Leonelo - You might be able to use -intreg- to fit an interval-censored
model for your log-transormed data.
Al Feiveson
-----Original Message-----
From: Leonelo Bautista [mailto:[email protected]]
Sent: Monday, May 19, 2003 8:56 AM
To: [email protected]
Subject: st: nondetectable values
Hello,
I'm analyzing a dataset with some nondetectable values (too low and too
large) in the outcome variable (log-normally distributed). I know that
maximum likelihood estimates of the mean and standard error could be
calculated (Hornung, R.W., and L.D. Reed: Estimation of average
concentration in the presence of nondetectable values. Appl. Occup. Environ.
Hyg. 5:46�51 (1990). Can anybody tell me how to get those estimates using
Stata? Also, I need to do a multiple regression analysis with this data. It
seems that one way to use the data on subjects with low non-detectable
values is to replace them by the limit of detection divided by 2 or by the
square root of 2. But, how should I handle high non-detectable values? Any
suggestions would be appreciated (including relevant references).
Thanks in advance,
Leonelo E. Bautista, MD, MPH, DrPH
Assistant Professor
Department Preventive Medicine & Biometrics
Uniformed Services University of the Health Sciences
4301 Jones Bridge Road
Bethesda, MD 20814
Room A1039
Phone: (301)2953712
Fax:(301)2951854
*
* 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/
*
* 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/
