-------------------------------------------------------------------------------
help for wtdset
-------------------------------------------------------------------------------

Declare data to be Waiting Time Distribution data

wtdset eventvar exitvar [if exp] [weight], start(date) end(date) [id(varname) robust cluster(varname) scale(#)]

wtd

fweights are allowed; see help weights.

Description

wtdset declares data to be Waiting Time Distribution (wtd) data; see help wtd.

wtd displays how the data are currently declared.

Caution: wtdset is destructive in the sense that it prunes and collapses data so that subsequent wtd commands can exploit frequency weighting. This improves estimation speed dramatically, but it is the users responsibility (exclusively!) to ensure preservation of original data, see help save.

Options for wtdset

start(date) specifies the starting date with syntax acceptable by the d() function, see help dfcn. If your earliest possible observation is 1 january 1997 then the starting date should be set to 31 december 1996. In other words the starting date is time zero, with observation of dates only possible for date 1, 2,...

end(date) specifies the date of the last possible observation time.

id(varname) specifies the subject-id variable. For each subject only the first observation in the observation period is kept. If id is not specified it is assumed that each observation corresponds to a single individual.

robust specifies that the Huber/White/sandwich estimator of variance is to be used as default in place of the traditional calculation in subsequent estimations with wtdml. Although set it may be overruled in the wtdml statement using option norobust, and if not set, it can be added in the wtdml statement. robust combined with cluster() further allows observations which are not independent within cluster (although they must be independent between clusters). See [U] 23.14 Obtaining robust variance estimates and help robust.

cluster(varname) specifies that data are clustered, ie. the observations are independent across groups (clusters) but not necessarily independent within groups. varname specifies to which group each observation belongs. Specifying cluster() implies default robust variance estimation.

scale(#) defines the scale of analysis time. This can be handy for making units of rate estimates more readable (such as converting person-days to person-years).

Examples

. wtdset event exit, start(31dec1996) end(31dec1997) id(id)

. wtdset event exit, start(31dec1996) end(31dec1997) id(id) scale(365)

. wtd

Also see

Manual: [ST] stset