RE: st: probable error, "weights invalid" using stset

["Brown, Elizabeth" <ebrown@prgs.edu>]

Hi Austin,  and thank you for your reply. Just a few follow-on questions; 1) Why do you recommend using first-observed and not last-observed weight? If weights adjust for attrition, wouldn't I want to use the last-observed weight for an individual, just prior to their leaving the sample? The last observed weight seems more likely to account for individual probability of leaving the sample, no? ; 2) As a clarification, you're suggesting that weights would be identified as probable error if they were 0 or missing - are there any other reasons you can think of?; 3) Last, thanks for the link to the Jenkins primer on survival analysis -- I've been traipsing all over these pages. Very helpful, thanks! 

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Austin Nichols
Sent: Monday, October 31, 2011 3:21 PM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: probable error, "weights invalid" using stset

Elizabeth <ebrown@prgs.edu>:
Use first-observed weights, not last, and make sure none are zero or missing. Or spend some more effort to define exactly what population you want your analysis sample to be representative of. Also note 2009 PSID data is available.  Also, you have discrete time data, not continuous time data--people are poor or not in one year, not at each point in time, at least using US definitions.  So use a discrete time hazard model; see e.g.
http://www.iser.essex.ac.uk/resources/survival-analysis-with-stata-module-ec968

On Mon, Oct 31, 2011 at 2:51 PM, Brown, Elizabeth <ebrown@prgs.edu> wrote:
> Dear all,
> I encountered a PROBABLE ERROR using stset that I'm unsure how to investigate/resolve. After stset, I get the following output, with a probable error with respect to 219 "weights invalid". What would cause weights to be invalid? How can I fix this?
>
> I'm using data from the Panel Study of Income Dynamics to examine multiple poverty transitions of families over the 1968 - 2007 period.
>
> With respect to how the weights were prepared, I first: 1) replaced any missing family weight values with family_weight[_n-1], and 2) made weights consistent within id by replacing  family weight with the last observed family weight value. In previous attempts to stset the data using weights, stata complained when weights were inconsistent within id.

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

__________________________________________________________________________

This email message is for the sole use of the intended recipient(s) and
may contain confidential information. Any unauthorized review, use,
disclosure or distribution is prohibited. If you are not the intended
recipient, please contact the sender by reply email and destroy all copies
of the original message.


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/