ABSTRACT: We design a field experiment in an actual workplace to examine the effects of coworker gender, and work-team gender composition, on multiple dimensions of employee productivity and other behaviors (interaction style, production speed, accuracy, team strategies) and worker selection patterns. The experiment will have 400-500 workers in a real workplace and will be run in both in-office/in-person and remote/online settings. The job is a 3-week data-entry position, where worker subjects enter information about images – basic details as well as more open questions requiring judgment calls about the image. The work takes place through a sophisticated web-based work platform currently under construction – this platform is the sole interface through which workers perform their work in both the remote and in-person work settings. Randomization occurs at two stages – the work stage and the recruiting stage. Worker subjects in the experiment will engage in both solo-work and two-person team-work, and the experiment will include randomized gender composition of teams, randomized peer review and feedback, and flexible labor supply over the 3-week period. The platform will collect a wealth of data on worker productivity measures, as well as raw video-chat recordings which will be processed into a novel and uniquely rich data set for an intricate view of the inner workings of workplace team interactions. Outcome variables of interest include coworker interruptions, talking time, dominance, tone of voice, team strategy, production speed, accuracy, labor supply choices, and others. The recruiting stage involves cross-subject variation in wage offer (low vs high) and firm culture messaging (pro-diversity vs. neutral), to study how these affect the distribution of unobserved worker characteristics among workers who select into the firm’s labor pool by accepting its employment offer. This project aims to crack open a critical black box. We believe it will be the first to closely examine the impacts of “Diversity and Inclusion” messaging on job-seeker selection, and the effects of gender composition of work teams in a real workplace on a broad range of around coworker interaction and labor behaviors, across both remote-work and in-person work settings.