Balázs Kovács – Yale School of Management
Amanda J Sharkey – Booth School of Business, The University of Chicago
Nick Mmbaga – University of Tennessee, Knoxville
Trey Lewis – University of Tennessee, Knoxville
Question 1. Your paper provides a novel context and challenges prior assumptions about the effects of possessing high status. In your context, having high status is beneficial but not optimal; we are curious to know, did you theorize these outcomes apriori or were you caught by surprise after you conducted your analysis? Assuming you were, how did the process of moving beyond the data to building a plausible story around these counterintuitive results unfold?
You are touching on an important point, and indeed, we did not expect this pattern to happen. Our very first reaction was a surprise and we thought that we might have done something wrong with the analysis. But the empirical pattern seemed very robust even after several additional analyses. That’s when we got excited because it’s always fun to figure out what cause unexpected results. Indeed, we were hoping to uncover a new theory. So, we sat down to work out alternative explanations that would explain the findings. These alternative explanations ended up as the hypotheses and the tests for alternative explanations. Of course, for such process not to be a post hoc rationalization, we needed to measure these mechanisms directly. Therefore, we created measures that would capture these mechanisms, and tested them, and demonstrated how they explain and mediate the main effect of status. Maybe one reason for the success of the paper is that we explain a counter-intuitive set of results with a set of plausible mechanisms.
It’s another story how to write up such a paper, and we think that the experience is worth sharing with the readers. We both teach research design courses and we know from our own experience, from talking to other researchers, and from the literature, that such an abductive research process is quite common in practice. Yet, almost all papers in management are written in the “hypothetical deductive” way, that is, by starting with hypotheses that are then tested. This is the norm in the field, especially with quantitative papers. We think that such a norm is often misleading because, as you see in our example, this is not how the research process has unfolded. We have been telling our PhD students the “behind the scenes” story about the process, but these are not reflected in the paper, and the process of discovery is not visible. Therefore, we decided to go against the tradition and write up the framing in the way we came up with the findings; we revealed the finding of a negative effect of status very early in the paper. We experienced some pushback from reviewers on this, and it was suggested that we write the framing in a more traditional way. We revised somewhat but ultimately stuck with our original approach. We hope the success of our paper might inspire others to write up their findings in a similar way.
Question 2. From our reading, there seems to be a perfect marriage between the context (book reviews) and your theory. Do you foresee similar result in other contexts? For instance, do you see potential for the IPO (Initial Public Offering) context to be used in future research based on the rationale that an IPO can be considered a positive “public status shock” that is ambiguous and subject to equivocal preferences? The crowdsourcing context seems like a natural fit as well. What are your thoughts? And are there other contexts that are ripe for future research?
We definitely think that our theory applies more widely. Take the audience-shift hypothesis: we believe that such a shift in audience is likely to happen in most cases when the person, organization, or product receives a public status shock. An audience shift is likely to lead to lower “fit.” We can envision our theory to apply in most cultural settings, such as movies and the Academy Awards. Another interesting case could be teaching ratings of elective classes: if the ratings of a professor go up (as compared to a similar quality professor), then next year maybe more student will take her class because they heard about the higher ratings and not necessarily because they were intrinsically interested in the topic. Therefore, we expect “fit” to go down, implying lower ratings next year. The IPO process could also be one example, although it is less clear to us how one would define “fit” between the stockholders and the company. If you think of fit as “how knowledgeable the shareholders are about the company,” then yes, the fit definitely decreases after an IPO.
Question 3. Your study raises an interesting question about how prize-granting institutions (e.g., Academy Awards, NSF, Nobel prize, Grammys, etc.) maintain their credibility in the presence of increasingly large and diverse audiences that challenge the judgements these institutions make. In your mind, how do institutions mitigate these potentially adverse effects and maintain their legitimacy? How do they ensure that there is a continuous fit with external evaluators’ tastes as audiences grow?
This is an interesting question. In some cases, it seems to us that the lack of fit between prize committee tastes and the tastes of average consumers works to the prize committees’ advantage. By this, we mean that people sometimes actually seem to enjoy debating whether or not a certain movie deserved an Oscar or not, and this kind of public debate sometimes drives individuals who have heard about the controversy to go see the movie and decide for themselves. Obviously, however, this has its limits. This process can work in cultural domains where there is quite a bit of ambiguity about what a “good” movie or book is, but we think such controversies would be more problematic for the legitimacy of an award in domains where quality is more unambiguous.
One other point to note is that we think this kind of legitimacy challenge is most likely to occur in settings where it is possible for a wide audience to gain experience with the prize-winning product. For example, in our case, people can easily buy a book, read it and form their own opinion. In other settings, like the Nobel prize in physics, a wide audience simply won’t have the knowledge to judge whether the prize committee “got it right.” Similarly, in settings where purchases prices are high, like with an award for an expensive car, people won’t be able to try out the prizewinning good and determine whether they think the award was deserved. So, in some sense, the legitimacy loss for an award is only a risk under certain conditions.
Question 4. Diff-Diff statistical models have been used extensively in economics and other quantitatively oriented disciplines. In light of the other methods you employ in your study, why did you select this particular statistical approach and how much emphasis did you put on using this well-established but less used technique in strategic management or organizational theory? Did reviewers openly receive this statistical approach? If not, how did you ease their apprehensions? If they did receive it well, do you recommend that doctoral students in OT, ENT, OB, SOC, or SM learn this technique?
We actually started the research with the differences-in-differences research design in mind; we had been inspired by the Azoulay, Stuart and Wang (2014) paper that used this set-up to look at prize-winning scientists. (Their research was a working paper at the time we were conducting our analyses). We were interested in testing the effects of status shocks, and we thought that with a diff-in-diff setup, we could get good empirical leverage on this issue. So, we looked for a setting in which we could reasonably match winners with non-winners. The book awards with the shortlists provided such a setting. We actually just published a follow-up paper (“The Many Gifts of Status: How Attending to Audience Reactions Drives the Use of Status”, Management Science 2018), in which we use the same setup.
We definitely recommend students to learn the differences-in-differences research design. It’s easy and intuitive and provides a reasonable strategy to get at causality. However, one thing to point out is that diff-in-diff requires that several specific assumptions hold in order for interpretation of the results as causal to be valid (e.g., similar pre-treatment trends in the treatment and control groups). As diff-in-diff has become more widely used, researchers haven’t always been explicit about whether and how they tested those assumptions. And we could have been more explicit in our study as well. We hope that future work will use the diff-in-diff method carefully.
Question 5. You demonstrate that it is not always beneficial to be at the highest strata of status. How can high status actors respond to negative evaluations that stem from positive “public status shocks”? From a practical perspective, what solution would you offer a high-status actor (Individual or organization), who’s stock (evaluation) has decreased due to their popularity?
Clearly, there are a lot of benefits to getting a status shock. As we show in the paper, many more people will read the award-winning books and sales go up. Yes, average ratings may go down, but ultimately it depends on what the actors care about: being known by many people or being liked by a few? This is a trade-off that presents itself in our findings. We are sure that both kinds of actors exist, and only the second set of people should be worried about the status shocks – future research could explore this question further!