Generating variable with conditions

[Laura Montolio]

Dear all,

I have a eu dummy variable that takes value 1 since the year of incorporation in the EU. I would like to generate a new dummy variable that takes value 1 if eu equals 1 for any year of the period, that is, a variable that indicates if the country is or not in the EU. I do not know how to indicate stata the "for any year of the period" issue.
I tried with the following code but got exactly the same eu variable.

gen EU_new_d=0
    replace EU_new_d=1 if eu_d==1&year==1995|eu_d==1&year==1996|eu_d==1&year==1997|eu_d==1&year==1998|eu_d==1&year==1999|eu_d==1&year==2000|eu_d==1&year==2001|eu_d==1&year==2002|eu_d==1&year==2003|eu_d==1&year==2004|eu_d==1&year==2005|eu_d==1&year==2006|eu_d==1&year==2007|eu_d==1&year==2008|eu_d==1&year==2009|eu_d==1&year==2010|eu_d==1&year==2011|eu_d==1&year==2012|eu_d==1&year==2013|eu_d==1&year

Thanks for any suggestions.

Best regards,
Laura

[Carlo Lazzaro]

Laura:
I'm not sure I've got you right, as it seems that you already have what you're after.
Anyway, a tempative answer would imply something along the following lines:

. set obs 10

. g  country="whatyouwant"

. g year=_n

. g eu=1 in 5/10

. replace eu=0 if eu==.

. bysort country year: gen new_eu=1 if eu==1

. replace new_eu=0 if new_eu==.

[Nick Cox]

Your code as copied here ends

& year 

and year as a condition by itself always counts as true (given that I guess you have no year 0).

Your code boils down to something more like

gen EU_new_d = eu_d==1 & inrange(year, 1995, 2014)

See for more guidance

https://www.stata-journal.com/sjpdf....iclenum=dm0058

https://www.stata.com/support/faqs/d...rue-and-false/

[Laura Montolio]

I´m sorry but I do not understand what you suggest. With the previous commnad my new EU dummy looks like the old one, as it is shown in the attached image. I would like it to take value 1 for the whole period (1995-2014) whenever my old eu variable takes the value 1 for any year of that period. Hope to have explained myself better now.
Best regards,
Laura

[Nick Cox]

I think #3 does you what you ask. Perhaps you're replying to Carlo in #2.

[Laura Montolio]

Hi Nick,

Thanks for your suggestion. However, when I apply what you suggest the new variable takes the same values as in the image I´ve attached. It should take the value 1 since 1995 to 2014 for the example shown in the image and not just for 2013-2014 as in the old eu variable.

[Nick Cox]

I don't see your point. If you want as one necessary condition

eu_d == 1 

then the values shown in the image (please don't show images; FAQ Advice #12) are exactly what are expected.

[Carlo Lazzaro]

Laura:
I propose another temptative answer:

set obs 10
g country="whatyouwant"
g year=_n
g eu=1 in 5/10
replace eu=0 if eu==.
bysort country: egen new_eu=max(eu)
list

     +----------------------------------+
     |     country   year   eu   new_eu |
     |----------------------------------|
  1. | whatyouwant      1    0        1 |
  2. | whatyouwant      2    0        1 |
  3. | whatyouwant      3    0        1 |
  4. | whatyouwant      4    0        1 |
  5. | whatyouwant      5    1        1 |
     |----------------------------------|
  6. | whatyouwant      6    1        1 |
  7. | whatyouwant      7    1        1 |
  8. | whatyouwant      8    1        1 |
  9. | whatyouwant      9    1        1 |
 10. | whatyouwant     10    1        1 |
     +----------------------------------+

[Laura Montolio]

Thank you so much Carlo. It provides exactly the variable I need.

Kind regards,
Laura