On 1 October at 10:00 via Zoom, Marco S. Minervini (with Tianyu He and Phanish Puranam) from INSEAD will give a seminar entitled

Does social interaction hurt search by groups? Evidence from Kaggle

Join Zoom Meeting:
https://aarhusuniversity.zoom.us/j/61171762570?pwd=RGVXbGx4WE9md0pvTExXWEpMRjc0Zz09

Meeting ID: 611 7176 2570
Passcode: ICOA

Abstract
The COVID-19 pandemic has created significant disruptions to existing patterns of social interaction. Are these restrictions likely to impede or enhance group creativity? While the modeling and experimental literature have reached a generally negative view on the effect of social interaction on search outcomes, the applicability of these results may be limited because of the specific nature of social interactions studied. Using data from an online platform hosting multiple highly incentivized competitions based on search on complex and noisy landscapes we show that, as compared to their “zero-interaction” counterfactuals, socially interacting groups (which are unrestricted in how they interact) do indeed exhibit less exploration (variation) in their search outcomes as noted in prior studies. However, they are also able to stimulate greater search efforts from their members, an effect that has so far remained unexamined. Further, both search efforts and exploration contribute positively to search performance. This implies that whether social interaction is beneficial for group search depends on problem attributes that make either exploration or effort relatively more important. For problems that require variation in search, managers can take advantage of the reduced social interaction, and, even engineer it, to ensure superior performance through broader search in the alternative generation phase. For problems that require more search effort (for instance in refining the quality of solutions) rather than covering a broad space of solutions, managers might adopt team designs that enhance social interactions.