Counting values of a variable as a proportion of total observations on the variable

[Donald Weatherburn]

Dear colleagues and mentors, I have a STATA file in which each row is a person convicted of an offence and I’m trying to determine the proportion of people appearing in a particular court that have a court’s workload that involves a particular case. Let Oij be the number of cases of type i in court j and let Tj be the total number of cases finalised by court j over the same time period. I want to create a variable for each court which takes the value Oij/Tj*100. Is there an easy way to do this?

[Andrew Musau]

assert !missing(cases)
bysort court: gen percent=_N
bys court type: replace percent= (_N/percent)*100

For your future posts, review FAQ Advice #10 for details on how to use the dataex command to provide data examples. Also see https://www.statalist.org/forums/help#spelling on spelling Stata.

[Nick Cox]

The percentage of a type is 100 times the mean over a (0, 1) indicator, where 1 indicates the type and 0 indicates anything else. That can be calculated on the fly, even if it does not exist as a variable. In the auto dataset, foreign is such an indicator and price > 5000 is an expression that is true or false.

Just watch out for missing values!

.

 
. sysuse auto, clear
(1978 automobile data)

. egen pcforeign = mean(100 * foreign), by(rep78)

. tabdisp rep78, c(pcforeign)

----------------------
Repair    |
record    |
1978      |  pcforeign
----------+-----------
        1 |          0
        2 |          0
        3 |         10
        4 |         50
        5 |   81.81818
        . |         20
----------------------

. tab rep78, missing su(foreign)

     Repair |        Summary of Car origin
record 1978 |        Mean   Std. dev.       Freq.
------------+------------------------------------
          1 |           0           0           2
          2 |           0           0           8
          3 |          .1   .30512858          30
          4 |          .5   .51449576          18
          5 |   .81818182   .40451992          11
          . |          .2    .4472136           5
------------+------------------------------------
      Total |    .2972973   .46018846          74

. tab rep78 foreign, row

+----------------+
| Key            |
|----------------|
|   frequency    |
| row percentage |
+----------------+

    Repair |
    record |      Car origin
      1978 |  Domestic    Foreign |     Total
-----------+----------------------+----------
         1 |         2          0 |         2 
           |    100.00       0.00 |    100.00 
-----------+----------------------+----------
         2 |         8          0 |         8 
           |    100.00       0.00 |    100.00 
-----------+----------------------+----------
         3 |        27          3 |        30 
           |     90.00      10.00 |    100.00 
-----------+----------------------+----------
         4 |         9          9 |        18 
           |     50.00      50.00 |    100.00 
-----------+----------------------+----------
         5 |         2          9 |        11 
           |     18.18      81.82 |    100.00 
-----------+----------------------+----------
     Total |        48         21 |        69 
           |     69.57      30.43 |    100.00 

. egen hi_price = mean(100 * (price > 5000)), by(rep78)

. tabdisp rep78, c(hi_price)

----------------------
Repair    |
record    |
1978      |   hi_price
----------+-----------
        1 |          0
        2 |         50
        3 |   43.33333
        4 |   66.66666
        5 |   54.54546
        . |         40
----------------------

[Andrew Musau]

Note: The FAQ Advice relating to the dataex command should be Advice #12 in post #2.