Currently I'm examining the duration of residence of households. I have a database at my disposal that indicates how long a specific household resides in its current home. I want to explain their duration of residence by the household characteristics, locational characteristics, etc. The database is updated every day.
The problem is that the structure of the dependent variable (length of stay) does not allow for a completed housing spell. Once a household moves out of its home (and thus completes its spell) it falls out of the database. So I only have information about households that currently live in a house that is followed by this database. In this way, I can only tell if the household has lived there until now for a month, 2 years or 6 years for example. However it is uncertain how long they keep living there in the future. In other words, every observation is right censored.
What type of duration/survival model suits this type of data the best? I have read about Cox proportional hazard models, Tobit models, etc., but I am not sure which one is the best applicable. I intend to use Stata as statistical program. I found an article* in which the authors seem to have coped with the same issue, but it didn't became clear to me how they did it.
Does anyone have a suggestion for me? Thanks in advance.
*Van der Vlist et al. (2002) Residential mobility and local housing-market differences, Environment and Planning A v34 n7 (7/2002): 1147-1164.
