Randomized Experiments to Detect and Estimate Social Influence in Networks

Taylor, Sean J.; Eckles, Dean

doi:10.1007/978-3-319-77332-2_16

Cited by 20 publications

(12 citation statements)

References 123 publications

(149 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, in many social, medical, and online settings the no-interference assumption may fail to hold (Rosenbaum 2007;Walker and Muchnik 2014;Aral 2016;Taylor and Eckles 2017). In such settings, peer and spillover effects can bias estimates of the global average treatment effect.…”

Section: Introductionmentioning

confidence: 99%

Regression Adjustments for Estimating the Global Treatment Effect in Experiments with Interference

Chin

2019

Journal of Causal Inference

View full text Add to dashboard Cite

Standard estimators of the global average treatment effect can be biased in the presence of interference. This paper proposes regression adjustment estimators for removing bias due to interference in Bernoulli randomized experiments. We use a fitted model to predict the counterfactual outcomes of global control and global treatment. Our work differs from standard regression adjustments in that the adjustment variables are constructed from functions of the treatment assignment vector, and that we allow the researcher to use a collection of any functions correlated with the response, turning the problem of detecting interference into a feature engineering problem. We characterize the distribution of the proposed estimator in a linear model setting and connect the results to the standard theory of regression adjustments under SUTVA. We then propose an estimator that allows for flexible machine learning estimators to be used for fitting a nonlinear interference functional form. We propose conducting statistical inference via bootstrap and resampling methods, which allow us to sidestep the complicated dependences implied by interference and instead rely on empirical covariance structures. Such variance estimation relies on an exogeneity assumption akin to the standard unconfoundedness assumption invoked in observational studies. In simulation experiments, our methods are better at debiasing estimates than existing inverse propensity weighted estimators based on neighborhood exposure modeling. We use our method to reanalyze an experiment concerning weather insurance adoption conducted on a collection of villages in rural China.

show abstract

Section: Introductionmentioning

confidence: 99%

Regression Adjustments for Estimating the Global Treatment Effect in Experiments with Interference

Chin

2019

Journal of Causal Inference

View full text Add to dashboard Cite

show abstract

“…Similar to offline field experiments, field experiments conducted on social media involve engaging with people's natural activities to investigate the impact of a treatment 'in the wild' -and as a result, there are important ethical considerations to take into account, above and beyond receiving approval from academic Institutional Review Boards (Desposato, 2015;Gallego, Martinez, Munger, & Vásquez-Cortés, 2019;Taylor & Eckles, 2018).…”

Section: Ethical Considerationsmentioning

confidence: 99%

Field Experiments on Social Media

Mosleh

Pennycook

Rand

2021

Curr Dir Psychol Sci

View full text Add to dashboard Cite

Online behavioral data, such as digital traces from social media, have the potential to allow researchers an unprecedented new window into human behavior in ecologically valid everyday contexts. However, research using such data is often purely observational, which limits its usefulness for identifying causal relationships. Here we review recent innovations in experimental approaches to studying online behavior, with a particular focus on research related to misinformation and political psychology. In hybrid lab-field studies, exposure to social-media content can be randomized, and the impact on attitudes and beliefs can be measured using surveys, or exposure to treatments can be randomized within survey experiments, and their impact on subsequent online behavior can be observed. In field experiments conducted on social media, randomized treatments can be administered directly to users in the online environment (e.g., via social-tie invitations, private messages, or public posts) without revealing that they are part of an experiment, and the effects on subsequent online behavior can then be observed. The strengths and weaknesses of each approach are discussed, along with practical advice and central ethical constraints on such studies.

show abstract

Section: Ethical Considerationsmentioning

confidence: 99%

Field experiments on social media

Mosleh¹,

Pennycook²,

Rand³

2021

Preprint

View full text Add to dashboard Cite

Online behavioral data, such as digital traces from social media, have the potential to allow researchers an unprecedented new window into human behavior in ecologically valid everyday contexts. However, research using such data is often purely observational, limiting its ability to identify causal relationships. Here we review recent innovations in experimental approaches to studying online behavior, with a particular focus on research related to misinformation and political psychology. In hybrid lab-field studies, exposure to social media content can be randomized, and the impact on attitudes and beliefs measured using surveys; or exposure to treatments can be randomized within survey experiments, and their impact observed on subsequent online behavior. In field experiments conducted on social media, randomized treatments can be administered directly to users in the online environment - e.g. via social tie invitations, private messages, or public posts - without revealing that they are part of an experiment, and the impacts on subsequent online behavior observed. The strengths and weaknesses of each approach are discussed, along with practical advice and central ethical constraints on such studies.

show abstract

Randomized Experiments to Detect and Estimate Social Influence in Networks

Cited by 20 publications

References 123 publications

Regression Adjustments for Estimating the Global Treatment Effect in Experiments with Interference

Regression Adjustments for Estimating the Global Treatment Effect in Experiments with Interference

Field Experiments on Social Media

Field experiments on social media

Contact Info

Product

Resources

About