Skip to contentScholarGate
LibraryBookshelfDeskReview StudioAssistant
Sign in
On this page
IntuitionHow it worksWhen to use itStrengths & limitationsCommon pitfallsApplicationsFrequently asked🔒 Read the full methodSourcesRelated methods
Cite this pageSpotted an issue on this page? Report or suggest a fix →
Home›Network analysis›Dynamic Exponential Random Graph Model
Machine learningNetwork science

Dynamic Exponential Random Graph Model

Dynamic Exponential Random Graph Model (Temporal ERGM) · Also known as: TERGM, Temporal ERGM, Dynamic ERGM, STERGM

The Dynamic Exponential Random Graph Model (TERGM / STERGM) extends the classic ERGM framework to panel network data, modeling how a network's ties form and dissolve over time as a function of structural tendencies, nodal attributes, and the network's own past state. It provides statistically principled inference about longitudinal network change.

ScholarGate
  1. Machine learning
  2. v1
  3. 2 Sources
  4. PUBLISHED
Cite this page →
Tools & resources
Download slides
Learn & explore

Read the full method

Members only

Sign in with a free account to read this section.

Sign in

Method map

The neighbourhood of related methods — select a node to explore.

Dynamic Exponential Random Graph Model
Dynamic Stochastic Block…Network Diffusion Analys…Stochastic Block ModelTemporal Network Analysis

When to use it

Use a dynamic ERGM when you have two or more observed snapshots of the same network over time and want to model the endogenous and exogenous drivers of tie formation and dissolution simultaneously. It is appropriate for relatively small to medium networks (up to a few thousand nodes) with binary ties and panel structure. Do not use it when the network has continuous edge weights (weighted networks require different frameworks), when only a single cross-sectional snapshot is available (use static ERGM instead), when the network has tens of thousands of nodes (computational cost becomes prohibitive), or when the temporal lag structure is unknown and theoretically undefined.

Strengths & limitations

Strengths
  • Jointly models tie formation and dissolution, separating the forces that create new ties from those that maintain existing ones.
  • Conditions on the observed past network, enabling causal-style inference about longitudinal network change.
  • Handles endogenous network statistics (reciprocity, triangles) that violate independence assumptions of standard regression.
  • Goodness-of-fit diagnostics via simulation reveal whether the model reproduces the observed network distribution.
  • Interpretable coefficients on structural and covariate terms support theory-driven network analysis.
Limitations
  • Computationally intensive: MCMLE can be slow or fail to converge on networks larger than a few thousand nodes.
  • Model degeneracy is a persistent risk — certain combinations of triangle and two-star terms can produce near-complete or near-empty networks.
  • Requires at least two network observations; more time points improve estimation but increase cost.
  • The choice of temporal lag (t to t-1) is theoretically driven; misspecifying the lag can bias all estimates.

Frequently asked

How is TERGM different from a static ERGM?

A static ERGM models one network snapshot as a function of structural statistics. A TERGM additionally conditions on the prior time point's network, allowing the model to capture how past ties influence future ones and to separately estimate formation versus dissolution processes.

What is STERGM and how does it relate to TERGM?

STERGM (Separable Temporal ERGM) is a variant developed by Krivitsky and Handcock (2014) that models tie formation and tie dissolution with separate parameter vectors, recognizing that the same structural forces need not drive both processes equally. It is the most widely used dynamic ERGM implementation, available in the R tergm package.

How many time points do I need?

A minimum of two observations is required. In practice, three or more time points improve parameter stability and allow cross-validation. Very short panels (two points) should be interpreted cautiously, especially for dissolution parameters.

What should I do if my model shows signs of degeneracy?

Degeneracy appears when simulated networks from the fitted model are mostly empty or mostly complete. Reduce the number of triangle or two-star terms, consider geometrically-weighted alternating statistics (GWESP, GWDEGREE) which are less prone to degeneracy, and re-run goodness-of-fit diagnostics after each revision.

Can I use TERGM on weighted or valued networks?

Standard TERGM implementations assume binary ties. Valued (weighted) dynamic networks require separate frameworks such as valued ERGMs or latent space models. Dichotomizing edge weights before fitting TERGM is an option but loses information and should be theoretically justified.

Sources

  1. Hanneke, S., Fu, W., & Xing, E. P. (2010). Discrete temporal models of social networks. Electronic Journal of Statistics, 4, 585–605. DOI: 10.1214/09-EJS548 ↗
  2. Krivitsky, P. N., & Handcock, M. S. (2014). A separable model for dynamic networks. Journal of the Royal Statistical Society: Series B, 76(1), 29–46. DOI: 10.1111/rssb.12014 ↗

How to cite this page

ScholarGate. (2026, June 3). Dynamic Exponential Random Graph Model (Temporal ERGM). ScholarGate. https://scholargate.app/en/network-analysis/dynamic-exponential-random-graph-model

Related methods

Dynamic Stochastic Block ModelNetwork Diffusion AnalysisStochastic Block ModelTemporal Network Analysis

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

  • Dynamic Stochastic Block ModelNetwork analysis↔ compare
  • Network Diffusion AnalysisNetwork analysis↔ compare
  • Stochastic Block ModelNetwork analysis↔ compare
  • Temporal Network AnalysisNetwork analysis↔ compare
Compare side by side →

Similar methods

Exponential Random Graph ModelDirected Exponential Random Graph ModelWeighted Exponential Random Graph ModelTemporal Social Network AnalysisBayesian Exponential Random Graph ModelBayesian Temporal Network AnalysisTemporal Stochastic Block ModelDynamic Stochastic Block Model

Related reference concepts

Time-Series Models • Dynamic Quantile Regressions • Dynamic Treatment Effect Models • Diffusion Processes • State Space ModelsTime-Series Models • Dynamic Quantile Regressions • Dynamic Treatment Effect Models • Diffusion ProcessesPanel Data Models • Spatio-temporal ModelsPanel Data Models • Spatio-temporal ModelsStructural Equation ModelingModel-Based Clustering

Spotted an issue on this page? Report or suggest a fix →

ScholarGate — Dynamic Exponential Random Graph Model (Dynamic Exponential Random Graph Model (Temporal ERGM)). Retrieved 2026-07-21 from https://scholargate.app/en/network-analysis/dynamic-exponential-random-graph-model · Dataset: https://doi.org/10.5281/zenodo.20539026
Quick facts
Originator
Hanneke, Fu & Xing; Krivitsky & Handcock
Year
2010–2014
Type
Probabilistic graphical model (temporal)
DataType
Longitudinal binary adjacency matrices (network panel data)
Subfamily
Network science
Related methods
Dynamic Stochastic Block ModelNetwork Diffusion AnalysisStochastic Block ModelTemporal Network Analysis
ScholarGate

A content-first reference library for research methods — what each one is, how it works, and where it comes from.

Open data (CC-BY)

Explore

  • Library
  • Search the library…
  • Browse by field
  • Fields
  • Journey
  • Compare
  • Which method?

Reference

  • Subjects
  • Atlas
  • Glossary
  • Methodology
  • Philosophy

Your tools

  • Bookshelf
  • Desk
  • Chat

Company

  • About
  • Pricing
  • Contact
  • Suggest a method

Entries are compiled from published sources for reference. Verifying the accuracy and suitability of any information for your own use remains your responsibility.

© 2026 ScholarGate · A research-method reference library
  • Privacy
  • Cookies
  • Terms
  • Delete account