I am using multiple-record-per-case data (stsplit at failures) to model
rates of exit and of change for kids in foster care using models of the
general form r(j)k(t)= exp[BX(t)] ((j) and second (t) are subscripted)
where X(t) is a vector of time-varying variables and B is a vector of
parameters indicating the effects of the variables on rate of exit or of
change. I would like to illustrate graphically the independent effect of
selected variables on the hazard rate using point estimates from my models.
However, I am having a difficult time doing this as I can't figure out how
to get STATA to use time on the x axis (I can only get a constant rate
curve using stcurve). For example, the coefficient associated with the
variable measuring time-of-duration in previous placement is positive and
significant, which I am interpreting as indicating that with a change there
initially is a drop in the risk of change. Then, the risk of change
increases more rapidly than would have occurred simply because of the aging
of the child.(time-of-duration in previous placement is measured by a
"duration clock", which records the elapsed time in days since last change;
it takes a value of 0 until a change, at which point it records the elapsed
time since the change. If another change occurs, the clock is reset to 0
and again records the elapsed time). As I say, I would like to illustrate
this effect graphically at a different duration times in foster care but
can't figure out how to do it. Any suggestions would be greatly
appreciated. Thanks
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
