whatifbandit - Analyzing Randomized Experiments Using Multi-Arm Bandits

Simulates response-adaptive experimental trials using Multi-Arm Bandits. Adaptive robust estimators defined in Hadad et al. (2021) <doi:10.1073/pnas.2014602118> and Offer-Westort et al. (2021) <doi:10.1111/ajps.12597> are used to robustly estimate conditional expectations and treatment effects. Provides significant simulation customization options for imperfect information, non-stationary bandits, and increased exploration strategies for assignments.