At Facebook, we have access to a plethora of research tools that should, hypothetically, enable us to design products for everyone. And yet, as we are tasked with building a product that works for a diverse group of many people, we often end up designing only for an imagined user, based on our own experiences and preconceptions. In this presentation, we discuss the main bias-related issues that pop up in our research (such as language, geography, accessibility, etc.), present case studies of how Facebook researchers have tackled various types of bias, and provide approaches/resources that can help researchers get around biases and take their research to the next level.
