James R. Detert – Johnson Graduate School of Management, Cornell University
Ethan R. Burris – McCombs School of Business, University of Texas at Austin
David A. Harrison – McCombs School of Business, University of Texas at Austin
Sean R. Martin – Johnson Graduate School of Management, Cornell University
Jonathan Keeney – Kenan-Flagler Business School, University of North Carolina
Elijah Wee – Robert H. Smith School of Business, University of Maryland
Article link: http://asq.sagepub.com/content/58/4/624
Question 1. Your work highlights the negative effects of lateral voice flows in terms of unit-level performance, but might such voice flows also have positive effects at the individual and group levels? For example, is it possible for employees (either those who speak up, or those who are spoken up to by others) to view voice as a mechanism to gain status within the group? Or, might lateral voice flows contribute to group psychological safety, or affect the quality of TMX and LMX?
The negative effects on performance of lateral voice flows were indeed, from our perspective, among the more interesting aspects of this work. We definitely think further study of the outcomes of lateral voice flows is necessary, and are currently doing some of this work. One possibility (which we’re currently exploring) is that there are factors that will mitigate the negative performance-type effects of sideways voice. For example, lateral voice may be less harmful to performance depending on who the sideways voice targets are. We can imagine, for instance, that if the majority of lateral voice is targeted toward those with the most competence, expertise, best relationship with those with power as a “second lieutenant,” etc., then the negative effects of talking to peers might be attenuated or eliminated altogether. However, another interesting possibility tilts in the other direction. If the internal target of lateral voice has an animosity tie with unit leader, then the negative impact could be accentuated.
As you note, it’s also important to consider the effects of sideways voice at different levels. How do individuals, for example, who speak up a lot sideways, perform? Does it help or hurt them in their own work outcomes? And, what are the outcomes for peers who are commonly chosen as targets for sideways voice? Do they perform better because they are a kind of information hub and hence either have more routes to direct performance improvement in their own work and/or more opportunities to impress their boss by selectively speaking up (on behalf of others) with the comments most likely to be favorably received? Social network theory strongly predicts this impact because of both the control and information benefits of being central, and it’s something we’re currently exploring.
Question 2. We especially appreciated your application of a structural perspective to the study of voice flows. Do you think the effects you observe will be stable over time or, taking the structure of the organization into account, might some voice flows have other downstream consequences? For example, some lateral voice flows might serve as “rehearsals” in which employees prepare and refine their message before engaging in upward voice, which could have positive long-term effects at the unit level. Alternatively, upward voice flow outside of an employee’s typical chain of command might have negative long-term effects by challenging the legitimacy of the employee’s direct supervisor and undermining group cohesion.
You point to several possibilities related to longer-term dynamics that we think are important and definitely worthy of study. Pursuing these questions would probably require not just an extended longitudinal design, but perhaps also a focus not just on overall voice flows but specific episodes, too. For instance, it seems that studying how voice is first vetted, then rehearsed, and then delivered to the powerholder who can act would probably be best studied by focusing on specific issues (i.e., by using issues, rather than the aggregation of issues at the flow level, as the level of analysis). That would require careful, even participant observation of smaller subsets of employees. One possibility is that, if the voice were delivered electronically (e.g., via SMS or e-mail), then the idea being transmitted could be tagged and followed.
Similarly, it would likely take different design and methodological choices to really understand the conditions under which speaking up outside the chain of command undermines employees’ immediate supervisor. For this issue, we would have to consider causality issues – e.g., was lack of trust or respect for the supervisor the cause of going elsewhere to speak up, and/or the result of doing so? And how does the overall culture of the organization determine whether such behavior is supported and normative, such that it isn’t seen as undermining or signaling incompetence but rather as a functional learning strategy that secure people welcome?
Question 3. It is fairly rare to see publications that draw on both qualitative and quantitative methods. According to Edmondson and McManus (2007), this hybrid approach is best suited when the state of prior theory development is intermediate, or positioned somewhere between nascent and mature. Could you share with us your decision process regarding the appropriate methodology to employ in this project? What led you to conclude that both induction and deduction were needed?
When we first designed the project, we felt that an important advance could be made in our understanding of voice by taking a structural approach that tapped voice flows to different types of targets. There is a fairly mature understanding of the antecedents of voice. And we knew that despite many anecdotes and logically plausible claims about the potential positive consequences of more voice, that was a much more nascent area, and it was important to empirically establish when and why this might be true. However, as our interactions with our company partners and our planning continued, we also realized that we didn’t know enough about when and why and about what issues people would choose different targets. This made it important to collect other kinds of data, too. In short, we knew we didn’t know enough at the very beginning to lay out deductively-derived hypotheses and never iterate on that starting point. But we did want to systematically develop and test ideas. Hence, a hybrid strategy made sense.
Question 4. What is one idea or concept in your original submission that you wish could have made it through the review process? Why is this idea important to you?
Rather than being very specific, which probably isn’t very helpful at this stage, let us answer that more generally. In early versions of our thinking and writing, we envisioned more of a comparative, “horse race” kind of paper wherein we’d lay out what might be considered a “prototypical network/structural perspective” and the “organizational behavior perspective” on communication/voice. We thought this might be a useful way to use our data to contribute to both network research and voice research. As we moved forward, it became clear that this only worked by creating two “strawman” views, which in the end means it didn’t really work. So, it’s probably fair to say that in the end we didn’t “adjudicate” anything, and we perhaps didn’t even really bring the two approaches or groups of scholars any closer together. But, hopefully we did show (as a few others have) the power and value of using a structural perspective and network methods to extend understanding of the kinds of constructs and phenomena of central interest to OB researchers.
Question 5. What are your hopes or expectations in terms of how voice scholars should apply and extend this research going forward? How should this change our approach to studying voice? What’s next? Can you share with us anything about your current work that extends research from this publication?
As noted above, we hope the approach we took encourages others to build on our work by continuing to explore voice networks – both their antecedents and their outcomes, and at both the unit level and individual node/employee level. We have many papers at the individual level, and some now that consider team- or unit-level climates for voice and silence. Not to say that we accomplished this, but it seems that some of the best understanding will come from research that simultaneously considers multiple kinds of influences at multiple levels. Voice obviously emanates from someone and is targeted to someone (making the study of both speakers and recipients at the individual/node level important), and it’s also obviously embedded in a broader environment (making understanding of the formal and informal structure of an organization so important). Thus, the more we can design studies that take both of these realities into account simultaneously, the more our findings are likely to really explain the phenomenon as it occurs in organizations.