Calsculate the Household production diversity

[Mwanga Ronald]

I have the following data hhid group
11001 Other fruits
11001 legumes nuts and seeds
11001 legumes nuts and seeds
11001 legumes nuts and seeds
11001 Oils and fats
11002 white roots and tubers
11002 starchy plants
11002 legumes nuts and seeds
11002 starchy plants
11002 starchy plants
11002 legumes nuts and seeds
11002 starchy plants
11011 starchy plants
11011 legumes nuts and seeds
11011 starchy plants
11016 Other fruits
11016 starchy plants
11016 legumes nuts and seeds
11016 starchy plants
11016 legumes nuts and seeds
11018 white roots and tubers
11018 starchy plants
11018 legumes nuts and seeds

I want to gen another variable production showing the different crops grown per household from the variable group which is the group of food for example the house household number 11001 should be 3. Then collapse by hhid

Thanks

[Mwanga Ronald]

Helo Statalists,
Using the dataex command the data is as follows
[CODE]
* Example generated by -dataex-. To install: ssc install dataex
clear
input float id long(hhid group)
1 11001 3
1 11001 7
1 11001 7
1 11001 7
1 11001 2
2 11002 9
2 11002 8
2 11002 7
2 11002 8
3 11002 8
3 11002 7
3 11002 8
4 11011 8
4 11011 7
4 11011 8
6 11016 3
6 11016 8
6 11016 7
7 11016 8
7 11016 7
8 11018 9
8 11018 8
8 11018 7
8 11018 2
9 11019 7
9 11019 9

I wish to count the number of categorised food groups per household and create the variable Household_Production_Score,then collapse by household.
Hope this is clear.
Thanks
Mwanga Ronald

[Hemanshu Kumar]

Welcome to Statalist! Here is some code that I think does what you ask. You can ignore the top part before the comment, that is just to re-create your data.

clear
input int hhid str30 group
11001 "Other fruits"
11001 "legumes nuts and seeds"
11001 "legumes nuts and seeds"
11001 "legumes nuts and seeds"
11001 "Oils and fats"
11002 "white roots and tubers"
11002 "starchy plants"
11002 "legumes nuts and seeds"
11002 "starchy plants"
11002 "starchy plants"
11002 "legumes nuts and seeds"
11002 "starchy plants"
11011 "starchy plants"
11011 "legumes nuts and seeds"
11011 "starchy plants"
11016 "Other fruits"
11016 "starchy plants"
11016 "legumes nuts and seeds"
11016 "starchy plants"
11016 "legumes nuts and seeds"
11018 "white roots and tubers"
11018 "starchy plants"
11018 "legumes nuts and seeds"
end

* start from here

egen byte tag = tag(hhid group)
collapse (sum) production = tag, by(hhid)

which produces:

  +------------------+
  |  hhid   produc~n |
  |------------------|
  | 11001          3 |
  | 11002          3 |
  | 11011          2 |
  | 11016          3 |
  | 11018          3 |
  +------------------+

Thank you for providing a sample of your data. But in the future, could you please do this using the dataex command, as advised in the Statalist FAQ? Thank you!

Edit: crossed with #2. Thanks for using dataex in your very first thread on Statalist!

[Andrew Musau]

egen tag= tag(hhid group)
bys hhid: egen wanted= total(tag)

The tag() function of egen will tag exactly one combination of household and group and the total() function will take the sum. See

h egen

[Nick Cox]

See also the article on distinct observations in the Stata Journal.

https://journals.sagepub.com/doi/pdf...867X0800800408

Andrew Musau's method is discussed on p.563.

[Mwanga Ronald]

Thanks alot Hemanshu Kumar your code worked and Nick Cox thanks for the reference.