counting total of variable based on other variable

[Thomas Kiekebosch]

Hello,

i want to count the total M&A's (labeled as 1 in maactive) of a firm (based on gvkey) en generate a new variable. I tried the code

egen totalma = count (maactive ==1), by (gvkey)

but if I use this code, it counts all of the observations on maactive based on gvkey but I only need the ''1'' of the variable maactive.

Does someone know what I am doing wrong?

[Ken Chui]

Use -sum- instead:

clear
input gvkey maactive
1078 1
1078 0
1078 0
1084 1
1084 1
1084 0
end

* egen totalma = sum(maactive), by(gvkey)
* ^ Revised due to sum being non-documented. Use total()

egen totalma = total(maactive), by(gvkey)

Results:

     +----------------------------+
     | gvkey   maactive   totalma |
     |----------------------------|
  1. |  1078          1         1 |
  2. |  1078          0         1 |
  3. |  1078          0         1 |
  4. |  1084          1         2 |
  5. |  1084          1         2 |
  6. |  1084          0         2 |
     +----------------------------+

[Thomas Kiekebosch]

Thanks! Works fine!

[Nick Cox]

You need total().

count() counts non-missing values. As the result of a true-or-false comparison is 1 or 0, that is non-missing either way, with the result you report.


sum() is an non-documented equivalent of total(), as from Stata 9. Given the choice, use documented features rather than what is not documented.

[Ken Chui]

You need total().

count() counts non-missing values. As the result of a true-or-false comparison is 1 or 0, that is non-missing either way, with the result you report.


sum() is an non-documented equivalent of total(), as from Stata 9. Given the choice, use documented features rather than what is not documented.

Thanks Nick, I revised my code answer.