Designing Experiments for Impact

This is a team-based course where students will work on a project to design and carry out an experiment intended to drive social impact in collaboration with a partner organization. The first few weeks will include lectures, hands-on tutorials, and labs designed to guide students through the process of experimental design in the digital context. Special topics include designing and selecting outcome measures that capture the impact of interventions; multi-stage experiments with applications to chatbots; learning how treatment effects vary across subgroups; adaptive experiments using bandits and artificial intelligence; and estimation of policies that target treatments based on subject characteristics. Experiments may be conducted on the customer base of a partner organization through their digital applications or on recruited subjects, such as subjects recruited to interactive chatbots. The teaching team will provide templates and technical assistance for designing and running the experiments. Students from different disciplinary backgrounds will be assigned roles to work in teams on the project. This course is part of the GSB's Action Learning Program, in which you will work on real business challenges under the guidance of faculty. In this intensive project-based course, you will learn research-validated foundations, tools, and practices; apply these tools and learnings to a real project for an external organization; create value for the organization by providing insights and deliverables; and be an ambassador to the organization by exposing them to the talent, values, and expertise of the GSB. You will also have the opportunity to gain practical industry experience and exposure to the organization, its industry, and the space in which it operates; build relationships in the organization and industry; and gain an understanding of related career paths. Prerequisites: Some experience with statistical analysis and the R statistical package. Students with less experience will have an opportunity to catch up through tutorials provided through the course. Non-GSB students are expected to have an advanced understanding of tools and methods from data science and machine learning as well as a strong familiarity with R, Python, SQL, and other similar high-level programming languages. Cardinal Course certified by the Haas Center.