Creating indicator variable that shows if food choices in HH were made by male or female individual

[Jacob Steckel]

Dear Statalists,
I am writing my thesis. Currently, I want to check out if one of my independent variables varies enough to conduct an analysis. To do so, I obviously have to construct that variable. The variable shall be a dummy variable indicating if the person that is responsible for food purchases of the household is female (1) or male (0).
The dataset is an unbalanced panel dataset with t=2 and I am using Stata 16.1.
I have the following information that should suffice to construct the variable, but I don't know how exactly and starting to get a bit mad after trying and searching for an answer for too long:

- hhid (household identifier)
- membid (identifier of household member)
- sex (female=1)
- respshop (membid of the person that is responsible for shopping)

Here you can find my dataex:

* Example generated by -dataex-. For more info, type help dataex
clear
input long hhid double membid int year float(sex respshop)
101015 1 2012 1 1
101045 1 2012 0 2
101060 2 2012 1 2
101075 1 2012 0 2
101075 2 2012 1 2
101092 2 2015 1 2
101092 2 2012 1 2
101105 2 2012 1 2
101120 1 2012 1 1
101125 1 2015 0 1
101133 2 2015 1 2
101150 1 2012 0 2
101150 2 2012 1 2
101152 1 2015 0 1
101152 2 2015 1 1
101160 1 2015 0 3
101160 3 2015 1 3
101165 1 2012 1 1
101180 1 2012 0 1
101187 1 2015 1 1
101232 2 2015 1 2
101233 1 2015 1 1
102012 1 2012 0 2
102012 2 2012 1 2
102042 1 2012 1 1
102042 1 2015 1 1
102057 1 2012 0 1
102072 1 2012 1 1
102117 1 2012 0 1
102147 1 2012 0 1
102177 1 2012 0 1
102192 2 2012 1 2
102198 1 2012 0 1
102207 1 2012 0 1
102224 1 2012 0 2
102224 2 2012 1 2
102224 2 2015 1 2
102235 1 2015 0 2
102235 2 2015 1 2
102237 1 2015 0 2
102237 1 2012 0 2
102237 2 2015 1 2
102237 2 2012 1 2
102238 1 2015 0 2
102238 2 2015 1 2
102242 2 2015 1 2
102248 1 2015 0 2
102248 2 2015 1 2
102251 2 2015 1 2
102252 2 2012 1 2
102253 2 2015 1 2
102255 2 2015 1 2
102258 2 2015 1 2
102259 1 2015 0 2
102259 2 2015 1 2
102261 1 2015 1 1
102262 1 2015 1 1
103005 1 2015 0 2
103005 1 2012 0 2
103005 2 2012 1 2
103005 2 2015 1 2
103020 1 2012 0 2
103020 2 2012 1 2
103020 2 2015 1 2
103035 1 2012 0 2
103035 2 2012 1 2
103050 1 2012 0 1
103050 1 2015 0 1
103079 2 2015 1 2
103079 2 2012 1 2
103094 2 2012 1 2
103109 1 2012 0 1
103270 1 2015 0 2
103270 2 2015 1 2
103273 2 2015 1 2
103286 1 2015 1 1
104011 1 2015 0 2
104011 2 2015 1 2
104014 1 2015 0 2
104014 1 2012 0 2
104014 2 2012 1 2
104014 2 2015 1 2
104023 1 2015 1 1
104028 1 2015 0 2
104028 2 2015 1 2
104038 2 2015 1 2
104045 1 2012 0 2
104045 2 2012 1 2
104045 2 2015 1 2
104051 2 2015 1 2
104053 1 2015 1 1
104058 1 2015 0 2
104058 2 2015 1 2
104059 1 2015 0 1
104059 1 2012 0 1
104061 1 2015 1 1
104071 2 2015 1 2
104073 2 2015 1 2
104083 1 2015 0 2
104083 2 2015 1 2
end
label values year year

I would really appreciate your help also if this might be very trivial for some of you, as I am currently stuck.
Kind regards, Jacob

[Nick Cox]

egen female_resp = max(sex & resp), by(hhid year)

will be 1 if any female is responsible and 0 otherwise,

Notes:

1. The code won’t check if a male as well as a female is declared responsible.

2. Your specification does not spell out whether you want to treat years separately, as I have guessed here. If you don’t, leave out year from the code.

[Jacob Steckel]

Thank you so so much! You made my day . To be honest I just did not think about using the max command for two variables at the same time...