Over the past 25 years, industrial and organizational (I-O) psychologists have made great strides forward in the area of teams research. They have developed and tested meso-level theories that explain and predict the behavior of individuals in teams and teams operating within and across organizations. The continued contributions of I-O psychologists to theory and research on teams require us to address the challenges—several of which were well described in the focal article (Guzzo, Fink, King, Tonidandel, & Landis, Reference Guzzo, Fink, King, Tonidandel and Landis2015)—and embrace the opportunities that are being ushered in by big and broad data streams (Hendler, Reference Hendler2013). We suggest that a principal unique value add of the I-O psychologist to the basic scientific endeavor of understanding small teams comes in the form of theory—theories that explain why, when, how, and to what end individuals form relationships needed for teams to function in unison toward the accomplishment of collective goals. Some have argued that the big data revolution means “the end of theory,” suggesting petabyte data render theoretical models obsolete (Anderson, Reference Anderson2008). On the contrary, we submit that big-data enabled social science holds the promise of rapid progress in social science theory, particularly in the area of teams.

As the focal article notes, big data about teams abounds, ranging from traces left as teams form and perform in online communities (Turek, Wierzbicki, Nielek, Hupa, & Datta, Reference Turek, Wierzbicki, Nielek, Hupa and Datta2010) to big data from computer simulations (Sullivan, Lungeanu, DeChurch, & Contractor, Reference Sullivan, Lungeanu, DeChurch and Contractor2015) to big data from wearable sensors (Kozlowski, Chao, Chang, & Fernandez, Reference Kozlowski, Chao, Chang, Fernandez, Tonidandel, King and Cortinain press; Pentland, Reference Pentland2000). Simultaneously, the “little team” is now capturing the attention of scholars in a wide range of fields and disciplines who are bringing big data to bear on the very same phenomena that I-O psychologists have been theorizing about and studying for decades. Climate scientists are engaging teams research as they build cyberinfrastructure tools to integrate data, tools, and methods from the disparate fields that study the land, sea, and air (Jacobs, Reference Jacobs2012). Biomedical scientists are engaging teams research as part of a new field dubbed “The Science of Team Science” (Cooke & Hilton, Reference Cooke and Hilton2015) in order to design organizational work systems such as those that led to the discovery of the Higgs boson (Aad et al., Reference Aad, Abajyan, Abbott, Abdallah, Abdel Khalek, Abdelalim and Zwalinski2012) or the sequencing of the human genome (Venter et al., Reference Venter, Adams, Myers, Li, Mural, Sutton and Zhu2001). Computer scientists are engaging teams research to understand forms of human collaboration enabled by the digital revolution such as peer production (Kittur et al., Reference Kittur, Nickerson, Bernstein, Gerber, Shaw, Zimmerman and Horton2013; Wilkinson, Reference Wilkinson2008) and collective intelligence (Smith, Reference Smith1994). Like the big data revolution in general, the engagement of these communities in the science of teams presents I-O psychologists studying teams with both challenges and opportunities.

The reality is that big data has brought the study of “little teams” to a crossroads. We wish to extend the discussion of big data to consider the opportunities to better accomplish the thing that we, as I-O psychologists, perhaps are the best positioned to do in the area of teams research: build and refine theory. We explore four key opportunities that arise when I-O psychologists pair their ability to develop, operationalize, and investigate social science theory with big data.

First, data-intensive research will enable social scientists to test predictions that derive from major theories “at scale.” For example, the team input–process–outcome (I–P–O) model was introduced to explain how certain inputs (e.g., team composition) could lead to team outcomes (Hackman, Reference Hackman and Lorsch1987; McGrath, Reference McGrath1984; Steiner, Reference Steiner1972). This model holds that team processes are the mediating link between inputs and subsequent outcomes. To date, hundreds of studies have tested team I–P–O relationships using samples ranging from 1 to 1,000 teams. Virtual organizational science has relied on this foundational thinking about teams (cf. Jarvenpaa, Shaw, & Staples, Reference Jarvenpaa, Shaw and Staples2004; Lurey & Raisinghani, Reference Lurey and Raisinghani2001; Martins, Gilson, & Maynard, Reference Martins, Gilson and Maynard2004). Newer models of team performance such as the input–mediator–output–input model (IMOI; Ilgen, Hollenbeck, Johnson, & Jundt, Reference Ilgen, Hollenbeck, Johnson and Jundt2005) suggest that the factors linking team inputs and outputs are often constructs other than behavioral processes, such as trust, cohesion, or shared cognition, which emerge over time and manifest at the team level. The way these constructs develop and shift over time is nonlinear and involves feedback loops whereby mediators and outputs at one moment become inputs at another (Ilgen et al., Reference Ilgen, Hollenbeck, Johnson and Jundt2005).

However, team I–P–O and IMOI research has barely scratched the surface of testing propositions about the influence of team context and environmental pacers on team internal functioning. Scant attention has been paid to exposing how processes unfold over time. To some degree, prior team theory and empirical research was limited by atomistic data—data centered on individuals or teams as entities. Measures of so-called “social behavior” and “team process” have more closely resembled “individuals’ perceptions” than coevolving behavioral repertoires. In contrast, big data is particularly useful for studying team dynamics given that these data are inherently relational and afford the investigation of more precisely specified relational theories (Macy, DellaPosta, & Shi, Reference Macy, DellaPosta and Shi2015). Mapping the emergence and impact of patterned team-level constructs over time (e.g., using implicit measures of relational constructs captured in real-time) is one way for research to delineate the IMOI cycles through which teams function. We offer this as an illustration of a key advancement that can be made possible by applying large-scale, data-intensive, computational social science (CSS; Lazer et al., Reference Lazer, Pentland, Adamic, Aral, Barabasi, Brewer and Van Alstyne2009) approaches to teams theory.

Second, data-intensive research may well expose new boundary conditions appropriate for incorporation into existing theories of teams. Over the past 2 decades, the social sciences have increasingly utilized meta-analysis as a technique for theory testing. Meta-analysis combines all available estimates of an effect size from published and unpublished studies in order to identify the boundary conditions, or moderators, of a given relationship (Hunter & Schmidt, Reference Hunter and Schmidt2004). In the past, this approach has proven particularly useful in the area of teams (e.g., Bell, Reference Bell2007; Mesmer-Magnus & DeChurch, Reference Mesmer-Magnus and DeChurch2009), given that access to large samples of teams data is typically untenable. A major limitation of relying on meta-analysis to understand teams, however, is that we are stitching together small data to construct big data, often with little knowledge of sources of variation between samples. CSS approaches, unfettered by small sample constraints, afford teams researchers a new way to chart the boundaries of their theories.

Third, data-intensive research invites new combinations of constructs, and the entry of altogether new constructs, into our theory space. For example, advancements in neuroimaging and sensor data are opening up the possibility that we can incorporate new constructs such as the “fluidity of team membership” (i.e., how and with whom collaboration occurs), “virtual proximity” (i.e., how reachable individuals are throughout the day), and “dominance” (i.e., a “protagonistic characteristic” influence over others in order to balance participation and derive consensus; Kim, McFee, Olguin, Waber, & Pentland, Reference Kim, McFee, Olguin, Waber and Pentland2012).

Fourth, there are bound to be entirely new phenomena identified through the use of big data. Thus, new theories about collaboration are needed to focus our attention on new or at least increasingly prevalent phenomena. An example of such a phenomenon is a virtual community (i.e., MediaMOO, Wikipedia) in which users come together voluntarily to connect and collaborate (Keegan, Gergle, & Contractor, Reference Keegan, Gergle and Contractor2012). With nothing more than a basic infrastructure in place, it is the users who develop the environment (Bruckman, & Resnick, Reference Bruckman and Resnick1995; Bryant, Forte, & Bruckman, Reference Bryant, Forte and Bruckman2005).

To recap, big data combined with computational social science methods (Lazer et al., Reference Lazer, Pentland, Adamic, Aral, Barabasi, Brewer and Van Alstyne2009) will define the future of teams research. The increasingly available digital streams afford an unprecedented and unparalleled opportunity for teams research. However, continuing to contribute to research on teams will require I-O psychologists to address the inherent challenges of big data and embrace the opportunities ushered in by big and broad data streams. Furthermore, the impact of data-intensive research in teams research can, perhaps, be seen most clearly when we consider the implications for teams theory.