you have at least two alternatives. One is to set the value to the upper (or lower) limit and then use another variable to indicate the variable was "censored". The second alternative is to convert them to a second type of missing value (say allows 28 different types, if I remember correctly). You can not ignore or delete these values. You're best bet is to use statistics that are designed to work on censored data. Fred

"Dennis G. To: SAS-L@LISTSERV.UGA.EDU Fisher" cc: <dfisher@CSULB.ED Subject: lab data??? U> Sent by: "SAS(r) Discussion" <SAS-L@LISTSERV.U GA.EDU>

12/10/2001 02:10 PM Please respond to "Dennis G. Fisher"

I have a question for people who work with data from lab tests. Some of the data I am working with is reported as >50000 or >850000 for examples. These are values that exceeded the ability of the lab to obtain exact values. I cannot put < or > as numerical values for analysis purposes as SAS will interpret them as character values. I am leaning toward just deleting those values (converting them to . or missing). Is this the right way to do this???? If not, why not and what should I do instead???? TIA

Dennis Fisher

-- Dennis G. Fisher, Ph.D. Director Center for Behavioral Research and Services 1090 Atlantic Avenue Long Beach, CA 90813 562-495-2330 562-983-1421 fax