Collaborative Research: Empirical Analysis of Social Networks with Unreported Links

点击翻译按钮获取中文摘要

In many social-economic contexts, an individual's behavior depend on his own characteristics, as well as the outcome and characteristics of others. Such dependence called a link; individuals with links are neighbors and a collection of neighbors is referred to as a network. A social network consists of linked individuals. This commonly occurs in applied economic research since links are often not well measured in the research data. This project will estimate the effects of social networks on individual outcomes when the links are either misclassified or not reported in data. The method proposed in this project is adaptable to a wide range of social networks. It also provide a general method for comparing various types of social effects given group characteristics. The project offers an efficient approach for policy analyses that resolves challenges due to data problems or measurement errors in network links. The results of this project provides a way to measure the effects of policies when there is no information on network structure. The results of this project will have a significant impact on empirical research on social networks and policies such as education. This will improve efficiency in business and policy decision making and in the process lead to improved well-being of U.S. citizens.This project identifies and estimates social network models when network links are either misclassified or unobserved. It first derives and characterizes conditions under which some misclassification of links does not interfere with the consistency or asymptotic properties of standard instrumental variable estimators of social effects. It then constructs a consistent estimator of social effects in a model where network links are not observed. This method does not require repeated observations of individual network members. The project will apply this estimator to data from Tennessee's Student/Teacher Achievement Ratio (STAR) Project. Without observing the latent network in each classroom, the research identifies and estimate peer and contextual effects on students' performance in mathematics. The results suggests that peer effects tend to be larger in bigger classes, and that increasing peer effects significantly improve students' average test scores. The results of this research will help businesses and policy makers account for social effects in decision making hence improve the living standards of U.S. citizens.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.