I had no clue this could happen and I am horrified it does...
When I write a simple line in stata 9.0
recode x 0=1 if z>2
I get all values of x that equal 0 recoded into 1 when z equals . or .a or .b
this means Stata understands missing values as large numbers...
I have tried a couple of times and it consistently recodes considering missing values as larger than life numbers
see the example of the dataset
var1 var2 var3 var4
1 0 1 0
1 0 1 0
1 0 1 0
10 0 1 0
10 0 1 0
10 0 1 0
100 0 1 0
100 0 1 0
100 0 1 0
1000 0 1 0
1000 0 1 0
1000 0 1 0
. 0 1 0
. 0 1 0
. 0 1 0
. 0 1 0
. 0 1 0
. 0 1 0
. 0 1 0
. 0 1 0
after running the following line, which should recode only 9 values, what I get is the dataset below.
recode var2 0=1 if var1>1
var1 var2 var3 var4
1 0 1 0
1 0 1 0
1 0 1 0
10 1 1 0
10 1 1 0
10 1 1 0
100 1 1 0
100 1 1 0
100 1 1 0
1000 1 1 0
1000 1 1 0
1000 1 1 0
. 1 1 0
. 1 1 0
. 1 1 0
. 1 1 0
. 1 1 0
. 1 1 0
. 1 1 0
. 1 1 0
Is it possible that no one noticed this? And if yes, why did stata not fix it yet?
The way around it was tricking the system by recoding the missing values into some very small or negative value, temporarily, so we can make the transformations without the giant missing value bug playing its tricks...
I am astonished.
Diego
Diego G. Bassani, M.Sc., Ph.D.
Epidemiologist - Post-Doctoral Fellow
Centre for Addiction and Mental Health
Health Systems Research & Consulting Unit
33 Russell Street, Tower - T308
Toronto, ON - M5S 2S1
[email protected]
http://www.camh.net/hsrcu
Phone: +01 (416) 535 8501 # 6465
Fax: +01 (416) 979 4703
*
* 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/
