Generating new variables based on criateria

[Muhammad Abid Shahzad]

Hi,

I want to generate a new variable based on criteria.

I have the data for, let's say, 20 different regions and each region has a percentage of child, young and old persons (which equals to 100%).
It looks like this:
Regions--child-- young--old
region 1--25--20--55
region 2--50--50--00
.
.
.
region 20-- 75--20--5

Now I want to divide regions into 3 different categories, i.e., regions with a majority of children, regions with a majority of young ones, and regions with a majority of old ones.
Here each category would be represented by a dummy (0.1).

Could you please tell me, how can I do this in stata?

[Ken Chui]

Since the post did not follow the FAQ (http://www.statalist.org/forums/help) and use -dataex-, I've converted the data into code form for those who would like to help:

* Example generated by -dataex-. For more info, type help dataex
clear
input float(region child young old)
 1 25 20 55
 2 50 50  0
20 75 20  5
end

Also, "majority" is a vague term. It may be helpful to define what is the threshold, and how would you like to resolve ties observed in region 2.

[Clyde Schechter]

In a situation like region 2, which is 50% child and 50% young, is this majority children, majority young, or not majority-anything?

Added: Crossed with #2 which raises the same question.

[Muhammad Abid Shahzad]

Ken Chui thank you very much.

[Muhammad Abid Shahzad]

Clyde Schechter thank you very much for your response.

Okay, let's call that case as "not majority-anything". However, these type of cases are rare in the dataset.

[Muhammad Abid Shahzad]

By majority I mean, the highest number of individuals.

[Clyde Schechter]

foreach v of varlist child_young_old {
    gen byte majority_`v' = (`v' > 50) & !missing(`v')
}