Crowds are increasingly invited to tackle complex innovation problems. However, our understanding of how crowds solve these problems has limited our ability to utilize them effectively. In this article, we examine how a crowd tasked with finding ways to preserve New Zealand’s native flora and fauna achieves solutions to this complex innovation problem through discussion. Our analysis reveals that discussions leading to innovative solutions differ from those leading to non-innovative solutions: innovation-yielding discussions (1) mobilize more Resources than Constraints, (2) converge on a single Goal, and (3) specify few (often one) responsible Subject(s); vis-à-vis, discussions leading to non-innovative solutions often surfaced multiple goals—thereby sustaining multiple problem framings—while subjects remained vague, implicit, or multiplied, and constraints frequently outnumbered resources. Moreover, we show how the specific weavings of these discursive elements support the crowd’s navigation of the complex innovation problem. Three discursive maneuvers – subject-goal anchoring, repurposing distant knowledge, and layering resources – are fundamental navigational tools used by the crowd to form what we call ‘action-oriented narratives:’ simple, dynamic, and evolving linguistic structures that enable collective problem exploration while giving room for localized improvisation. We contribute to the literature on problem solving by showing how specific linguistic activities alter individuals’ capacity to contribute to and solve complex problems. We also contribute to the literature on linguistic structures supporting innovation by showing that the types of discussions that support innovation depend on the openness of the setting.
Technological change reshapes labor markets not only by changing labor demand but also by altering labor supply as workers reallocate effort across tasks and opportunities in response to technological advances. Yet research focuses more on demand-side effects than on how workers themselves adapt. In this research, we examine worker repositioning on a large online labor market following the launch of ChatGPT in November 2022. We argue that the launch of ChatGPT reduced expected returns to labor in exposed skill domains and develop and test a series of hypotheses regarding how this shock is associated with participation, horizontal positioning, and vertical positioning among workers on the platform. Using proprietary data on incumbent freelancers and a difference-in-differences design that leverages cross-sectional variation in preperiod AI exposure and skill level, we show that freelancers more exposed to AI shift away from prior domains and reorient bidding toward higher value contracts. In contrast, high skill-level freelancers, who face greater adjustment costs, are less likely to reduce participation on the platform, reposition across domains, and move upmarket. These findings highlight how technological shocks reshape labor markets through supply-side adaptation and how domain-specific human capital constrains worker responses. Funding: This work was supported by the Mack Institute for Innovation Management, Wharton School, University of Pennsylvania and the Smith-Richardson Foundation. Supplemental Material: The online appendix is available at https://doi.org/10.1287/orsc.2025.20566 .
Existing research on interorganizational network dynamics has largely treated network change as gradual, local, and primarily oriented toward private organizational advantage. Yet contemporary cooperative interorganizational systems are highly interconnected across industrial sectors and geographies and are routinely exposed to shocks that simultaneously disable multiple organizations or relationships among them. These shocks stem from pandemics, natural disasters, wars, sanctions, and cascading partner failures, and can rapidly reconfigure whole‑network architectures and undermine collective outcomes. In light of this new reality, we reconceptualize interorganizational networks as loci of frequent, punctuated change at the whole network level, with primary implications for collective outcomes. We develop this reconceptualization through a robustness–resilience lens, defining network robustness as an interorganizational network’s capacity to maintain its collective outcomes in the immediate aftermath of a shock, and network resilience as its capacity to recover and potentially surpass pre‑shock collective outcomes over the longer term. We then outline the theoretical and methodological opportunities that this robustness–resilience perspective opens for the study of network change.
This study examines how foreign firms strategically manage the salience of foreignness when facing unfavorable domestic evaluators. Foreign firms do so by altering the ethnic composition of the agents representing them before the evaluators. We examine this question in the context of Asian firms’ selection of legal counsel in U.S. federal courts. Prior research suggests that foreign firms may anticipate less favorable evaluations in courts that have historically favored domestic over foreign firms and in those with conservative judges. We argue that, in such evaluative settings, Asian firms adjust the ethnic composition of their external legal teams to manage the salience of foreignness, and that such adjustments are especially likely in courts that combine judge conservatism with pro-domestic ruling histories. We test these arguments using data on U.S. patent litigation from 2000 to 2016 and validated measures of courts’ political ideology and pro-domestic ruling histories. We find that Asian firms reduce the proportion of Asian-origin lawyers and increase the proportion of White lawyers when litigating in conservative courts, particularly when these courts also have pro-domestic ruling histories. By contrast, the ethnic composition of lawyers representing domestic firms remains stable across courts. Evidence of implicit bias against Asians among legal professionals provides additional support consistent with the theorized mechanism. Post-hoc analyses reveal that adjusting the ethnic composition of external legal teams attenuates the likelihood of unfavorable outcomes in conservative courts. These findings contribute to research on liability of foreignness by identifying agent selection as a specific mitigation mechanism in evaluative arenas.
As teams and team membership become more dynamic, colleagues from the same prior team increasingly transition together to a new team as a cohort: a phenomenon known as co-moving. Drawing on theories of shared team experience and team receptivity, we develop a multilevel, multi-theory, and multi-period model to examine how co-moving affects individual and team performance. We theorize that co-movers carry forward proximal shared team experience which they can leverage immediately to support their own performance relative to solo movers, who must build new working relationships. However, we also theorize that co-movers’ proximal shared team experience hinders team performance, though these detrimental effects attenuate over time. We test our theory using archival data from the National Basketball Association, supplemented by qualitative interviews with individuals involved in professional basketball. Our findings reveal that co-movers exhibit better individual performance than solo movers on their new teams and in our robustness checks we find that greater shared tenure on the prior team further enhances this advantage. At the team level, teams with co-movers perform worse than teams with solo movers, and we find partial support that this disadvantage attenuates over time. Additional analyses show that co-movers’ individual performance advantage persists over time in the new team. The findings from our interviews illuminate how these processes unfold both on and off the court. Our research contributes to the literature on shared team experience and newcomer integration by highlighting a multilevel tradeoff: while co-moving benefits individuals, it can initially undermine collective performance.
