Towards Sensitivity Analysis: A Workflow

Establishing causal claims is one of the primary endeavors in sociological research. Statistical causal inference is a promising way to achieve this through the potential outcome framework or structural causal models, which are based on a set of identification assumptions. However, identification assumptions are often not fully discussed in practice, which harms the validity of causal claims. In this article, we focus on the unmeasurededness assumption that assumes no unmeasured confounders in models, which is often violated in practice. This article reviews a set of papers in two leading sociological journals to check the practice of causal inference and relevant identification assumptions, indicating the lack of discussion on sensitivity analysis methods on unconfoundedness in practice. And then, a blueprint of how to conduct sensitivity analysis methods on unconfoundedness is built, including six steps of proper choices on practices of sensitivity analysis to evaluate the impacts of unmeasured confounders.