This report contends that most military-manpower research that aims to estimate causal effects suffers from flaws that render the research invalid. I discuss the objectives and ethics of research, and I give a low-math quick guide to the main pitfalls of research, While I do not test the claim that most manpower research is flawed, I show that several relatively prominent examples of manpower research turns out to be invalid, using the pitfalls I describe. This includes some of my own prior research. I end with recommendations on how to improve the research process for the Department of Defense, which includes: requiring training of researchers (beyond what graduate school taught and which could have helped me avoid my research mistakes), having review panels assess a proposed study’s methods before funding is granted, reviewing incentives at research organizations to ensure they promote high-quality research, improving quality control at research organizations, and periodically reviewing the quality of an organization’s research.