preparing longitudinal survey data for survival analysis

[Rune Lomholt]

Hello anyone

I am using stata for social science research.

I want to conduct survival analysis form longitudinal survey data on university students. I understand the theory of the analysis, but i am confused about an aspect of preparing the data for analysis. I have collected data about reasons for dropout in a survey, distributed to the same students two times during their study period. In this period, they either persist og drop out.

1. How do i merge the to survey datasets, telling stata that data from the second dataset is a second measurement of the same variables as the one in the first dataset, and;
2. how do i include the change in these variables, in my survival analysis?

I have not included examples, since this is merely a questions about preparing the data, and not how to conduct the analysis itself.

[David Radwin]

If you have observations from only two points in time, it sounds more like a cross-sectional method like logit or probit rather than a survival analysis method like the Cox proportional hazards model (stcox). Either way, the first step is to merge the two datasets together using a variable (typically a number, but perhaps a name or something else) that uniquely identifies each student in each dataset. If you have more than two time points per individual, the next step is to reshape the data from wide format to long format.