I am an Associate Professor of Organizational Behavior at the Graduate School of Business at Stanford University. My main areas of research include social networks (especially studying them experimentally in real world settings), entrepreneurship, careers and work, and information technology as well as their intersection.
At Stanford I have served as PhD program liaison (running the Macro-OB Program from the faculty side). I also serve as Associate Editor for Management Science, the leading quantitative journal in management. I have taught classes in Social Networks at the MBA and PhD level, Organization Theory, and Data and Decision Making.
In addition to my academic work, I also have collaborated with industry, government and non-profit organizations to apply cutting edge social science and data analytics to problems of relevance to these organizations.
My doctoral dissertation Social networks, stratification and careers in organizations examined two central issues in network analysis: (1) how networks form and evolve among previously disconnected individuals and (2) how individuals’ social networks affect their performance and career outcomes.
Because the data I used in my dissertation (observational and mainly cross-sectional), I found it difficult to make strong causal claims about the mechanisms that drive network formation, network effects, and career progression. This got me thinking.
Recognizing this limitation in my dissertation work, I began collecting data on networks and academic performance (with Surendrakumar Bagde) on a cohort of students at a college in India. I knew that I could use one feature of the university–the fact that students were randomly assigned to dormitory rooms–to better develop and test theories of network formation and network effects.
In The mechanics of social capital and academic performance in an Indian college (Am Soc Rev 2013) we examine how network processes affect academic achievement. In Peers and network growth: Evidence from a natural experiment (Man Sci.) we study how networks form and change in response to interacting with well-connected others. We are also examining how inter-caste contact affects the formation and dynamics of inter-caste friendships.
Network Field Experiments
In recent research, in collaboration with Rembrand Koning my former doctoral student now Assistant Professor at Harvard Business School, I have been working on developing new approaches for conducting field experiments to study network processes. We recently conducted a field experiment embedded in an organization we founded called Innovate Delhi Entrepreneurship Academy. The study examines the causal role that networks play in the innovation and entrepreneurship process. A recent working paper from this study called Randomizing Networks in the Field examines the extent to which networks in off-line field settings can be experimentally treated, thus carving the way for rigorously studying network effects in a variety of field settings.
I have also done more conceptual research on network formation and evolution. Both the ties that make up networks and the network “positions” people occupy are dynamic. In Group based trajectories of network formation and dynamics ( Soc Net 2012 ), I propose the use of an existing mixture-model (called DTA) to describing network trajectories–at both the dyadic and node levels.
Careers and Work
The second set of questions I am interested in concern careers and labor markets. One stream of this research with John-Paul Ferguson focuses on the role of skill specialization and diversification over a worker’s career. We find in Specialization and career dynamics: Evidence from the Indian Administrative Service (Admin Sci Q 2013) that contrary to prior research, specialization benefits workers throughout their career, but the reasons why specialization matters differs between early and late career stages.
While studying specialization, we realized that we were making a strong assumption about the opportunity structure of jobs through which individuals move. In particular, studies of work often assume stable job structures where experience is accumulated. However, jobs in organizations are created and destroyed, and this this instability may impact workers. Before studying the impact of a job structure ’s dynamics, we realized that it would be fruitful to first understand the sources of change (and stability) in organizational jobs. In “The lives and deaths of jobs: Technical interdpendence and survival in a job structure” (Org. Sci 2015) John-Paul Ferguson, Rembrand Koning and I develop a theory of job stability that depends on how the job is linked to other jobs in the organization through ties of task overlap and coordination. We theorized that jobs that are linked together with other jobs through these dependences are less likely to die. Testing such a theory requires detailed data on personnel and distinct jobs over decades. Using data collected in 2008 through a series of FOIA applications to the University of Michigan, we amassed information on the entire UM workforce between 1979 and 2009 and all job descriptions from the 1950’s to 2004. We tested our theory using this data and new text mining approaches and found that jobs linked together through ties of technical interdependings—task overlap and coordination—survive longer, and that such protection is contingent on the otherpresence of types of restructuring within the organization.
Before switching my field to organizational behavior in my third year of graduate school, I studied information systems and health informatics. My main interest at that time was on how we could use information technology to improve health care outcomes and processes. In my second-year paper at Carnegie Mellon, I developed a machine learning approach to address the “Medication Reconciliation” problem. I published a version of this research as Towards a collaborative filtering approach to medication reconciliation (Proc Am Med Info Assoc 2008) and later as a full length journal article with a more developed methodology and framework in Automatic detection of omission in medication lists (J. Amer Med Info Assoc 2011).