Over the last decade, there have been many suggestions to improve how scientists answer their questions, but far fewer attempt to improve the questions scientists are asking in the first place. The goal of the paper is then to examine and summarize synthesize the evidence on how to ask the best questions possible. First is a brief review of the philosophical and empirical literature on how the best science is done, which implicitly but not explicitly mentions the role of psychology and especially cognitive conflict. Then we more closely focus on the psychology of the scientist, finding that they are humans, engaged in a meaning making process, and that cognitive conflict is a necessary input for any learning or change in the system. The scientific method is, of course, a specialized meaning making process. We present evidence for this central role of cognitive conflict in science by examining the most discussed scientific papers between 2013 and 2017, which are, in general, controversial and about big problems (e.g., whether vaccines cause autism, how often doctors kill us with their mistakes). Toward the end we discuss the role of science in society, suggesting science itself is an uncertainty reducing and problem solving enterprise. From this basis we encourage scientists to take riskier stances on bigger topics, for the good of themselves and society generally.