Collective Problem-Solving of Groups Across Tasks of Varying Complexity

Abstract

As organizations gravitate to group-based structures, the problem of improving performance through judicious selection of group members has preoccupied scientists and managers alike. However, it remains poorly understood under what conditions groups outperform comparable individuals, which individual attributes best predict group performance, or how task complexity mediates these relationships. Here we describe a novel two-phase experiment in which individuals were evaluated on a series of tasks of varying complexity; then randomly assigned to solve similar tasks either in groups of different compositions or as individuals. We describe two main sets of findings. First, while groups are more efficient than individuals and comparable “nominal group” when the task is complex, this relationship is reversed when the task is simple. Second, we find that average skill level dominates all other factors combined, including social perceptiveness, skill diversity, and diversity of cognitive style. Our findings illustrate the utility of a “solution-oriented” approach to identifying principles of collective performance.

Empirica: a virtual lab for high-throughput macro-level experiments

“…It has long been recognized that the internet presents researchers with new opportunities to recruit remote participants for behavioral, social, and economic experiments (Grootswagers, 2020). For instance, remote participation allows researchers to solve some of the issues that limit laboratory research, such as (1) recruiting more diverse samples of participants than are available on college campuses or in local communities (Reips, 2000;Berinsky, Huber, & Lenz, 2012); (2) increasing statistical power by enabling access to larger samples (Awad et al 2018;Reips, 2000); and (3) facilitating longitudinal and other multiphase studies by eliminating the need for participants to repeatedly travel to the laboratory (Almaatouq, Yin, & Watts, 2020;Reips, 2000). The flexibility around time and space that is afforded by remote participation has enabled researchers to design experiments that would be difficult or even impossible to run in a physical lab.…”

“…While early online experiments often required extensive up-front customized software development, a number of virtual lab software packages and frameworks have now been developed that reduce the overhead associated with building and running experiments. As a result, it is now easier to implement designs in which dozens of individuals interact synchronously in groups (Arechar, Gächter, & Molleman, 2018;Almaatouq, Yin, & Watts, 2020;Whiting, Blaising, et al, 2019) or via networks (Becker et al, 2017), potentially comprising a mixture of human and algorithmic agents (Ishowo-Oloko et al 2019;Traeger, Sebo, Jung, Scassellati, & Christakis, 2020;Shirado & Christakis, 2017).…”

“…Empirica has an active and growing community of contributors, including professional developers, methodfocused researchers, question-driven social scientists, and outcome-oriented professionals. Although Empirica is under steady development, it has already been used to build (at least) 31 experiments by more than 18 different research teams across 12 different institutions, generating at least 12 manuscripts between 2019 and 2020 (Feng, Carstensdottir, El-Nasr, & Marsella, 2019;Pescetelli, Rutherford, Kao, & Rahwan, 2019;Becker, Porter, & Centola, 2019;Becker, Guilbeault, & Smith, 2019;Almaatouq, Noriega-Campero, et al, 2020;Houhton 2020a, b;Becker, Almaatouq, & Horvat, 2020;Almaatouq, Yin, & Watts, 2020;Noriega et al 2020;Feng 2020;Guilbeault, Woolley, & Becker, 2020;Jahani et al 2020).…”

“…The "Room Assignment" game (Almaatouq, Yin, & Watts, 2020) 12 explored how factors such as task complexity Fig. 4 This screenshot of the "Detective Game" shows the view that participants use to categorize mystery clues as either Promising Leads (which are shared with their social network neighbors) or Dead Ends (which are not).…”

“…For example, the platform explicitly separates experiment design and administration from implementation, promoting the development of reliable, replicable, and extendable research by enabling "experimentation-as-code." This modular structure encourages strategies such as multifactor (Almaatouq, Noriega-Campero, et al, 2020), adaptive (Letham, Karrer, Ottoni, & Bakshy, 2019;Balietti et al, 2020b;Paolacci et al, 2010;Balandat et al, 2020), and multiphase experimentation designs (Mao, Dworkin, Suri, & Watts, 2017;Almaatouq, Yin, & Watts, 2020), which dramatically expand the range of experimental conditions that can be studied. Additionally, the platform provides built-in data synchronization, concurrency control, and reactivity to natively support multi-participant experiments and support the investigation of macro-scale research questions.…”