Modeling Relational Events
Autor: | Guido Conaldi, Alessandro Lomi, Marco Tonellato, Eric Quintane |
---|---|
Rok vydání: | 2014 |
Předmět: |
Class (computer programming)
050402 sociology Exploit Event (computing) Computer science Strategy and Management 05 social sciences Statistical relational learning General Decision Sciences Context (language use) computer.software_genre Data science Set (abstract data type) Empirical research 0504 sociology Management of Technology and Innovation 0502 economics and business Relational model Data mining computer 050203 business & management |
Zdroj: | Organizational Research Methods |
ISSN: | 1552-7425 1094-4281 |
DOI: | 10.1177/1094428113517007 |
Popis: | Sequences of relational events underlie much empirical research on organizational relations. Yet relational event data are typically aggregated and dichotomized to derive networks that can be analyzed with specialized statistical methods. Transforming sequences of relational events into binary network ties entails two main limitations: the loss of information about the order and number of events that compose each tie and the inability to account for compositional changes in the set of actors and/or recipients. In this article, we introduce a newly developed class of statistical models that enables researchers to exploit the full information contained in sequences of relational events. We propose an extension of the models to cater for sequences of relational events linking different sets of actors. We illustrate the empirical application of relational event models in the context of a free/open source software project with the aim to explain the level of effort produced by contributors to the project. We offer guidance in the interpretation of model parameters by characterizing the social processes underlying organizational problem solving. We discuss the applicability of relational events models in organizational research. |
Databáze: | OpenAIRE |
Externí odkaz: | |
Nepřihlášeným uživatelům se plný text nezobrazuje | K zobrazení výsledku je třeba se přihlásit. |