Estimating peer effects in networks with peer encouragement designs

Heterogeneous peer effects of college roommates on academic performance

Understanding how student peers influence learning outcomes is crucial for effective education management in complex social systems. The complexities of peer selection and evolving peer relationships, however, pose challenges for identifying peer effects using static observational data. Here we use both null-model and regression approaches to examine peer effects using longitudinal data from 5,272 undergraduates, where roommate assignments are plausibly random upon enrollment and roommate relationships persist until graduation. Specifically, we construct a roommate null model by randomly shuffling students among dorm rooms and introduce an assimilation metric to quantify similarities in roommate academic performance. We find significantly larger assimilation in actual data than in the roommate null model, suggesting roommate peer effects, whereby roommates have more similar performance than expected by chance alone. Moreover, assimilation exhibits an overall increasing trend over time, suggesting that peer effects become stronger the longer roommates live together. Our regression analysis further reveals the moderating role of peer heterogeneity. In particular, when roommates perform similarly, the positive relationship between a student's future performance and their roommates' average prior performance is more pronounced, and their ordinal rank in the dorm room has an independent effect. Our findings contribute to understanding the role of college roommates in influencing student academic performance.

Political reinforcement learners

Politics can seem home to the most calculating and yet least rational elements of humanity. How might we systematically characterize this spectrum of political cognition? Here, we propose reinforcement learning (RL) as a unified framework to dissect the political mind. RL describes how agents algorithmically navigate complex and uncertain domains like politics. Through this computational lens, we outline three routes to political differences, stemming from variability in agents' conceptions of a problem, the cognitive operations applied to solve the problem, or the backdrop of information available from the environment. A computational vantage on maladies of the political mind offers enhanced precision in assessing their causes, consequences, and cures.

Heuristics in vaccination Decision-Making for newly developed Vaccines: Understanding the public's imitative behavior

This study aims to investigate the extent to which the public engages in imitative behavior in their vaccination decisions for newly developed vaccines in the Chinese context. Given the crucial role of newly developed vaccines in preventing and controlling the COVID-19 pandemic, a better understanding of how people make decisions about vaccination with new vaccines is important for overcoming vaccine hesitation and promoting widespread adoption of the vaccines. Our results indicate that the public's decision-making about the newly developed vaccine is influenced by a range of heuristics, including a privileged information heuristic, competence heuristic, and consensus heuristic. Specifically, individuals are more likely to imitate the vaccination behavior of those with privileged information, such as insiders, and those with perceived competence in the field, such as experts. Our findings also demonstrate the impact of majority influence, as the popularity of new vaccines leads to an increase in vaccination uptake through herd behavior. Our data highlights the importance of the first movers who are insiders with privileged information or experts with competence, as their behavior can significantly shape the vaccination decisions of others. Our study provides valuable insights into the complex interplay of heuristics and imitative behavior in vaccination decision-making for newly developed vaccines.

Social media feedback and extreme opinion expression

On popular social media platforms such as Twitter, Facebook, Instagram, or Tiktok, the quantitative feedback received by content producers is asymmetric: counts of positive reactions such as 'likes,' or 'retweets,' are easily observed but similar counts of negative reactions are not directly available. We study how this design feature of social media platforms affects the expression of extreme opinions. Using simulations of a learning model, we compare two feedback environments that differ in terms of the availability of negative reaction counts. We find that expressed opinions are generally more extreme when negative reaction counts are not available than when they are. We rely on analyses of Twitter data and several online experiments to provide empirical support for key model assumptions and test model predictions. Our findings suggest that a simple design change might limit, under certain conditions, the expression of extreme opinions on social media.

Interdependent program evaluation: Geographic and social spillovers in COVID-19 closures and reopenings in the United States

In an interconnected world, understanding policy spillovers is essential. We propose a program evaluation framework to measure policy spillover effects and apply that framework to study the governmental responses to COVID-19 in the United States. Our analysis suggests the presence of social spillovers. We estimate that while state closures directly reduced mobility by 3 to 4%, all other states locking down further decreased mobility in the focal state by 8 to 14%. Similarly, while reopening directly increased mobility by 2 to 3%, all other states' reopening increased mobility in the focal state by 12 to 21%. Our analysis also suggests geographic spillovers: Travel from locked down origins to open destinations increased by 12 to 29%. In contrast, travel from reopened origins to locked down destinations decreased by 6 to 7% for nearby counties and by 14 to 18% for distant counties. Despite its limitations, we believe that our approach takes the first steps toward creating a framework for interdependent program evaluation across policy domains.

A computational reward learning account of social media engagement

Social media has become a modern arena for human life, with billions of daily users worldwide. The intense popularity of social media is often attributed to a psychological need for social rewards (likes), portraying the online world as a Skinner Box for the modern human. Yet despite such portrayals, empirical evidence for social media engagement as reward-based behavior remains scant. Here, we apply a computational approach to directly test whether reward learning mechanisms contribute to social media behavior. We analyze over one million posts from over 4000 individuals on multiple social media platforms, using computational models based on reinforcement learning theory. Our results consistently show that human behavior on social media conforms qualitatively and quantitatively to the principles of reward learning. Specifically, social media users spaced their posts to maximize the average rate of accrued social rewards, in a manner subject to both the effort cost of posting and the opportunity cost of inaction. Results further reveal meaningful individual difference profiles in social reward learning on social media. Finally, an online experiment (n = 176), mimicking key aspects of social media, verifies that social rewards causally influence behavior as posited by our computational account. Together, these findings support a reward learning account of social media engagement and offer new insights into this emergent mode of modern human behavior.

Realistic modelling of information spread using peer-to-peer diffusion patterns

In computational social science, epidemic-inspired spread models have been widely used to simulate information diffusion. However, recent empirical studies suggest that simple epidemic-like models typically fail to generate the structure of real-world diffusion trees. Such discrepancy calls for a better understanding of how information spreads from person to person in real-world social networks. Here, we analyse comprehensive diffusion records and associated social networks in three distinct online social platforms. We find that the diffusion probability along a social tie follows a power-law relationship with the numbers of disseminator's followers and receiver's followees. To develop a more realistic model of information diffusion, we incorporate this finding together with a heterogeneous response time into a cascade model. After adjusting for observational bias, the proposed model reproduces key structural features of real-world diffusion trees across the three platforms. Our finding provides a practical approach to designing more realistic generative models of information diffusion.

Use of a controlled experiment and computational models to measure the impact of sequential peer exposures on decision making

It is widely believed that one's peers influence product adoption behaviors. This relationship has been linked to the number of signals a decision-maker receives in a social network. But it is unclear if these same principles hold when the "pattern" by which it receives these signals vary and when peer influence is directed towards choices which are not optimal. To investigate that, we manipulate social signal exposure in an online controlled experiment using a game with human participants. Each participant in the game decides among choices with differing utilities. We observe the following: (1) even in the presence of monetary risks and previously acquired knowledge of the choices, decision-makers tend to deviate from the obvious optimal decision when their peers make a similar decision which we call the influence decision, (2) when the quantity of social signals vary over time, the forwarding probability of the influence decision and therefore being responsive to social influence does not necessarily correlate proportionally to the absolute quantity of signals. To better understand how these rules of peer influence could be used in modeling applications of real world diffusion and in networked environments, we use our behavioral findings to simulate spreading dynamics in real world case studies. We specifically try to see how cumulative influence plays out in the presence of user uncertainty and measure its outcome on rumor diffusion, which we model as an example of sub-optimal choice diffusion. Together, our simulation results indicate that sequential peer effects from the influence decision overcomes individual uncertainty to guide faster rumor diffusion over time. However, when the rate of diffusion is slow in the beginning, user uncertainty can have a substantial role compared to peer influence in deciding the adoption trajectory of a piece of questionable information.

Testing for arbitrary interference on experimentation platforms

Experimentation platforms are essential to large modern technology companies, as they are used to carry out many randomized experiments daily. The classic assumption of no interference among users, under which the outcome for one user does not depend on the treatment assigned to other users, is rarely tenable on such platforms. Here, we introduce an experimental design strategy for testing whether this assumption holds. Our approach is in the spirit of the Durbin–Wu–Hausman test for endogeneity in econometrics, where multiple estimators return the same estimate if and only if the null hypothesis holds. The design that we introduce makes no assumptions on the interference model between units, nor on the network among the units, and has a sharp bound on the variance and an implied analytical bound on the Type I error rate. We discuss how to apply the proposed design strategy to large experimentation platforms, and we illustrate it in the context of an experiment on the LinkedIn platform.

A Trial of Student Self-Sponsored Peer-to-Peer Lending Based on Credit Evaluation Using Big Data Analysis

There is still no effective approach to overcome the problem of credit evaluation for Chinese students. In absence of a reliable credit evaluation system for students, the university students have to only apply through online peer-to-peer (P2P) loan platforms because Chinese financial institutions typically reject students' loan applications. Lack of students' financial records hinders financial institutes and banks to routinely evaluate the students' credit status and assign loans to them. Hence, this paper attempted to benefit from university students' diversified daily behavior data, and logistic regression (LR) and gradient boosting decision tree (GBDT) algorithms were also used to develop robust credit evaluation models for university students, in which the validation of the proposed models was assessed by a real-time P2P lending platform. In this study, the students' overdue behavior in returning books to university library was used as an index. With training 17838 samples, the proposed models performed well, while GBDT-based model outperformed in identification of “bad borrowers.” Based on the proposed models, a self-sponsored peer-to-peer loan platform was established and developed in a Chinese university for ten months, and the achieved findings demonstrated that adopting such credit evaluation models can effectively reduce the default ratio.

Big data uncertainties

Big data-the idea that an always-larger volume of information is being constantly recorded-suggests that new problems can now be subjected to scientific scrutiny. However, can classical statistical methods be used directly on big data? We analyze the problem by looking at two known pitfalls of big datasets. First, that they are biased, in the sense that they do not offer a complete view of the populations under consideration. Second, that they present a weak but pervasive level of dependence between all their components. In both cases we observe that the uncertainty of the conclusion obtained by statistical methods is increased when used on big data, either because of a systematic error (bias), or because of a larger degree of randomness (increased variance). We argue that the key challenge raised by big data is not only how to use big data to tackle new problems, but to develop tools and methods able to rigorously articulate the new risks therein.

Affective stimuli in behavioural interventions soliciting for health check-up services and the service users' socioeconomic statuses: a study at Japanese pachinko parlours

Editor’s note

The study reported in this article examines a health intervention that uses gendered stereotypes of the nursing profession and suggestive uniforms that play on women’s sexuality to encourage people to engage in health checkups. The intervention was not under the control of the authors and the study was approved by an institutional research ethics board. The Journal of Epidemiology & Community Health condemns the use of sexism, gender and professional stereotypes and other forms of discriminatory or exploitative behaviour for any purpose, including health promotion programs. In light of concerns raised about this paper (see eLetters with this paper), we are conducting an audit of our review process and will put in place measures to ensure that the material we publish condemns sexism, racism and other forms of discrimination and embodies principles of inclusion and non-discrimination.

Reviewer bias in single- versus double-blind peer review

Scientific peer review has been a cornerstone of the scientific method since the 1600s. Debate continues regarding the merits of single-blind review, in which anonymous reviewers know the authors of a paper and their affiliations, compared with double-blind review, in which this information is hidden. We present an experimental study of this question. In computer science, research often appears first or exclusively in peer-reviewed conferences rather than journals. Our study considers full-length submissions to the highly selective 2017 Web Search and Data Mining conference (15.6% acceptance rate). Each submission is simultaneously scored by two single-blind and two double-blind reviewers. Our analysis shows that single-blind reviewing confers a significant advantage to papers with famous authors and authors from high-prestige institutions.

Peer review may be “single-blind,” in which reviewers are aware of the names and affiliations of paper authors, or “double-blind,” in which this information is hidden. Noting that computer science research often appears first or exclusively in peer-reviewed conferences rather than journals, we study these two reviewing models in the context of the 10th Association for Computing Machinery International Conference on Web Search and Data Mining, a highly selective venue (15.6% acceptance rate) in which expert committee members review full-length submissions for acceptance. We present a controlled experiment in which four committee members review each paper. Two of these four reviewers are drawn from a pool of committee members with access to author information; the other two are drawn from a disjoint pool without such access. This information asymmetry persists through the process of bidding for papers, reviewing papers, and entering scores. Reviewers in the single-blind condition typically bid for 22% fewer papers and preferentially bid for papers from top universities and companies. Once papers are allocated to reviewers, single-blind reviewers are significantly more likely than their double-blind counterparts to recommend for acceptance papers from famous authors, top universities, and top companies. The estimated odds multipliers are tangible, at 1.63, 1.58, and 2.10, respectively.