Abstract
In many contexts we may be interested in understanding whether direct connections between agents, such as declared friendships in a classroom or family links in a rural village, affect their outcomes. In this paper, we review the literature studying econometric methods for the analysis of linear models of social effects, a class that includes the ‘linear-in-means’ local average model, the local aggregate model, and models where network statistics affect outcomes. We provide an overview of the underlying theoretical models, before discussing conditions for identification using observational and experimental/quasi-experimental data.
Highlights
Researchers and policymakers are often interested in identifying whether and the extent to which direct connections between agents affect their outcomes
In the absence of information on interactions within a network, identification of social effect parameters is greatly complicated by the so-called reflection problem, a form of simultaneity where it is not possible to identify who influences whom within the network or reference group (Manski 1993)
In this paper, we provide an overview of methods to identify social effects in linear social effect models using a single cross section of data
Summary
Researchers and policymakers are often interested in identifying whether and the extent to which direct connections between agents affect their outcomes. The individual heterogeneity parameter, πi,g(Xg, Gi,g), can be modelled as a linear function of observed and unobserved individual and network characteristics: Ng πi,g Xg , Gi,g = xi,g γ + Gij,g xj,g δ + zg η + νg + εi,g j=1 Substituting for this in Eq 3, we obtain the following best response function for individual outcomes: Ng yi,g = β Gij,g yj,g +xi,g γ + Gij,g xj,g δ+zg η+νg +εi,g j=1 j=1. N πi,g Xg , Gi,g = xi,g δ + Gij,g xj,g γ + zg η + νg + εi,g j=1 so that individual heterogeneity is a function of a node’s own characteristics, the average characteristics of its neighbours, network-level observed characteristics, and some unobserved network- and individual-level terms This model shares some features with the model of Blume et al (2015), as different network matrices are used to capture the effects of neighbours’ outcomes and characteristics, which helps to ease identification. To account for the fixed effect, a within-transformation is applied, as in the “Local average models” section
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
