problem with reg for fweight

[caroline phn]

Hi,
I want to input a number of people by age classes, but then my code give me one observation for each age classe. I try with the reg function, who say I have 66 millions of observations, but when after I try "tab AGECL" that say again that I have one observation for each classe group. Do somebody know how can I solve this problem ?
Thanks !

clear
input byte (AGECL) POPULATION
1 11965004
2 11740895
3 12486039
4 13345975
5 11133961
6 6286470
end
label values AGECL AGECL
label def AGECL 1 "0-14", modify
label def AGECL 2 "15-29", modify
label def AGECL 3 "30-44", modify
label def AGECL 4 "45-59", modify
label def AGECL 5 "60-74", modify
label def AGECL 6 "plus de 75", modify
label var AGECL "âge en classe"
label var POPULATION "Population par groupe d'âge en France"

reg AGECL [fweight=POPULATION]


      Source |       SS       df       MS              Number of obs =66958344
-------------+------------------------------           F(  0,66958343) =    0.00
       Model |           0     0           .           Prob > F      =       .
    Residual |   16878110466958343  2.52068819           R-squared     =  0.0000
-------------+------------------------------           Adj R-squared =  0.0000
       Total |   16878110466958343  2.52068819           Root MSE      =  1.5877

------------------------------------------------------------------------------
       AGECL |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _cons |   3.280807    .000194  1.7e+04   0.000     3.280427    3.281188
------------------------------------------------------------------------------



. tab AGECL

     âge en |
     classe |      Freq.     Percent        Cum.
------------+-----------------------------------
       0-14 |          1       16.67       16.67
      15-29 |          1       16.67       33.33
      30-44 |          1       16.67       50.00
      45-59 |          1       16.67       66.67
      60-74 |          1       16.67       83.33
 plus de 75 |          1       16.67      100.00
------------+-----------------------------------
      Total |          6      100.00

[Rich Goldstein]

I'm not sure I completely follow; however, you used "[fweight=POPULATION]" but not in your -tab- command - if you add it to your -tab- command, you get a total of "66,958,344"

[caroline phn]

Thanks for your reply, I try it, but I have the same problem

tab POPULATION

 Population |
 par groupe |
   d'âge en |
     France |      Freq.     Percent        Cum.
------------+-----------------------------------
    6286470 |          1       16.67       16.67
   1.11e+07 |          1       16.67       33.33
   1.17e+07 |          1       16.67       50.00
   1.20e+07 |          1       16.67       66.67
   1.25e+07 |          1       16.67       83.33
   1.33e+07 |          1       16.67      100.00
------------+-----------------------------------
      Total |          6      100.00

[Andrew Musau]

tab AGECL [fw=POPULATION]

[caroline phn]

Thanks ! Do you know if it's possible to save it without having to make [fw= POPULATION] at each command ? Because I would like to use this population as a reference population for a standardisation, so i would like to save the data (the data that I've got when I make [fw=POPULATION]) in a .dta file

[Andrew Musau]

The only thing that tabulate is giving you are the percentages and cumulative percentages, which you can calculate. Otherwise, the frequencies are already present in the dataset.

clear
input byte (AGECL) POPULATION
1 11965004
2 11740895
3 12486039
4 13345975
5 11133961
6 6286470
end
label values AGECL AGECL
label def AGECL 1 "0-14", modify
label def AGECL 2 "15-29", modify
label def AGECL 3 "30-44", modify
label def AGECL 4 "45-59", modify
label def AGECL 5 "60-74", modify
label def AGECL 6 "plus de 75", modify
label var AGECL "âge en classe"
label var POPULATION "Population par groupe d'âge en France"

tab AGECL [fw=POPULATION]

format POPULATION %14.0f
egen double pct= total(POPULATION)
replace pct= (POPULATION/pct)*100
gen cum_pct=sum(pct)

Res.:

. tab AGECL [fw=POPULATION]

     âge en |
     classe |      Freq.     Percent        Cum.
------------+-----------------------------------
       0-14 | 11,965,004       17.87       17.87
      15-29 | 11,740,895       17.53       35.40
      30-44 | 12,486,039       18.65       54.05
      45-59 | 13,345,975       19.93       73.98
      60-74 | 11,133,961       16.63       90.61
 plus de 75 |  6,286,470        9.39      100.00
------------+-----------------------------------
      Total | 66,958,344      100.00


.
. format POPULATION %14.0f

.
. egen double pct= total(POPULATION)

.
. replace pct= (POPULATION/pct)*100
(6 real changes made)

.
. gen cum_pct=sum(pct)

. l, sep(0)

     +----------------------------------------------+
     |      AGECL   POPULA~N         pct    cum_pct |
     |----------------------------------------------|
  1. |       0-14   11965004   17.869325   17.86933 |
  2. |      15-29   11740895   17.534626   35.40395 |
  3. |      30-44   12486039   18.647473   54.05142 |
  4. |      45-59   13345975   19.931758   73.98318 |
  5. |      60-74   11133961   16.628191   90.61137 |
  6. | plus de 75    6286470   9.3886282        100 |
     +----------------------------------------------+

.

[caroline phn]

Thanks you very murch this is exactly what I wanted to do !