Adam M. Kleinbaum, Tuck School of Business, Dartmouth
Chen Zhang, Ross School of Business, University of Michigan
Karyn Dossinger, Carlson School of Management, University of Minnesota
Article link: http://asq.sagepub.com/content/57/3/407
Question 1. We think what is very interesting in your article is the examination of individuals’ dynamic career trajectories over time and in particular the notion that some people’s career trajectories could be “atypical” in their organizations. In this respect, we think your article can provide a good example both for idea generation and for method design. Idea-wise, what inspired you to look into career path typicality versus atypicality in relation to a social network outcome? Method-wise, besides using large archival data sets, what do you think could be some promising ways of inquiry (e.g., research methods, data opportunities) to gain insights on people’s career processes over time as well as on outcomes of such dynamic processes?
There were two factors that directed my attention to typical and atypical careers. The first came from the field research that underlies the quantitative data. In searching for a dissertation, I spent some time sitting in on corporate offsites, in various BigCo offices, and interviewing managers and executives there. One guy in particular really influenced my thinking. He held a PhD in the humanities (English literature, I think), but went on to become BigCo’s global head of financial services sales. I remember telling my wife about interviewing him and her very perceptive reaction was, “How did he end up in a job like that?” I hadn’t really thought about it, but I filed it away: “atypical career”.
At the same time, I was reading interesting research from the ecological tradition about categories, category-spanning, and the downsides of atypicality. Ezra Zuckerman’s work was particularly generative of my thinking. I began to wonder if there might be some upsides to atypicality and I started to consider whether atypical people might have networks that are valuable precisely because of the atypical way in which they were constructed.
There is a long and rich history of research on careers, but the emergence of large, longitudinal, electronic data sets is allowing us to study things that couldn’t readily be studied before. Until recently, the only way to collect longitudinal data on careers was to follow a sample of people for years and years. Archival electronic data makes it possible to study larger, more representative samples of careers. It’s an exciting time to be a social scientist.
Question 2. We are very intrigued by the potential paradox in the practical implications of your article. Your findings suggest that “organizationally misfit” career paths could help individuals accrue social network brokerage benefits. On the other hand, you also noted that such individuals might find others questioning their legitimacy when they make these unusual moves within an organization. In addition, when more individuals become interested in pursuing organizationally misfit career paths for social network benefits, these career trajectories might no longer be atypical or unique in the first place, and individuals’ brokerage advantages would be diminished. Therefore, it seems tricky to conclude that individuals should always actively pursue atypical career trajectories. So we are wondering: (a) At the individual level, what might you say to people who wish to draw some useful implications from your article for their own career designs? (b) At the organizational level, what advice would you give to organizations that wish to capitalize on the benefits of career misfits while minimizing their potential drawbacks?
In my experience talking about this work with students and executives, there is very little risk of anyone finding my results so persuasive that they are going to completely embrace a “misfit” identity! Most people – and certainly career-oriented MBA students – are pretty risk-averse about their careers and prefer to stick pretty close to the well-trodden paths. So the action, I think, is at the “typical” end of the continuum – people should be more willing to make an atypical move once in a while.
For organizations, I think the implication is a bit more profound. Lots of companies spend lots of time thinking systematically about job rotations. For example, management training programs often move people systematically from one part of the organization to another. But in virtually every company I’ve talked to, these programs are premised on a logic of human capital – giving people a broad base of experience for their future careers. My work suggests that the social capital benefits of mobility are very significant as well.
There are several implications of this. First, companies might significantly improve the returns on their investments in job rotation if they tweak the programs just a little to focus on building social capital as well as human capital. Second, when filling key positions, companies might benefit from considering not only the human capital of the candidate, but the network they bring to the table as well. If a job requires working closely with another part of the company, a candidate who has existing ties to people there from earlier in their careers might, all else being equal, be a better choice than a candidate whose network lacks such ties.
Question 3. In your article, you used a magnificent dataset of employees’ email correspondence in this organization BigCo. Gaining access to data of this scope and magnitude seemed critical for answering your research questions. Based on your experience with BigCo and on other projects, what might be some suggestions you would give, perhaps particularly to doctoral students, for how to build relationships with organizations that allow for accessing such intensive data? Once formed, what strategies have you found useful for managing ongoing relationships with organizations so as to make such collaborations mutually beneficial?
Unfortunately, it’s very hard to get a company to agree to provide data like this. In my limited experience, doing so successfully requires a strong personal relationship with a senior executive at the organization. So the first suggestion is to leverage your and your advisers’ personal networks to forge strong, trusting relationships with executives at possible research sites. Once you’re able to initiate a serious conversation about data collection, the next suggestion is to move as quickly as possible – the longer things drag on, the more likely it is that someone, somewhere in the organization will say no and kill the data collection.
Ironically, I’ve found that the more privacy protections that get put in place, the LESS likely it is that the data collection is ultimately going to succeed. The reason is, very simply, that if the company feels it needs lots of legalistic protections on the data – non-disclosure agreements, privacy policies, etc. – the sponsoring executive probably doesn’t trust you enough to really go to bat for you when the inevitable questions arise and make sure the data collection happens.
Having said all that, I think the world is changing very quickly. People and firms are getting used to the idea that there is no privacy with electronic data. I wouldn’t be surprised to see more data sets like this one becoming available to researchers.
Question 4. You made a very impressive effort in the paper to account for endogeneity in the relationship between career paths and social network brokerage. The issue of endogeneity seems especially relevant for social network studies. Do you have any thoughts or suggestions to share with other researchers who are also interested in investigating social network-related questions but might be concerned about endogeneity issues?
Yes, endogeneity is an important challenge in observational research and is especially difficult in network studies. In this case the problem was particularity acute because my theory concerns the effects of careers on networks – precisely the opposite direction of causality as that in so many earlier studies. In this paper, I used a version of propensity score estimators to try to account for the fact that mobility is an endogenous result of earlier networks in my estimates of the effect of mobility on subsequent network structure. In a current working paper, I am exploiting a rare natural experiment, in which formal structure is randomly and exogenously assigned, to examine the effects of organizational structure on network dynamics. These are just two of numerous possible identification strategies available in this work.
I think one of the issues confronting our field right now is a perception that there is a trade-off between interestingness and identification, between rigor and relevance. In the “straw man” characterization, there are people who ask interesting questions, but pay so little attention to the challenges of causal inference that their results are meaningless, and there are others who have such high standards for identification that the only questions they can ask are too narrow to be relevant. In my view, this represents a false dichotomy. The best papers I know find ways to ask interesting questions and to address them with analytical rigor. Doing this necessarily requires thoughtful research design and compromises on both sides, but ultimately leads to results that are both interesting and true.
Question 5. We are also interested in further research following your article. In your discussion, you noted that it could be an open question whether these results about brokerage benefits of atypical career paths would generalize to inter-organizational career mobility. For those individuals who make many moves across different organizations, do you think it would be more challenging for them to convert other people’s perceptions of them as uncommitted ‘job hoppers’ into someone who holds a valuable brokerage position? Are you conducting any further work, or aware of other people’s further work, that addresses this issue of atypical career trajectories across organizations?
Yes, Ming Leung has a terrific paper in ASR that looks at a very similar question in the external labor market for freelancing services. He finds that people with highly erratic job histories are penalized in the labor market. Together, I think these two studies can help to shed light on the boundary conditions surrounding the costs and benefits of atypical careers. In his empirical setting, firms and freelancers transact in a largely free, efficient market, where networks don’t seem to be all that important in determining the next transaction. In my setting within a large organization, networks matter more. So people with atypical or erratic career trajectories may be viewed as dilettantes and they may also assemble broad, far-reaching networks; how these two mediating mechanisms lead to performance outcomes depends on the setting.