Discretion Within Constraint: Homophily and Structure in a Formal Organization

Abstract

H omophily in social relations results from both individual preferences and selective opportunities for interaction, but how these two mechanisms interact in large, contemporary organizations is not well understood. We argue that organizational structures and geography delimit opportunities for interaction such that actors have a greater level of discretion to choose their interaction partners within business units, job functions, offices, and quasi-formal structures. This leads us to expect to find a higher proportion of homophilous interactions within these organizational structures than across their boundaries. We test our theory in an analysis of the rate of dyadic communication in an email data set comprising thousands of employees in a large information technology firm. These findings have implications for research on homophily, gender relations in organizations, and formal and informal organizational structure.

Intraorganizational Network Dynamics in Times of Ambiguity

“…Between these two poles lies semiformal organization, shared affiliations 6 afforded (but not necessarily mandated) by participation in the myriad project teams, task forces, and employee groups that exist within organizations (Biancani, McFarland, and Dahlander 2014;Feld 1981;Kleinbaum, Stuart, and Tushman 2013;Yakubovich and Burg 2014).…”

“…First, interpersonal relationships and communications tend to hew to the contours of organizational structure-for example, the boundaries defined by departments and functions (Kleinbaum, Stuart, and Tushman 2013;Han 1996;Hinds and Kiesler 1995). Indeed, as Allen (1977: 211) noted, "The real goal of formal organization is the structuring of communication patterns."…”

“…Prior research attempting to identify semiformal affiliations among people using email data has assumed that, when two people receive the same mass email, they also have a semiformal affiliation (Kleinbaum, Stuart, and Tushman 2013). In effect, this approach infers the existence of an email distribution list based on messages exchanged.…”

Knowing Me, Knowing You: Inventor Mobility and the Formation of Technology-oriented Alliances

“…We address the non-independence problem by estimating robust standard errors that are clustered on three dimensions: on the level of each dyad i-j, the level of firm i, and the level of competitor j. Cameron, Gelbach, and Miller (2011) develop an approach that allows for simultaneous clustering on both firms in a dyad as well as the dyad itself that has been implemented in a Stata ado-file (Kleinbaum, Stuart & Tushman, 2013). 13 We compute marginal effects at the mean (MEM) for our key independent variables based on the results obtained from the three-way clustered estimations using the delta-method (Wooldridge, 2010).…”

Seeing Social Structure: Assessing the Accuracy of Interpersonal Judgments About Social Networks

“…First, we rely on self-reported network data, which are susceptible to various forms of reporting bias (Marsden 2011). It would be useful in future studies to include more objective measures of targets' networks such as those derived from email archives (Kleinbaum, Stuart, and Tushman 2013;Srivastava 2015;Goldberg et al 2016;Srivastava et al 2017). Second, we used laptop webcams to gather videos of targets.…”