Publications
Publications
- 2007
Mind Over Matter? Similarities and Differences Between Perceived and Observed Networks
Abstract
In spite of the rapid development of new methods for network analysis—relying on electronic data sources and sophisticated computational analysis—organizational scholars continue to rely largely on more traditional survey-based methods. We believe that the organizations field has been slow to adopt these new methods in part because our field doesn't know how to interpret the "ties" that exist in electronic data. In this paper, we employ the quadratic assignment procedure (QAP) and exponential random graph models (ERG or p* models) to analyze both a survey-based and an e-mail-based network describing relations among the same actors in the same organization. We demonstrate that e-mail networks are substantially similar to traditional survey networks (with inter-method correlations as high as .43), and so should be trusted as informative; but are also systematically different in particular, predictable ways that enable different sorts of research questions to be answered. Surveys measure respondents' perceptions of their social networks, which are biased from their observed networks in substantively important ways. As such, we suggest that e-mail methods are a useful complement to traditional survey methods of social network analysis.
Keywords
Citation
Quintane, Eric, and Adam M. Kleinbaum. "Mind Over Matter? Similarities and Differences Between Perceived and Observed Networks." 2007. (Under Review.)