Our team was part of a broader effort by researchers at UC-Irvine (Gloria Mark, Stephan Mandt) and UC-Boulder (Sidney D’Mello). The focus of our group was with respect to the back-end software that handled the respective beacon, phone agent, and wearable data with much of our codebase driven by Dr. Gonzalo Martinez. Although the various security keying / licensing (esp. for the Garmin wearables) prevents us from sharing the code via a public repository, we are happy to privately share upon request all of the code, particularly the more difficult back-end code with respect to processing secure HTTPS POST operations from the Garmin Health API. Our code has now been used for roughly 8+ years starting with our Tessarae project and culminating in this particular project.
For code sharing, please reach out to Prof. Aaron Striegel and we would be happy to share code as appropriate or determine if you can collaborate for testing purposes (e.g. try it out to see if it works for what you had in mind). We also have a wide variety of past data from Tesserae that might be worth looking into for various experiments.
Study Overview
We believe that the future of information work will be characterized by new organizational structures, the integration of new technologies such as AI into the workplace, and new models of work. We expect that for teams, non-routine work, increased diversity, and the need to be adaptable to fluid team membership and changing work structures will require new skills in decision-making and in navigating such complex conditions. The challenge is enhanced especially as the amount of digital information to be managed increases. Our work is situated in the context of cognitively demanding information work in high tech domains where teams of the future will need to be effective and efficient problem solvers. Our main hypothesis is thus, that the future of teamwork will require integration of technological advances to facilitate team performance as work becomes increasingly nonroutine and complex, as teams become more diverse, and as the workplace becomes increasingly dynamic. The primary goals of this project are (1) to develop models of team states (e.g., team affect/mood; team cohesion) and team processes (e.g., team communication, coordination) that are associated with team performance in real-world contexts. Machine learning will be used to develop the models; and (2) to integrate the machine learning models into an intelligent team facilitator (AI-TF) that will dynamically intervene with targeted strategies based on sensed team states and processes. our approach for the facilitators has been (1) obtaining basic research insights; (2) conducting interviews with participants; and (3) machine learning of team states/processes.
In year 1, we developed our research procedures and made necessary adjustments to our protocols for the recruitment and enrollment of teams due to COVID-19. For example, the original study was set up to recruit teams from the UCI Beall Applied Innovation Center. However, when work moved to the home in the wake of the outbreak, we expanded the scope and created recruitment and enrollment materials (e.g. instructional videos, a website, ad campaigns on social media for recruitment). We also started collecting data in year 1. In year 2, we expanded our recruitment efforts, collected more data, and conducted analysis on team diversity (published), sleep alignment (published), and work location (original submission not accepted). IIn year 3, we conducted additional data on participant’s calendars and focused on publishing one manuscript on work location (published) and initiating research on emotional contagion. In what follows, we summarize our activities in our no cost-extension year 4 which mainly focused on an analysis of emotional contagion among teams (also see Supplement).