scholarly journals Spillover Effects in the Presence of Unobserved Networks

2020 ◽  
pp. 1-30
Author(s):  
Naoki Egami
Keyword(s):  
Sensitivity Analysis ◽  
Spillover Effect ◽  
Spillover Effects ◽  
Unbiased Estimation ◽  
Potential Influence ◽  
Face To Face ◽  
Multiple Networks ◽  
Omitted Variable Bias ◽  
Experimental Subjects ◽  
Variable Bias

Abstract When experimental subjects can interact with each other, the outcome of one individual may be affected by the treatment status of others. In many social science experiments, such spillover effects may occur through multiple networks, for example, through both online and offline face-to-face networks in a Twitter experiment. Thus, to understand how people use different networks, it is essential to estimate the spillover effect in each specific network separately. However, the unbiased estimation of these network-specific spillover effects requires an often-violated assumption that researchers observe all relevant networks. We show that, unlike conventional omitted variable bias, bias due to unobserved networks remains even when treatment assignment is randomized and when unobserved networks and a network of interest are independently generated. We then develop parametric and nonparametric sensitivity analysis methods, with which researchers can assess the potential influence of unobserved networks on causal findings. We illustrate the proposed methods with a simulation study based on a real-world Twitter network and an empirical application based on a network field experiment in China.

scholarly journals The Impossible Mediation Test: A Method for Dealing with Plausibly Confounded Mediation

2017 ◽  
Author(s):  
David Scott Yeager ◽  
Jon Krosnick
Keyword(s):  
Sensitivity Analysis ◽  
Correlational Analysis ◽  
Causal Process ◽  
Psychology Research ◽  
Strong Basis ◽  
Spurious Association ◽  
Omitted Variable Bias ◽  
Variable Bias ◽  
As If

In much psychology research, mediators are measured, not manipulated. Therefore, the paths from mediators to outcomes—the so-called b paths—can be confounded by omitted variable bias, as in any other correlational analysis. The present research builds on the logic of falsification tests in econometrics and sensitivity analysis in statistics to propose the impossible mediation test, which can quantify the amount of confounded mediation. Researchers can add to an experiment a condition in which the dependent variable is measured first, then the manipulation is implemented, and finally, the posited mediator is measured. This allows for assessment of the spurious association between the dependent variable and the mediator, and statistics can be estimated as if the dependent variable was measured after the manipulation was implemented, to assess whether the spurious association is sufficiently strong to yield the false appearance of mediation. This estimate of “impossible” mediation can be compared to the results obtained from data where the dependent variable is actually measured in the conventional order after the mediator, to determine whether evidence of mediation is stronger in the latter case than the former. Evidence of mediation that survives the impossible mediation test constitutes a strong basis for a claim about mediation of a causal process. The paper illustrates this procedure with an empirical example.

scholarly journals Identification of average effects under magnitude and sign restrictions on confounding

10.3982/qe689 ◽  
2019 ◽  
Vol 10 (4) ◽  
pp. 1619-1657 ◽  
Author(s):  
Karim Chalak
Keyword(s):  
Sensitivity Analysis ◽  
Direct Effect ◽  
Instrumental Variables ◽  
Average Treatment Effect ◽  
Structural Systems ◽  
Omitted Variables ◽  
Average Effect ◽  
Omitted Variable Bias ◽  
Marginal Effects ◽  
Variable Bias

This paper studies measuring various average effects of X on Y in general structural systems with unobserved confounders U, a potential instrument Z, and a proxy W for U. We do not require X or Z to be exogenous given the covariates or W to be a perfect one‐to‐one mapping of U. We study the identification of coefficients in linear structures as well as covariate‐conditioned average nonparametric discrete and marginal effects (e.g., average treatment effect on the treated), and local and marginal treatment effects. First, we characterize the bias, due to the omitted variables U, of (nonparametric) regression and instrumental variables estimands, thereby generalizing the classic linear regression omitted variable bias formula. We then study the identification of the average effects of X on Y when U may statistically depend on X and Z. These average effects are point identified if the average direct effect of U on Y is zero, in which case exogeneity holds, or if W is a perfect proxy, in which case the ratio (contrast) of the average direct effect of U on Y to the average effect of U on W is also identified. More generally, restricting how the average direct effect of U on Y compares in magnitude and/or sign to the average effect of U on W can partially identify the average effects of X on Y. These restrictions on confounding are weaker than requiring benchmark assumptions, such as exogeneity or a perfect proxy, and enable a sensitivity analysis. After discussing estimation and inference, we apply this framework to study earnings equations.

Effects of tourism development on economic growth: An empirical study of China based on both static and dynamic spatial Durbin models

2021 ◽  
pp. 135481662110211
Author(s):  
Honghong Liu ◽  
Ye Xiao ◽  
Bin Wang ◽  
Dianting Wu
Keyword(s):  
Economic Growth ◽  
Spillover Effect ◽  
Tourism Development ◽  
Spillover Effects ◽  
Policy Implications ◽  
Long Term Effects ◽  
Spatial Spillover ◽  
Domestic Tourism ◽  
Inbound Tourism ◽  
Spatial Spillover Effect

This study applies the dynamic spatial Durbin model (SDM) to explore the direct and spillover effects of tourism development on economic growth from the perspective of domestic and inbound tourism. The results are compared with those from the static SDM. The results support the tourism-led-economic-growth hypothesis in China. Specifically, domestic tourism and inbound tourism play a significant role in stimulating local economic growth. However, the spatial spillover effect is limited to domestic tourism, and the spatial spillover effect of inbound tourism is not significant. Furthermore, the long-term effects are much greater than the short-term impact for both domestic and inbound tourism. Plausible explanations of these results are provided and policy implications are drawn.

scholarly journals Regional Disparity and Patients Mobility: Benefits and Spillover Effects of the Spatial Network Structure of the Health Services in China

2021 ◽  
Vol 18 (3) ◽  
pp. 1096
Author(s):  
Liping Fu ◽  
Kaibo Xu ◽  
Feng Liu ◽  
Lu Liang ◽  
Zhengmin Wang
Keyword(s):  
Health Economic ◽  
Health Services ◽  
Network Structure ◽  
Spillover Effect ◽  
Spillover Effects ◽  
Health Economy ◽  
Spatial Network ◽  
Spatial Effects ◽  
Medical Resource ◽  
Medical Resources

Background: The distribution of medical resources in China is seriously imbalanced due to imbalanced economic development in the country; unbalanced distribution of medical resources makes patients try to seek better health services. Against this backdrop, this study aims to analyze the spatial network characteristics and spatial effects of China’s health economy, and then find evidence that affects patient mobility. Methods: Data for this study were drawn from the China Health Statistical Yearbooks and China Statistical Books. The gravitational value of China’s health spatial network was calculated to establish a network of gravitational relationships. The social network analysis method was used for centrality analysis and spillover effect analysis. Results: A gravity correlation matrix was constructed among provinces by calculating the gravitational value, indicating the spatial relationships of different provinces in the health economic network. Economically developed provinces, such as Shanghai and Jiangsu, are at the center of the health economic network (centrality degree = 93.333). These provinces also play a strong intermediary role in the network and have connections with other provinces. In the CONCOR analysis, 31 provinces are divided into four blocks. The spillover effect of the blocks indicates provinces with medical resource centers have beneficial effects, while provinces with insufficient resources have obvious spillover effects. Conclusion: There is a significant gap in the geographical distribution of medical resources, and the health economic spatial network structure needs to be improved. Most medical resources are concentrated in economically developed provinces, and these provinces’ positions in the health economic spatial network are becoming more centralized. By contrast, economically underdeveloped regions are at the edge of the network, causing patients to move to provinces with medical resource centers. There are health risks of the increasing pressure to seek medical treatment in developed provinces with abundant medical resources.

scholarly journals Spatial Spillover Effect and Influencing Factors of Information Flow in Urban Agglomerations—Case Study of China Based on Baidu Search Index

2021 ◽  
Vol 13 (14) ◽  
pp. 8032
Author(s):  
Chengzhuo Wu ◽  
Li Zhuo ◽  
Zhuo Chen ◽  
Haiyan Tao
Keyword(s):  
Influencing Factors ◽  
Information Flow ◽  
Spillover Effect ◽  
Spillover Effects ◽  
Urban Agglomeration ◽  
Coordinated Development ◽  
Spatial Spillover ◽  
Urban Agglomerations ◽  
Search Index ◽  
Spatial Spillover Effect

Cities in an urban agglomeration closely interact with each other through various flows. Information flow, as one of the important forms of urban interactions, is now increasingly indispensable with the fast development of informatics technology. Thanks to its timely, convenient, and spatially unconstrained transmission ability, information flow has obvious spillover effects, which may strengthen urban interaction and further promote urban coordinated development. Therefore, it is crucial to quantify the spatial spillover effect and influencing factors of information flows, especially at the urban agglomeration scale. However, the academic research on this topic is insufficient. We, therefore, developed a spatial interaction model of information flow (SIM-IF) based on the Baidu Search Index and used it to analyze the spillover effects and influencing factors of information flow in the three major urban agglomerations in China, namely Beijing–Tianjin–Hebei (BTH), the Yangtze River Delta (YRD), and the Pearl River Delta (PRD) in the period of 2014–2019. The results showed that the SIM-IF performed well in all three agglomerations. Quantitative analysis indicated that the BTH had the strongest spillover effect of information flow, followed by the YRD and the PRD. It was also found that the hierarchy of cities had the greatest impact on the spillover effects of information flow. This study may provide scientific basis for the information flow construction in urban agglomerations and benefit the coordinated development of cities.

scholarly journals Omitted Variable Bias in GLMs of Neural Spiking Activity

2018 ◽  
Vol 30 (12) ◽  
pp. 3227-3258 ◽  
Author(s):  
Ian H. Stevenson
Keyword(s):  
Neural Activity ◽  
Single Neuron ◽  
Linear Models ◽  
Theory And Practice ◽  
Parameter Estimates ◽  
Omitted Variables ◽  
History Effects ◽  
Omitted Variable Bias ◽  
Wide Range ◽  
Variable Bias

Generalized linear models (GLMs) have a wide range of applications in systems neuroscience describing the encoding of stimulus and behavioral variables, as well as the dynamics of single neurons. However, in any given experiment, many variables that have an impact on neural activity are not observed or not modeled. Here we demonstrate, in both theory and practice, how these omitted variables can result in biased parameter estimates for the effects that are included. In three case studies, we estimate tuning functions for common experiments in motor cortex, hippocampus, and visual cortex. We find that including traditionally omitted variables changes estimates of the original parameters and that modulation originally attributed to one variable is reduced after new variables are included. In GLMs describing single-neuron dynamics, we then demonstrate how postspike history effects can also be biased by omitted variables. Here we find that omitted variable bias can lead to mistaken conclusions about the stability of single-neuron firing. Omitted variable bias can appear in any model with confounders—where omitted variables modulate neural activity and the effects of the omitted variables covary with the included effects. Understanding how and to what extent omitted variable bias affects parameter estimates is likely to be important for interpreting the parameters and predictions of many neural encoding models.

Sign in / Sign up

Export Citation Format

Share Document