Counting non-missing values - long panel data

[Bertha Rohenkohl]

Hello,

I am currently struggling with the following issue:

I have a big dataset with around 12,000 individuals. For each individual, I have information on 18 consecutive years. Each individual is matched to his/her father.
I would like to create a variable that allows me to assess the number of individuals (small n) who have non-missing values for income (of the individual) and non-missing values of father's income at the same time.

In summary, I would like to calculate the small n for certain variables, applying certain restrictions. The product should be something like the small n from xttab, but I need to do it manually.
I thought of creating dummy variable that will be equal to 1 if the individual has at least one observation for income (over the 18 years) and if there is at least one observation for the income of the father (over 18 years); or 0 if income and father's income are missing for the 18 years.

I am already struggling to create the dummy for each case separately, never mind with two restrictions.
Could anyone help?

Thank you.

[Nick Cox]

Assume variables

id year income fincome

then you want -- it seems --

egen totalbyyear = total(!missing(income) & !missing(fincome)), by(year)
egen totalbyid = total(!missing(income) & !missing(fincome)), by(id)

If father's income is not aligned, see e.g. http://www.statalist.org/forums/foru...in-a-household

If this doesn't help (enough), please go back to FAQ Advice especially #12 and provide a data example.

I don't see that you need dummies at all. You can count satisfactory observations directly. You just need to build on true = 1, false = 0.

[Bertha Rohenkohl]

Hello Nick,

Thanks for the quick reply. It works perfectly!