This case traces the steps that we took to bring our study from initial conception to publication. The study uses a unique dataset that links the official voting records of 18- to 30-year-olds and their parents to data about their social background characteristics. We explain how we went from testing a loosely connected set of expectations based on the political socialization literature to framing our study as a test of two major theories about the parent–child link in voting. This process necessitated a thorough re-thinking of our research design. We describe how we progressed from estimating straightforward logistic regression models to performing a causal mediation analysis. We emphasize the importance of understanding the assumptions that underlie causal mediation analysis and of incorporating robustness checks. Finally, our case illustrates how the peer-review process stimulated us to develop a stronger theoretical framework and a more sophisticated research design.