st: egen rowsum, and missing values

[Nirali Shah <nmshah@jhsph.edu>]

Hi,

I find the egen commands to be very useful, especially the rowsum  
operation, which treats missing values as 0.  However, I accidentally  
discovered an aspect of this operation which has me baffled.

In my data set, I have two types of missing values - missing due to  
non-response, and missing due to a skip pattern.  I'd like to keep  
them separate, for tabulations and other purposes.  The missing due to  
a skip pattern I have coded as ".a", and the regular missing values as  
"."  (As numbers, not text).

Egen rowsum appears to treat my ".a" missing values as the number 56.   
Does anyone know why it does this? Do you have any suggestions for  
getting around this and still allowing me to keep the two different  
types of missing data codes?

Much appreciated,
Nirali
____________________________________________________
Nirali Shah
PhD Candidate, International Health, Health Systems Division
Bloomberg School of Public Health
Johns Hopkins University

*
*   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/