Incorporating Parameters from Estimation Outputs into a Mathematical Formula automatically

[Simwinga Simwinga]

Hello,
Could you assist me in creating a code that initially computes the means and standard deviations of SBA marks and Exam marks by School_Code, performs a regression of Raw_SBA marks on Exam Scores, and subsequently utilizes the means, standard deviations, and the regression coefficient in a mathematical formula to determine the moderated mark?
The formula is "Moderated SBA Score = Overal_Mean_Raw_SBA + RegCoef(Sch_mean_Exam - Overal_Mean_Exam) + (Raw_SBA_Student - Sch_mean_SBA)(Sch_SD_Exam/Sch_SD_SBA)"

* Example generated by -dataex-. For more info, type help dataex
clear
input float Indentification long School_Code int(Subject_Code Raw_SBA) float Exam
  1 10002 208 37 75
  2 10002 208 38 52
  3 10002 208 38 52
  4 10002 208 40 72
  5 10002 208 37 77
  6 10005 208 32 56
  7 10005 208 34 66
  8 10005 208 34 68
  9 10005 208 34 66
 10 10005 208 34 62
 11 10013 208 36 58
 12 10013 208 38 59
 13 10013 208 36 51
 14 10013 208 36 51
 15 10013 208 38 53
 16 10018 208 39 68
 17 10018 208 38 84
 18 10018 208 34 63
 19 10018 208 35 38
 20 10018 208 39 72
 21 10032 208 34 67
 22 10032 208 38 69
 23 10032 208 37 68
 24 10032 208 34 61
 25 10032 208 36 75
 26 10175 208 36 60
 27 10175 208  0  5
 28 10175 208 36 52
 29 10175 208 34 58
 30 10175 208 36 60
 31 10496 208 37 57
 32 10496 208 37 71
 33 10496 208 37 57
 34 10496 208 35 57
 35 10496 208 36 63
 36 10558 208  0 67
 37 10558 208  0 61
 38 10558 208  0 55
 39 10558 208  0 59
 40 10558 208  0 53
 41 10837 208 32 49
 42 10837 208 13 36
 43 10837 208 36 55
 44 10837 208 33 59
 45 10837 208 36 56
 46 10838 208 32 34
 47 10838 208 32 37
 48 10838 208 35 51
 49 10838 208 36 38
 50 10838 208 36 48
 51 20005 208 31 51
 52 20005 208 30 44
 53 20005 208 29 53
 54 20005 208 31 51
 55 20005 208 29 53
 56 20010 208 39 76
 57 20010 208 36 68
 58 20010 208 34 74
 59 20010 208 37 80
 60 20010 208 33 71
 61 20023 208 36 55
 62 20023 208 36 55
 63 20023 208 36 55
 64 20023 208 36 55
 65 20023 208 36 55
 66 20080 208 36 76
 67 20080 208 39 55
 68 20080 208 40 71
 69 20080 208 36 60
 70 20080 208 36 59
 71 20085 208 37 59
 72 20085 208 37 60
 73 20085 208 36 49
 74 20085 208 38 43
 75 20085 208 36 58
 76 20130 208 37 74
 77 20130 208 35 57
 78 20130 208 32 58
 79 20130 208 37 59
 80 20130 208 34 66
 81 20153 208 33 46
 82 20153 208 24 42
 83 20153 208 33 58
 84 20153 208 24 44
 85 20153 208 33 60
 86 20187 208 33 51
 87 20187 208 34 43
 88 20187 208 30 40
 89 20187 208 34 54
 90 20187 208 33 51
 91 20195 208 30 56
 92 20195 208 29 54
 93 20195 208 34 76
 94 20195 208 34 76
 95 20195 208 29 44
 96 20235 208 38 69
 97 20235 208 36 60
 98 20235 208 38 71
 99 20235 208 35 57
100 20235 208 36 59
end

[Simwinga Simwinga]

The means and SDs are to be calculated by School_Code and Overall (National).

[Andrew Musau]

If you need to use more than 1 mean or SD, you use a local macro to hold these:

sysuse auto, clear
foreach var of varlist mpg turn disp{
    qui summarize `var'
    local mean_`var'= r(mean)
    local SD_`var'= r(sd)
}
di `mean_mpg'
di `SD_turn'
di `mean_displacement'

Res.:

. di `mean_mpg'
21.297297

.
. di `SD_turn'
4.3993537

.
. di `mean_displacement'
197.2973

To see how to reference a coefficient, run the regression with the option -coeflegend-, e.g.,

sysuse auto, clear
regress price mpg weight turn i.rep78, coeflegend

Res.:

. regress price mpg weight turn i.rep78, coeflegend

      Source |       SS           df       MS      Number of obs   =        69
-------------+----------------------------------   F(7, 61)        =      6.34
       Model |   242819042         7  34688434.6   Prob > F        =    0.0000
    Residual |   333977917        61  5475047.82   R-squared       =    0.4210
-------------+----------------------------------   Adj R-squared   =    0.3545
       Total |   576796959        68  8482308.22   Root MSE        =    2339.9

------------------------------------------------------------------------------
       price | Coefficient  Legend
-------------+----------------------------------------------------------------
         mpg |  -86.24749  _b[mpg]
      weight |    3.39851  _b[weight]
        turn |  -321.7209  _b[turn]
             |
       rep78 |
          2  |   1143.126  _b[2.rep78]
          3  |   1074.757  _b[3.rep78]
          4  |   1541.853  _b[4.rep78]
          5  |   2813.323  _b[5.rep78]
             |
       _cons |   9030.873  _b[_cons]

So after the regression, when the estimation results are active, you can specify your expression, e.g.,

display "wanted = " `=(`mean_mpg'*_b[mpg])+ (`mean_weight'*_b[weight])/ `SD_turn''

But note that if you have missing values, not all observations are used in the regression, and you will need to run the regression first and compute the statistics using the estimation sample, e.g.,

sysuse auto, clear
regress price mpg weight turn i.rep78, coeflegend
mean mpg if e(sample)