One of the key challenges graduate students face is how to come up with a good rationale for their theses. Unfortunately, the methods literature in and beyond political science does not provide much advice on this important issue. While focusing on how to conduct research, this literature has largely neglected the question of why a study should be undertaken. The limited discussions that can be found suggest that new research is justified if it (1) fills a ‘gap’; (2) addresses an important real-world problem; and/or (3) is methodologically rigorous. This article discusses the limitations of these rationales. Then, it proposes that research puzzles are more useful for clarifying the nature and importance of a contribution to existing research, and hence a better way of justifying new research. The article also explores and clarifies what research puzzles are, and begins to devise a method for constructing them out of the vague ideas and questions that often trigger a research process.