define size dummy of city

[River Huang]

Dear All, the following question is from here (https://bbs.pinggu.org/forum.php?mod...=1#pid57703787). The data set is

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(id year pop)
3 2003  200
3 2004  500
3 2005  600
3 2006  700
1 2003  600
1 2004  800
1 2005  900
1 2006 1000
2 2003  600
2 2004  900
2 2005 1200
2 2006 1300
4 2003  400
4 2004  500
4 2005  600
4 2006  700
5 2003  900
5 2004  950
5 2005 1000
5 2006 1100
end

For each id, a "big" dummy is defined to be 1 (for all the observations in this id) if the population (pop) is larger than 500 in the year 2016, 0 otherwise. Similarly, a "medium" dummy is defined if 200 < pop < 500, and a "small" dummy is defined as pop < 200. Any suggestions? Thanks.

[Nick Cox]

Technique is shown by

egen big = total(year == 2016 & inrange(pop, 500, .)), by(id)

as the condition

year == 2016 

is presumably satisfied at most once for each city -- not at all in the example data -- and

inrange(pop, 500, .) 

must be satisfied at the same time.


Note that max() should work as total() does.

Note that this is an FAQ!

How do I create a variable recording whether any members of a group (or all members of a group) possess some characteristic?

https://www.stata.com/support/faqs/d...ble-recording/

See also https://www.stata-journal.com/sjpdf....iclenum=dm0055 for a survey of related technique.

[River Huang]

Dear Nick, Thanks again for your helpful suggestions.