Data manipulation: Calculating the mean of two or more numbers and putting it back into the data set

[Max Shepard]

Dear Statalists,

i generally want to calculate the AAR from AR. In order to do so i have generated a Cluster_ID which match two firms with overlapping event days. Now i want Stata to take the value of AR_MM2 if AAR2_cluster matches and calculate the mean of this and then give it back in another variable. See table for the two different companies, both have the same cluster_id the model adds up their AR_MM2 values and gives back the average to both firms.

co_id	AR_MM2	Cluster_ID	AAR2_MM
1	5	1	3
2	1	1	3

How do i calculate this?

My code did not work:

gen AAR_MM2=AR_MM2 if Cluster_ID_2==.
by AAR2_cluster: replace AAR_MM2= (sum(AR_MM2)/2)

Here the dataset for more infos.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(id AR_MM2R AAR2_cluster)
511            .  1
113            .  1
511            .  2
113            .  2
511            .  3
113            .  3
511            .  4
113            .  4
113            .  5
511            .  5
113            .  6
511            .  6
511            .  7
113            .  7
511            .  8
113            .  8
113   .005302138  9
511  -.003760204  9
113  -.008428667 10
511  .0040853457 10
113  -.032759704 11
511 -.0007772278 11
113   .014768788 12
511   .004597272 12
113  -.018418241 13
511    .00586128 13
113            . 14
511            . 14
113            . 15
...
558            . 27
575            . 28
558            . 28
575            . 29
558            . 29
558 -.0035255456 30
575   .003658384 30
575   .007337487 31
558   .004893153 31
558   .008426867 32
575  -.016584055 32
575   .006924237 33
558  -.003346003 33
558   .003431862 34
575  -.006588891 34

end

Thank you for your help!
Best,
Max

[Andrew Musau]

bys Cluster_ID: egen AAR2_MM= mean(AR_MM2)

[Max Shepard]

Worked perfectly! Thank you.