st: RE: Treat zeros in logarithms

["Verkuilen, Jay" <JVerkuilen@gc.cuny.edu>]

Doing half DL replacement shows up as a simple but generally inadvisable
strategy to handle left-endpoint non-detects. 

Use -intcens- and choose a reasonable detection limit for the censoring
point. Then vary that limit to make sure you're results aren't sensitive
to the choice.

I am (finally) finishing up my paper on this topic but the lognormal
distribution is VERY sensitive to the replacement value you pick. Other
distributions are much less sensitive and generally you should consider
them; -intcens- provides easy access to them and lets you run a
left-censored model, which is probably what you want anyway.



*
*   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/