This study shows that residual variation can cause problems related to scaling in exponential random graph models (ERGM). Residual variation is likely to exist when there are unmeasured variables in a model—even those uncorrelated with other predictors—or when the logistic form of the model is inappropriate. As a consequence, coefficients cannot be interpreted as effect sizes or compared between models and homophily coefficients, as well as other interaction coefficients, cannot be interpreted as substantive effects in most ERGM applications. We conduct a series of simulations considering the substantive impact of these issues, revealing that realistic levels of residual variation can have large consequences for ERGM inference. A flexible methodological framework is introduced to overcome these problems. Formal tests of mediation and moderation are also proposed. These methods are applied to revisit the relationship between selective mixing and triadic closure in a large AddHealth school friendship network. Extensions to other classes of statistical work models are discussed.
adamsjimiSchaeferDavid R.. 2016. “How Initial Prevalence Moderates Network-based Smoking Change: Estimating Contextual Effects with Stochastic Actor-based Models.” Journal of Health and Social Behavior57:22–38.
2.
AgrestiAlan. 2002. Categorical Data Analysis. New York: Wiley.
3.
AiChunrongNortonEdward C.. 2003. “Interaction Terms in Logit and Probit Models.” Economics Letters80:123–29.
4.
AllisonPaul D.1999. “Comparing Logit and Probit Coefficients across Groups.” Sociological Methods and Research28:186–208.
5.
BesagJulian E.1972. “Nearest-neighbour Systems and the Auto-logistic Model for Binary Data.” Journal of the Royal Statistical Society Series B34:75–83.
6.
BesagJulian E.1974. “Spatial Interaction and the Statistical Analysis of Lattice Systems.” Journal of the Royal Statistical Society Series B36:192–236.
7.
BollenKenneth A.StineRobert. 1990. “Direct and Indirect Effects: Classical and Bootstrap Estimates of Variability.” Sociological Methodology20:115–40.
8.
Box-SteffensmeierJanet M.ChristensonDino P.MorganJason W.. 2018. “Modeling Unobserved Heterogeneity in Social Networks with the Frailty Exponential Random Graph Model.” Political Analysis26:3–19.
9.
BreenRichardKarlsonKristian BerntHolmAnders. 2013. “Total, Direct, and Indirect Effects in Logit and Probit Models.” Sociological Methods and Research42:164–91.
10.
ButtsCarter T.2008. “A Relational Event Framework for Social Action.” Sociological Methodology38:155–200.
11.
CramerJan Salomon. 2007. “Robustness of Logit Analysis: Analysis for Unobserved Heterogeneity and Mis-specified Disturbances.” Oxford Bulletin of Economics and Statistics69:545–55.
12.
CranmerSkyler J.LeifeldPhilipMcClurgScott D.RolfeMeredith. 2017. “Navigating the Range of Statistical Tools for Inferential Network Analysis.” American Journal of Political Science61:237–51.
13.
DesmaraisBruce A.CranmerSkyler J.. 2012. “Statistical Inference for Valued-edge Networks: The Generalized Exponential Random Graph Model.” PLoS One7:e30136.
14.
DumitrescuElena IvonaHurlinChristophe. 2012. “Testing for Granger Non-causality in Heterogenous Panels.” Economics Modeling29:1450–60.
15.
DuxburyScott. 2019. ergMargins: Process Analysis for Exponential Random Graph Models. Comprehensive R Archive Network.
16.
ErdosPaulRenyiAlfred. 1959. “On Random Graphs.” Publicationes Mathematicae6:290–97.
17.
FrankOveStraussDavid. 1986. “Markov Graphs.” Journal of the American Statistical Association81:832–42.
18.
GelmanAndrewSternHal. 2006. “The Difference between “Significant” and “Not Significant” Is Not Itself Statistically Significant.” The American Statistician60:328–31.
19.
GeyerCharles J.ThompsonElizabeth A.. 1992. “Constrained Monte Carlo Maximum Likelihood for Dependent Data.” Journal of the Royal Statistical Society B54:657–99.
20.
GoodreauSteven M.KittsJames A.MorrisMartina. 2009. “Birds of a Feather, or Friend of a Friend? Using Exponential Random Graph Models to Investigate Adolescent Social Networks.” Demography46:103–25.
21.
HandcockMark S.RobinsGarrySnijdersTom A. B.MoodyJimBesagJulian. 2003. “Assessing Degeneracy in Statistical Models of Social Networks.” Journal of the American Statistical Association76:33–50.
22.
HannekeSteveFuWenjieXingEric P.. 2010. “Discrete Temporal Models of Social Networks.” Electronic Journal of Statistics4:585–605.
23.
HollandPaul W.LeinhardtSamuel. 1981. “An Exponential Family of Probability Distributions for Directed Graphs.” Journal of the American Statistical Association76:33–50.
HunterDavid R.2007. “Curved Exponential Family Models for Social Networks.” Social Networks29:216–30.
26.
KarlsonKristian BerntHolmAndersBreenRichard. 2012. “Comparing Regression Coefficients between Same-sample Nested Models Using Logit and Probit: A New Method.” Sociological Methodology42:286–313.
27.
KimYeajiLeonardoAntenangeliKirklandJustin. 2016. “Measurement Error and Attenuation Bias in Exponential Random Graph Models.” Statistics, Politics, and Policy7:29–54.
28.
KoskinenJohanDaraganovaGalina. 2013. “Dependence Graphs and Sufficient Statistics.” Pp. 77–90in Exponential Random Graph Models for Social Networks, edited by LusherDeanKoskinenJohanRobinsGarry, chapter 7. Cambridge: Cambridge University Press.
29.
KreagerDerek A.YoungJacob T. N.HaynieDana L.BouchardMartinSchaeferDavid R.ZajacGary. 2017. “Where “Old Heads” Prevail: Inmate Hierarchy in a Men’s Prison Unit.” American Sociological Review82:685–718.
30.
KrivitskyPavel N.2012. “Exponential-family Random Graph Models for Valued Networks.” Electronic Journal of Statistics6:1100–1128.
31.
LewisKevin. 2013. “The Limits of Racial Prejudice.” Proceedings of the National Academy of Sciences110:18814–819.
32.
LongJ. Scott. 1997. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage.
33.
LongJ. ScottMustilloSarah A.. 2018. “Using Predictions and Marginal Effects to Compare Groups in Regression Models for Binary Outcomes.” Sociological Methods & Research. doi:10.1177/0049124118799374.
34.
LusherDeanKoskinenJohanRobinsGarry. 2013. Exponential Random Graph Models for Social Networks. Cambridge: Cambridge University Press.
35.
MackinnonDavid P.2008. Introduction to Statistical Mediation Analysis. Abingdon, UK: Routledge.
36.
MackinnonDavid P.LockwoodChondra M.BrownHendricksWangWei. 2007. “The Intermediate Endpoint Effect in Logistic and Probit Regression.” Clinical Trials4:499–513.
37.
MackinnonDavid P.LockwoodChondra M.WilliamsJason. 2004. “Confidence Limits for the Indirect Effects: Distribution of the Product and Resampling Methods.” Multivariate Behavioral Research39:99–128.
38.
MeleAngelo. 2017. “A Structural Model of Dense Network Formation.” Econometrica85:825–50.
39.
MizeTrentonDoanLongScott LongJ.. 2019. “A General Framework for Comparing Predictions and Marginal Effects across Models.” Sociological Methodology49(1):1–38.
40.
MoodCarina. 2010. “Logistic Regression: Why We Cannot Do What We Think We Can Do, and What We Can Do about It.” European Sociological Review26:67–82.
41.
NewmanMark E.J.2010. Networks: An Introduction. Oxford, UK: Oxford University Press.
42.
PapachristosAndrew V.BastomskiSara. 2018. “Connected in Crime: The Enduring Effect of Neighborhood Networks on the Spatial Patterning of Violence.” American Journal of Sociology124:517–68.
43.
PapachristosAndrew V.HureauDavid M.BragaAnthony A.. 2013. “The Corner and the Crew: The Influence of Geography and Social Networks on Gang Violence.” American Sociological Review78:417–47.
44.
SchweinbergerMichael. 2020. “Consistent Structure Estimation of Exponential Family Random Graph Models with Block Structure.” Bernoulli26:1205–33.
45.
SchweinbergerMichaelHandcockMark S.. 2015. “Local Dependence in Random Graph Models: Characterization, Properties, and Statistical Inference.” Journal of the Royal Statistical Society, Series B77:647–76.
46.
SnijdersTom A. B.2001. “The Statistical Evaluation of Social Network Dynamics.” Sociological Methodology31:361–95.
47.
SnijdersTom A. B.2002. “Markov Chain Monte Carlo Estimation of Exponential Random Graph Models.” Journal of Social Structure3: 2–37.
48.
SnijdersTom A.B.PattisonPhillipa E.RobinsGarry L.HandcockMark S.. 2006. “New Specifications for Exponential Random Graph Models.” Sociological Methodology36:99–150.
49.
SobelMichael E.1986. “Direct and Indirect Effects in Linear Structural Equation Models.” Sociological Methods and Research16:155–76.
50.
StewartJonathanSchweinbergerMichaelBojanowskiMichalMorrisMartina. 2019. “Multilevel Network Data Facilitate Statistical Inference for Curved ERGMs with Geometrically Weighted Terms.” Social Networks59:98–119.
51.
ThiemichenS.FrielN.CaimoA.KauermannG.. 2016. “Bayesian Exponential Random Graph Models with Nodal Random Effects.” Social Networks46:11–28.
52.
van DuijnMarijtje A.SnijdersTom A. B.ZijlstraBonne J. H.. 2004. “p2: A Random Effects Model with Covariates for Directed Graphs.” Statistica Neerlandica58:234–54.
53.
WassermanStanPattersonPhillipa. 1996. “Logit Models and Logistic Regressions for Social Networks I: An Introduction to Markov Graphs and p*.” Psychometrika61:401–25.
54.
WattsDuncan J.StrogatzSteven H.. 1998. “Collective Dynamics of ‘Small-world’ Networks.” Nature393:440–42.
55.
WeesieJeroen. 1999. “Seemingly Unrelated Estimation and Cluster-adjusted Sandwich Estimator.” Stata Technical Bulletin9:231–48.
56.
WimmerAndreasLewisKevin. 2010. “Beyond and Below Racial Homophily: ERG Models of a Friendship Network Documented on Facebook.” American Journal of Sociology2:583–642.
57.
WinshipChristopherMareRobert D.. 1983. “Structural Equations and Path Analysis for Discrete Data.” American Journal of Sociology89:54–110.
58.
WooldridgeJeff. 2002. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: MIT Press.
59.
YoungJacob. 2011. “How Do They ‘End Up Together’? A Social Network Analysis of Self-Control, Homophily, and Adolescent Relationships.” Journal of Quantitative Criminology27:251–73.
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
