 2020 MapleMeth Conference
Chapter 2: Conditional on what? Disentangling intention and equality motivations behind conditional cooperation
Why citizens engage in costly political participation is one of the most persistent puzzles in political science. They overcome collective action problems, for example by voting and turning out to protest. Extensive work using public goods games show conditional cooperators, those who cooperate with other cooperators, are critical to overcoming such dilemmas. But we know little about what motivates these cooperators. Are they swayed because they perceive those around them to have cooperative intentions? Or are they driven by general concerns for equity? Using an incentivized experiment, I find that cooperative intentions sustain conditional cooperation rather than a desire to maintain an equitable distribution of resources. This suggests that in order to effectively motivate cooperation in domains such as turning out to vote, we should emphasize the number of people already successfully doing so.
Poster presented at the 2019 CSAP American Politics Conference
Paper presented at the 2019 EITM Summer Institute
Paper presented at the 2020 Winter Experimental Social Science Institute (WESSI)
Chapter 3: Who Are They Fighting For? The Principal-Agent Problem of Environmental Activism
Activists have the potential to overcome the problem of low political participation. However, this creates a principal-agent problem that is exacerbated when activists identify and act on an issue that is distinct from the issue identified by the larger groups they represent. I illustrate this problem using the case of climate change mitigation. In this instance, the principals are those who support climate change mitigation policies but are not politically active, while the agents are environmentalists. In order for environmentalists to serve as effective agents, (1) there actually have to be individuals who identify as environmentalists, (2) environmentalists must be politically active, and (3) environmentalists need to advocate for climate change mitigation policies. Using a novel scale to identify environmentalists, I find that environmentalists exists and are politically active. However, instead of prioritizing the problem of climate change, they prioritize environmental conservation.
Chapter 4: (Antisocial) Punishment is Equally Effective Across cultures
Revise & Resubmit
A stable minority of individuals engage in antisocial punishment: inflicting costs on cooperators. Computational models suggest this undermines the evolution of cooperative strategies, and there appears to be massive cross-cultural variability in punishment behavior. Using data from existing work which conducted public goods games in industrialized countries throughout the world (Herrmann, Thöni, and Gächter 2008), I find this isn’t the case. Antisocial punishment is as effective as any other form of punishment in increasing cooperation. Furthermore, the relationship between how much someone cooperates and how much they are punished does not vary across cultures. Instead, this apparent variability is a downstream byproduct of variability in people’s willingness to cooperate. This suggests that future work should focus on why first order defectors are willing to engage in second order cooperation.
Poster presented at the 2020 NYU CESS conference
Pre-Prints Available for All Projects on Request
Who do you trust? Inefficiency Incentives in Climate Change Mitigation
Talbot M. Andrews, Andrew W. Delton, & Reuben Kline
Revise & Resubmit
Climate change is an extremely polarized issue in the United States, with leaders across the political spectrum sending very different messages about whether and how we should implement mitigation policies. Do citizens have the tools necessary to distinguish between helpful and unhelpful information about mitigation policies? Leaders have different incentives which constrain their support for or opposition to mitigation spending. Here we test whether citizens are sensitive to different institutions which may give leaders an incentive to misrepresent the cost of providing public goods like mitigation or disaster prevention. We use an incentivized experiment to do so, specifically using a modified collective risk social dilemma. In this public goods game, players must contribute enough money to prepare for an ongoing disaster. Leaders know the exact cost of damage prevention, and send signals to the other players about the cost. We show that people are sensitive to institutional differences: when leaders have a stake in inefficiency, citizens trust the leader less and contribute less to the public good. In the midst of bleak research on mitigation policy support, we provide optimistic evidence of people’s ability to differentiate between helpful and unhelpful information about mitigation policies
Too Many Ways to Help: How to Promote Climate Change Mitigation Behaviors
Talbot M. Andrews, Reuben Kline, Yanna Krupnikov, & John Barry Ryan
Revise & Resubmit
What are the most effective messages to mobilize people to engage in climate change mitigation behaviors? One common strategy is to tell individuals about many easy ways they can get involved. However, psychological theories of choice suggest this communication strategy might backfire: when presented with too many options, people become less likely to make any choice at all. Here we conduct a two-wave survey experiment to see if a similar phenomenon occurs with regards to mitigation. In the first wave, we randomly assigned subjects to see messages encouraging either 1, 5, 10, or 20 pro-environmental behaviors drawn from a set which was pre-tested to get a rating of how difficult they were to carry out. Consistent with a theory we call “mitigation overload”, we find that subjects who saw a message suggesting 20 easy ways they could engage in climate change mitigation felt less efficacious. One week later, these subjects also reported engaging in fewer mitigation behaviors compared to those who saw fewer ways to mitigate. But, introducing more difficult items to the list alleviated mitigation overload and increased efficacy. These results suggest more isn’t always better when communicating ways individuals can help stop climate change.
Anticipating Moral Hazard Undermines Climate Mitigation in an Experimental Geoengineering Game
Talbot M. Andrews, Andrew W. Delton & Reuben Kline
Geoengineering is often touted as a solution to climate change but will only be successful in conjunction with other mitigation strategies. This creates a potential for a “moral hazard”: If people think geoengineering alone will mitigate climate change, there is little incentive for them to support more piecemeal but necessary mitigation efforts. We test this in a series of economic games where players in groups must prevent a simulated climate disaster. One player, the “policymaker,” decides whether to implement geoengineering. The rest are “citizens” who decide how much to contribute to more piecemeal mitigation efforts. We find that citizens contribute to mitigation even when the policymaker uses geoengineering. Despite this, policymakers expect that citizens will engage in moral hazard. This expectation leads policymakers to not use geoengineering even though everyone would be better off if they used it. Moral hazard anticipation undermines climate change mitigation, rather than the moral hazard itself.
Close to Home: How Local Media Shapes Climate Change Attitudes
Talbot M. Andrews, Cana Kim, & Jeong Hyun Kim
Highlighting the local impacts of climate change has the potential to increase the public’s awareness of and engagement with climate change. However, information about local impacts is only effective when delivered by trusted sources such as co-partisan political leaders. Is information about climate change conveyed by local media sources similarly beneficial? We argue that local media are well positioned to communicate the local implications of climate change, thereby enhancing the public’s risk perceptions of climate change and willingness to take climate action. We further hypothesize that climate coverage by local media, the media type that is more trusted across party lines, will have more profound effects on Republicans. Using the case of Louisiana, we first demonstrate that local and national newspapers cover climate change in substantially different ways, with local media more consistently focused on local impacts. Our survey experiment of Louisiana residents reveals that Republicans viewed the coverage of a hurricane in the region more positively when it came from a local newspaper rather than a national newspaper. Furthermore, local newspapers’ climate coverage increased Republicans’ willingness to take action to mitigate climate change. These results provide insights into the effective communication of climate change to the public and the role of local media in mitigating partisan polarization.
The road to reelection is paved with good intentions: Experiments on the role of outcomes and intentions in voting behavior
Talbot M. Andrews & Scott Bokemper
Politicians signal their intended policy outcomes to voters, such through campaign promises. However, theories of retrospective voting suggest these intentions don’t impact vote choice – so why do politicians spend resources on signaling their intentions? Here, we propose an outcome-intention theory of voting in which voters weigh both policy outcomes and intentions in their choice between candidates. We conducted four incentivized experiments using a novel election game. In these experiments, participants took the role of voters who experienced a better or worse policy enacted by an incumbent and then decided whether to reelect the incumbent or vote in a challenger. We manipulated whether the voters knew whether the incumbent intended to enact the outcome that voters experienced. Across all four experiments, we found intentions matter. Voters rewarded incumbents who purposefully enacted good outcomes and forgave incumbents for bad outcomes that were not intended.
Contextual Constraints on Discussion Partner Choice and Correct Voting
John Barry Ryan & Talbot M. Andrews
Using a group based experiment, we demonstrate how contextual constraints in discussion partner choice can aid the efficacy of discussion as an information shortcut. Previous research has shown that political agreement is more important than expertise when individuals seek to use discussion as an information shortcut. In contrast, individuals typically place more weight on expertise than disagreement when choosing political discussion partners. As a result, the efficacy of political discussion improves when the supply of discussion partners is constrained to likeminded individuals. In effect, this prevents individuals from choosing a knowledgeable discussion partner who may provide biased information that misleads the individual. Ultimately these constraints result in more accurate information sharing and more individual-level correct voting when individuals are sorted into neighborhoods based on political preference. Further, the electorate’s preferred candidate – typically the less extreme candidate – is more likely to win when voters are sorted into ideologically distinct neighborhoods.
Summer 2019 & 2020
(Online, cross listed as a masters and undergraduate course)
Environmental policy is one of the most important and contentious areas of politics today. In this course we will cover how environmental policy is made, what major federal policies exist today, and how policies relate to public opinion and individual behavior. Overall, by addressing these three topics, this course will give students the tools to understand and evaluate the complex environmental policymaking environment.
Winter 2019 (Online)
The US Congress is, literally, the most powerful legislative body in the world. In fact, the US Congress has more authority over policy-making than the legislatures of most other democratic nations. What does this power mean for American democracy and what does this power mean for the policy process? In this course, we will consider the inner workings of Congress through three questions: What happens during Congressional elections? How do members legislate? And do outside forces have influence on Congressional politics?