*Getting example data

use http://dss.princeton.edu/training/Panel101.dta, clear

*To see original data

browse

*Collapse by country

collapse (mean) y (sum) y_bin (max) x1 (min) x2 (median) x3, by(country)

*To see collapsed data

browse

/*
country	y	y_bin	x1	         x2	           x3
A	1.728e+09	8	0.499805	-1.107956	 0.9879055
B	2.139e+08	5	0.8801672 	 1.423646	-0.0129679
C	1.345e+09	7	1.446412	-1.344218	 0.4669301
D	3.641e+09	9	0.4863668	 1.541221	 0.0801717
E	7.231e+08	9	1.39114 	 1.189055	-0.0894756
F	3.559e+09	9	1.396892	-0.7556524	 0.2844574
G	1.706e+09	9	1.257879	-1.621761	 2.533169
*/

*Categorical variables need to be separated by category, this is 
*each category becomes a dummy variable (0/1). Then you can either sum or
*take the mean (giving the percentage of those in 1).
*To separate a categorical variable into different dummies type

tab race, gen(race)

*See an example here, page 57
* http://dss.princeton.edu/training/StataTutorial.pdf#page=57

*Lets say 'race' has 3 categories, your data will have three variables race1, race2 and race3,
*each corresponding to one category. In the collapse you will add

collapse (mean) y (sum) y_bin (max) x1 (min) x2 (median) x3 ///
         (mean) race1 (mean) race2 (mean) race3, by(country)

*The '///' means the command continues in the next line

*Make sure to save the data with a new name after collapsing

*NOTE:
*If you want to collapse by more than one variable you can type:

collapse .... , by(country state)

*For more details type

help collapse