How to calculate ratio with two conditions

[Aqib Yousuf]

I want to calculate completion rate for two conditions.

* Example generated by -dataex-. For more info, type help dataex
clear
input float(grid treatment_status enrolled) byte dropout
301 1 1 .
209 1 1 1
209 1 1 1
209 1 1 1
206 1 1 1
301 1 1 1
301 2 1 .
204 2 1 .
207 2 1 1
209 2 1 .
end
label values treatment_status treatment_lbl
label def treatment_lbl 1 "[T1] Information Only", modify
label def treatment_lbl 2 "[T2] Out of Village Mobilization (OVM)", modify

I want to be able to compute the completion rate (dropout != 0 ) for each grid in each treatment arm.

Thank you!

[Carlo Lazzaro]

Aqib:
do you mean something along the following lines?:

. bysort grid treatment_status: egen wanted=sum( enrolled ) if dropout!=.

. list

     +-----------------------------------------------------------------------------+
     | grid                         treatment_status   enrolled   dropout   wanted |
     |-----------------------------------------------------------------------------|
  1. |  204   [T2] Out of Village Mobilization (OVM)          1         .        . |
  2. |  206                    [T1] Information Only          1         1        1 |
  3. |  207   [T2] Out of Village Mobilization (OVM)          1         1        1 |
  4. |  209                    [T1] Information Only          1         1        3 |
  5. |  209                    [T1] Information Only          1         1        3 |
     |-----------------------------------------------------------------------------|
  6. |  209                    [T1] Information Only          1         1        3 |
  7. |  209   [T2] Out of Village Mobilization (OVM)          1         .        . |
  8. |  301                    [T1] Information Only          1         .        . |
  9. |  301                    [T1] Information Only          1         1        1 |
 10. |  301   [T2] Out of Village Mobilization (OVM)          1         .        . |
     +-----------------------------------------------------------------------------+

.

[Aqib Yousuf]

No. For each grid, I want the completion rate in each treatment arm. So if [T2] is able to stay enrolled and not dropout. I want the rates for all [T2] that were succesfully able to stay enrolled. I hope I am clear now.

[Carlo Lazzaro]

Aqib:
sorry, but I do not follow you.
In your example, there's no such a thing as -dropout=0-.
In addition, could you please create a column with the values you 'd want to see? Thanks,

[Aqib Yousuf]

Yes, sorry. That is a mistake. Please ignore the dropout !=0.

I want to see completion rate of each grid in each treatment arm. Specifically, if a grid and treatment arm is able to have dropout ==. (missing) then should have a completion rate of 1 or 100%, and so on. I hope this makes things clear now.

[Carlo Lazzaro]

Aqib:

.  bysort grid treatment_status: egen wanted=sum( enrolled ) if dropout==.


. list

     +-----------------------------------------------------------------------------+
     | grid                         treatment_status   enrolled   dropout   wanted |
     |-----------------------------------------------------------------------------|
  1. |  204   [T2] Out of Village Mobilization (OVM)          1         .        1 |
  2. |  206                    [T1] Information Only          1         1        . |
  3. |  207   [T2] Out of Village Mobilization (OVM)          1         1        . |
  4. |  209                    [T1] Information Only          1         1        . |
  5. |  209                    [T1] Information Only          1         1        . |
     |-----------------------------------------------------------------------------|
  6. |  209                    [T1] Information Only          1         1        . |
  7. |  209   [T2] Out of Village Mobilization (OVM)          1         .        1 |
  8. |  301                    [T1] Information Only          1         .        1 |
  9. |  301                    [T1] Information Only          1         1        . |
 10. |  301   [T2] Out of Village Mobilization (OVM)          1         .        1 |
     +-----------------------------------------------------------------------------+

.