2017-2018 Winter Colloquia

ABSTRACT: Majorities of many nations around the world believe that climate change is a major threat. But individuals don’t typically think they can’t make a difference in the face of such an overwhelming threat. They may feel guilty about their “carbon footprint” but don’t understand their leverage to exert change. In the era of big data, machine learning, and social networks, individuals can be engaged and harnessed not only to reduce their personal impact, but also to cause governments and corporations to act on climate change. The Climate Action Platform will be a collective action machine that can transform this collective action problem. CAP is designed to harness the power of millions of people to take on climate change. In partnership with leading NGOs around the world and companies who sell low carbon goods and services, CAP will build a network of profiled users and present each with free, customized reports on their climate risk exposure and actions they can take both individually and jointly as voters, consumers, and shareholders to address climate change while capturing benefits in their daily lives. Utilizing advanced engagement and behavioral design techniques, CAP will attract and organize these users into syndicated campaigns that prod government and corporations toward unprecedented levels of action.

ABSTRACT: American politics overrepresents the wealthy. But what policies do the wealthy support? Many accounts implicitly assume the wealthy are monolithically conservative and that increases in their political power will increase inequality. Instead, we argue there is substantial heterogeneity by industry, wherein the wealthy from an industry can share a distinctive set of political preferences. Consequently, how increases in the wealthy’s influence affect inequality depends on which industries’ rich are gaining influence and which issues are at stake. We demonstrate our argument with three original surveys, including the two largest surveys of wealthy Americans to date: one of technology entrepreneurs—a burgeoning wealthy demographic—and another of political campaign donors. We show that technology entrepreneurs support liberal redistributive, social, and globalistic policies but conservative regulatory policies—a bundle of preferences rare among other wealthy individuals. Consistent with our theoretical argument, we also present evidence that suggests these differences arise from their distinctive predispositions.

ABSTRACT: Pre-registration of analysis plans involves making data-analysis decisions before the analysis is run in order to prevent flexibly re-running the data until a specific results appears (p-hacking). The complement to p¬-hacking, null-hacking, is the use of the same questionable research practices to re-analyze open data to return a null finding. We provide a vocabulary for null-hacking and introduce the threat it poses to the spirit of the open science movement and pre-registration in particular. Null-hacking forces us to introduce considerations of model fit to compare pre-registered and ‘alternative’ models. The reason null-hacking cannot be ignored is that a null-hacked model can easily provide better fit to the data than a pre-registered one. Model fit, however, is a precarious problem and in the extreme challenges the very purpose of pre-registration. Solutions to the null-hacking problem that focus on model fit eliminate the purpose of pre-registration, while giving preference to pre-registered results ignores how well our models can represent the data. We provide a tentative solution to retain the advantage and purpose of pre-registration, include model fit, and protect against null-hacking. We call this Fully Informed Model Pre-registration and it can either take a supervised machine-learning form or a strict, heavily tested and justified pre-registered confirmatory analysis plan. This solution eliminates exploratory analyses and the ignoring of classic statistical assumptions that are not reported, eliminating the only ground null-hacking has for being accepted.

ABSTRACT: Much coverage of and speculation about the 2016 U.S. presidential election has focused on social media, but little of it has been empirical in nature. Digital inequality suggests adoption of various platforms is one way that differentiated Internet use manifests itself. Twitter and Reddit were two platforms often associated with hateful messages during the elections. Based on national survey data collected in the summer before the election, the paper compares Twitter and Reddit use among Clinton and Trump supporters finding those intending to vote for Trump were much less likely to use these social media than those favoring Clinton. The paper considers the implications of these findings for the potential role of social media in the 2016 presidential election.

ABSTRACT: Keeping adolescents engaged with their schoolwork is a perennial challenge facing U.S. public schools, but in recent years randomized trials have found that interventions based on insights from social and developmental psychology have shown promising effects on adolescents’ motivation and behavior. For example, growth mindset interventions–which involve time-limited (sometimes 30-minute) activities conveying that the brain’s abilities can grow stronger–have resulted in a desire to take on challenges and, among lower-achieving adolescents, higher GPAs. A question often asked about these findings is: shouldn’t psychological intervention effects depend on the school context? In theory, a growth mindset intervention should depend on whether a school provides adequate learning opportunities for students, and on whether or not the school climate allows an initial change in mindset beliefs to become self-sustaining. However, progress in this area has stalled because the available data have not been up to the task of evaluating competing predictions. This talk discusses results from the National Study of Learning Mindsets, which delivered a direct-to-student, online growth mindset intervention in a national probability sample of 76 U.S. public high schools and measured features of the school environment that might interact with the treatment. By replicating growth mindset intervention effects in a generalizable sample, the National Study is able to provide policy-relevant effect size estimates. By identifying sources of treatment effect heterogeneity, the findings contribute to theory about the mechanisms that sustain effects of behavior-change interventions. Finally, by identifying settings where growth mindset interventions are more or less effective, the findings are relevant to the practical question of how to target interventions more strategically.

ABSTRACT: 
Researchers doing pre-election polling have traditionally believed that a random sample of a population includes both voters and non-voters, so predicting the outcome of an election requires ignoring the responses of people who will not vote. The most prominent method for doing so was developed by Paul Perry of the Gallup Organization in the 1970s. This paper reports one of few published evaluations of the effectiveness of this method and comparison with others. Data from the 2008 American National Election Studies (ANES) Time Series Survey (collected via face-to-face interviewing from a probability sample) were used to compare the accuracy of various approaches using three criteria for assessing accuracy: (1) post-election reports of turnout (compared to the government statistic), (2) post-election reports of candidate vote share (compared to the government statistics), and (3) the demographics of voters (compared with results from exit polls). A simple self-report of turnout intentions was a surprisingly successful measure. Furthermore, in an independent replication of recent problems observed by the Gallup Organization, the popular Perry method performed substantially worse than all other methods evaluated. These results may have a clear and useful implication: trust respondents’ self-reports.

ABSTRACT: 
While married mothers and fathers now divide childcare more evenly than in the past, sharp increases in divorce and single parenthood over the past fifty years have led to a growing gender gap in children’s living arrangements in the United States. One major reason why this gap persists is that many people continue to hold traditional views on parenthood, and these expectations play an key role in determining the rewards and punishments mothers and fathers receive for investing time in either parenting or paid labor. In this study, I employ a conjoint survey experiment (N=1,006) to demonstrate the mechanisms leading to a maternal preference in child custody. Findings demonstrate a maternal preference arises in cases where the mother and father have previously divided care evenly. Fathers who deviate from the breadwinner model by staying home are also significantly less likely to be given custody of their children. These findings show that Americans still hold traditional views on parenthood, expecting mothers to act as primary caregivers and fathers to act as providers. The results of this study contribute to a growing literature on how the uneven division of care work contributes to broader economic gender gaps.

ABSTRACT: This talk provides an overview of Deliberative Polling, both past applications and future plans. It will draw informally on my new book Democracy When the People Are Thinking (Oxford 2018). It covers criteria for evaluation and the contexts that have facilitated policy impact. Drawing on cases from 28 countries it will discuss deliberative applications in elections, for constitutional change, and for policy making. Both local and national cases will be discussed. The introduction, the cases in parts III and IV and the speculations about new applications in sections 14 and 15 of Part IV may be of particular interest.