Longitudinal studies can provide significant insights into how students develop competence in a topic or subject area over time. However, there are many barriers, such as retention of students in the study and the complexity of data analysis, that make these studies rare. Here, we present how a statistical framework, discrete-time survival analysis, can help overcome these barriers to longitudinal assessment studies using data from our research on students' understanding of structure-property relationships in chemistry. In the study presented, we administered the "Implicit Information from Lewis Structures Instrument" (IILSI)--an instrument designed to elicit from students what information can be predicted using a Lewis structure--to three cohorts of students at five time points over a two-year period, throughout their general chemistry and organic chemistry courses, to determine the ways in which student ideas about structure-property relationships evolved. Using survival analysis, we were able to identify both "whether" and "when" learning occurred over the two-year period. The single model also allowed us to construct trajectories to determine the ways in which students developed the ideas that underlie structure-property relationships in chemistry.