Statistical power of the social network autocorrelation model

    Authors

    W. Wang; E. J. Neuman;D. A. Newman

    Comments

    Authors: contact us about adding a copy of your work at STARS@ucf.edu

    Abbreviated Journal Title

    Soc. Networks

    Keywords

    Network autocorrelation model; Social network analysis; Statistical; power; AUTO-CORRELATION; AUTOREGRESSIVE MODELS; DEPENDENCY TESTS; ESTIMATION; BIAS; PSYCHOLOGY; TOPOLOGY; DYNAMICS; BEHAVIOR; DENSITY; Anthropology; Sociology

    Abstract

    The network autocorrelation model has become an increasingly popular tool for conducting social network analysis. More and more researchers, however, have documented evidence of a systematic negative bias in the estimation of the network effect (rho). In this paper, we take a different approach to the problem by investigating conditions under which, despite the underestimation bias, a network effect can still be detected by the network autocorrelation model. Using simulations, we find that moderately-sized network effects (e.g., rho=.3) are still often detectable in modest-sized networks (i.e., 40 or more nodes). Analyses reveal that statistical power is primarily a nonlinear function of network effect size (rho) and network size (N), although both of these factors can interact with network density and network structure to impair power under certain rare conditions. We conclude by discussing implications of these findings and guidelines for users of the autocorrelation model. (C) 2014 Elsevier B.V. All rights reserved.

    Journal Title

    Social Networks

    Volume

    38

    Publication Date

    1-1-2014

    Document Type

    Article

    Language

    English

    First Page

    88

    Last Page

    99

    WOS Identifier

    WOS:000337015400008

    ISSN

    0378-8733

    Share

    COinS