1 / 10

Presentation of the Groningen group plan

Presentation of the Groningen group plan. Marijtje van Duijn ECRP Meeting Ljubljana, Feb. 2, 2012. Who is ( who ?) in Groningen. Marijtje van Duijn Christian Steglich Christophe Stadtfeld Tom Snijders Mark Huisman Filip Agneessens …. What do we do – for ECRP?.

jalene
Télécharger la présentation

Presentation of the Groningen group plan

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Presentation of the Groningen group plan Marijtje van Duijn ECRP Meeting Ljubljana, Feb. 2, 2012

  2. Who is (who?) in Groningen • Marijtje van Duijn • Christian Steglich • ChristopheStadtfeld • Tom Snijders • Mark Huisman • Filip Agneessens • …

  3. What do we do – for ECRP? • Statistics, models, methods, simulation/computing, applications … • ECRP project: Peer Influence in Social Networks: Comparing and Evaluating Methods across Domains. • Focused on influence – and disentangling it from selection (contagion vs. homophily)

  4. Influence vs. selection • Hot topic! • Requirements • Longitudinal data • Adequate statistical model/method • SAOM; networkautocorrelation; GEE; SEM; … • Whydifficult? (interesting!) • Resultsseem to dependon model used • Adequacy of model assumptions • Notstraightforward in longitudinalsocialnetwork data • No ‘independent’ observations… • Lots of debate

  5. The Groningen contribution Comparing and evaluatingmethod • Twodimensions • Models (usedfor the analysis) • Data (and theirproperties) • SAOM (RSiena) is the starting/reference point • Analysis • Data generation • Simulationstudies • Aimed at comparison of models/methods

  6. General purpose of a simulationstudy • Show that (estimation) methodworkswell • Generate data according to model • Estimate model parameters • Show that these are ‘good’ (unbiased, etc.) • Show that model estimation is sensitive to assumptionviolation and/ormisspecification • Generate‘wrong’ data • Use‘wrong’ model forestimation • Evaluateestimationresuls

  7. Rough design of ECRPsimulationstudy • Generate data and analyzeitwith • the ‘true’ model and • other‘wrong’ models • Compare the results of both analyses • Works for SAOM (can do both) • More difficultforother models/methods • E.g. GEE has implicitway of controllingfordependence in the network data • Howuseitfor data generation? What are trueparameter values? • What are realistic parameter choices, i.e. leading to plausiblenetwork data (structures)

  8. Input from/output to otherprojects • Otherprojectsmay provide interesting and realistic data configurations • For our data generating • Ourmethodsmaybeusefulforotherprojects • Adviceforchoosinganalysismethod • For analysis of your data • Greatforcollaboration!

  9. Whynow (and notearlier)? • Interestingquestion… • Somepossibleanswers • Focus on model development (as a start) • Variousapproaches to model development (different groups; workingonsimilarproblems; with different modelingsolutions) • Nowthatmodels have been developed, a wider view is possible • …???

  10. First start • Christoph’spresentationtomorrow • Questions and suggestions …???

More Related