Longitudinal research designs involve data collection at multiple time points to measure change over time. Therefore, identification of the same respondents is essential at each time point so that data from the same respondents can be matched for comparison over time. Subject-generated identification codes permit an anonymous means to track respondents over multiple data collection points. This article describes the evolution of subject-generated identification codes, techniques to improve respondent match rates, and the authors' experience using this mechanism in a longitudinal study of staff registered nurses working in hospitals. Challenges, recommendations, and implications for using subject-generated identification codes are discussed. (Contains 5 tables and 1 figure.)