Hierarchical Network Models


  About us

  Papers & Talks

  Data Sets & Examples




Intervention Effects on Social Networks in Education Research

March 2014, Spring Meeting of the Society for Research on Educational Effectiveness (SREE)

Tracy Sweet, University of Maryland
Andrew Thomas & Brian Junker, Carnegie Mellon University

Experimental and observational studies in education are sometimes focused not on the effects of changing curriculum, teaching and learning materials, or classroom technique, but rather on changes in the way students -- or teachers, teaching coaches and administrators -- interact with one another (e.g., relationship between transformational leadership and leaders' professional social connections, influence of peer group structure on student behavior and aggression, diffusion of innovation and reform initiatives in schools, advice giving/receiving and social capital among teachers, etc.). Many whole school initiatives encourage some type of social structural change, be it an increase in collaboration, distribution of leadership or a push toward small learning communities: in short, they encourage changes in the social networks of students and of professionals in school systems.

Hierarchical Network Models (HNMs) enable you to model and detect the effects of interventions and other covariates on the structure of social networks, by pooling information across ensembles of social networks (teachers’ professional networks across multiple school buildings, students’ peer networks across multiple geographic areas, etc.).   We will walk you through hands-on data analyses, using software we have developed, to model and understand influences on social network structure in education settings.  A brief summary of the technical aspects of HNMs will be included but most of the workshop will focus on substantive research questions and useful interpretations.

All data analyses will be conducted in the free statistical package R; participants should bring a laptop with the current version of R installed and functioning correctly.  Other data and software will be provided in the workshop.  There will be an optional "R bootcamp" from 8am to 9am, and the main workshop will run from 9am to noon. 

For more information about the workshop, or to find out what will be covered in the R bootcamp, please see 2014-03 SREE WORKSHOP.

Participants wishing to brush up on R in advance may consult the websites

and/or the books

  • Teetor, P. (2011). R cookbook. O'Reilly.
  • Chang, W. (2012). R graphics cookbook. O'Reilly.


  1. Ennett, S. and Bauman, K. (1993). Peer group structure and adolescent  cigarette smoking: a social network analysis.  Journal of Health and  Social Behavior, 226--236.
  2. Frank, K. A., Zhao, Y., & Borman, K. (2004). Social capital and the  diffusion of innovations within organizations: The case of computer  technology in schools. Sociology of Education, 77, 148--171.
  3. Low, S., Polanin, J.R., Espelage, D.L. (2013). The Role of Social  Networks in Physical and Relational Aggression Among Young  Adolescents.  Journal of Youth and Adolescence, 42(7), 1078--1089.
  4. Penuel, W. R., Frank, K. A., & Krause, A. (2006). The distribution of  resources and expertise and the implementation of schoolwide reform  initiatives. In Proceedings of the 7th international conference on  Learning sciences, ICLS '06, (pp. 522--528). International Society of  the Learning Sciences.
  5. Spillane, J., Kim, C., and Frank, K. (2011). Instructional Advice and  Information Providing and Receiving Behavior in Elementary Schools:  Exploring Tie Formation as a Building Block in Social Capital  Development.  Institute for Policy Research, Northwester University,  Working Paper Series.
  6. Sweet, T. M., Thomas, A. C., and Junker, B. W. (2013) Hierarchical  network models for education research: hierarchical latent space  models. Journal of Educational and Behavioral Research, 33, 295--318
  7. Thomas, A., Dabbs, B., Sadinle, M, Sweet, T., & Junker, B. (June  2013). Conditionally independent dyad network models; an integrative  framework for modeling and computing.  Working paper.
  8. Thomas, A., Dabbs, B., & Sweet, T. (August 2013).  CIDnet: An R  software package for inference with conditionally independent dyad  network models. In preparation.
  9. Weinbaum, E., Cole, R., Weiss, M., & Supovitz, J. (2008). Going with  the flow: Communication and reform in high schools. In J. Supovitz &  E.Weinbaum (Eds.), The implementation gap: understanding reform in  high schools (pp. 68--102). Teachers College Press.