Biostatistics Seminar Series: Michael Baiocchi, PhD

Title: Several major wins in the fight against sexual assault - statistical insights from the front lines.Abstract: In the spring of 2013, Brad Efron approached me with a challenge. An NGO operating in the slums of Nairobi, Kenya had collected data for what appeared to be a promising classroom-based sexual assault prevention program. They passionately believed their program could cut the rate of rape in half (for context: the rate stands at 25% per year in the toughest slums in Nairobi) but few believed them. This initial project escalated rapidly into a much more involved line of research - notable events including: having to design a randomized trial that could withstand a contested presidential election and an election-nullification, being present for a shootout, having lunch with a gang leader who is famous for kidnapping a politician, and being asked to move to Nairobi and become the head of an NGO in order to complete our randomized trial. None of those events will be addressed in this talk. Instead, we will focus on how a small team of statisticians has found itself present at the early stages of an academic discipline that is transitioning from being predominately about advocacy and theory into a discipline grounded in empirical science. I'll introduce the intervention and sketch out the two cluster-randomized trials we have completed (this will be the first time we share results from our newest trial). We'll describe the handful of methodological innovations we've had to make in order to do our work: (i) randomization in the presence of interference, (ii) explicitly addressing one of the biggest problems with randomization in cluster-randomized trials, and (iii) developing a framework for causal inference with free-text. Inspired by Tom Ten Have's admonishment that "statisticians must be the guardians of the scientific method," a running theme in the talk will be the role that statisticians can play in unifying disparate disciplines and establishing best practices through lowering barriers to experiencing, replicating, and deploying our methods. Time permitting, I'll discuss how our efforts in Kenya have translated to a program we rolled out at Stanford in the spring of 2018.