Announcement

Collapse
No announcement yet.
X
Collapse
  • Page of 1
  • Filter
  • Time
  • Show
Clear All
new posts
  • #1

    Getting means from a regression with clustered standard errors and exporting them to TEX file

    Hi,

    I need to create a table similar to that in the below picture and export it to a TEX file (for use in LaTeX).

    Click image for larger version Name: Skærmbillede 2023-03-30 183902.png Views: 1 Size: 40.0 KB ID: 1707843

    It is obviously similar to a ttest, but I cannot use ttest, as I need to cluster standard errors by the District1 variable and weigh them by "withinwt".

    I get the results using

    Code:
     reg elections Quota [aweight=withinwt], vce(cluster District1)
    Is there a way to get Stata to save the means and produce the above table in a TEX file using the regression output? Or do you have other suggestions for how to do it?

    My data looks as follows

    Code:
    * Example generated by -dataex-. For more info, type help dataex
clear
input float elections byte Quota float withinwt str14 District1
4 0 1.1162767 "D0309"
4 0 1.1162767 "D0309"
4 0 1.1162767 "D0309"
4 0 1.1162767 "D0309"
1 0 1.1162767 "D0308"
1 0 1.1162767 "D0309"
4 0 1.1162767 "D0308"
3 0 1.1162767 "D0308"
1 1  .9732189 "B0706"
4 1  .9732189 "B0706"
4 1  .9732189 "B0706"
4 1  .9732189 "B0706"
3 1  .9732189 "B0706"
3 1  .9732189 "B0706"
1 1  .9732189 "B0706"
4 1  .9732189 "B0706"
4 0  .6487038 "A0308"
2 0  .6487038 "A0308"
4 0  .6487038 "A0308"
4 0  .6487038 "A0308"
2 0  .6487038 "A0308"
4 0  .6487038 "A0308"
4 0  .6487038 "A0308"
2 0  .6487038 "A0308"
1 0  .9725117 "E1208"
1 0  .9725117 "E1208"
4 0  .9725117 "E1208"
3 0  .9725117 "E1208"
2 0  .9725117 "E1208"
3 0  .9725117 "E1209"
3 0  .9725117 "E1209"
2 0  .9725117 "E1209"
3 0  .9393414 "C1009"
4 0  .9393414 "C1009"
3 0  .9393414 "C1010"
4 0  .9393414 "C1010"
2 0  .9393414 "C1010"
4 0  .9393414 "C1010"
3 1  .6602322 "J1106"
. 1  .6602322 "J1106"
1 1  .6602322 "J1106"
4 1  .6602322 "J1106"
4 1  .6602322 "J1106"
1 1  .6602322 "J1106"
2 1  .6602322 "J1106"
3 1  .6602322 "J1106"
1 1 1.0847074 "C1505"
3 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
1 1 1.0847074 "C1505"
2 1 1.0847074 "C1505"
3 0   .895265 "G0609"
3 0   .895265 "G0609"
1 0 1.0417415 "E0206"
1 0 1.0417415 "E0206"
1 0 1.0417415 "E0206"
4 0 1.0417415 "E0206"
3 0 1.0417415 "E0206"
2 0 1.0417415 "E0206"
1 0 1.0417415 "E0206"
3 0 1.0417415 "E0206"
2 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
1 1 1.1030433 "F0406"
2 1 1.1030433 "F0406"
1 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
1 1  .6517281 "G0802"
1 1  .6517281 "G0802"
4 0   .674803 "F1307"
1 0   .674803 "F1307"
4 0   .674803 "F1307"
. 0   .674803 "F1307"
end
    Tags: None
  • #2
    Use margins to get the adjusted means. The following uses estout from SSC, with some manipulation of column names to achieve the desired format. No separate column for p-values is created as the significance stars indicate this.

    Code:
    ssc install estout, replace
    Code:
    * Example generated by -dataex-. For more info, type help dataex
clear
input float elections byte Quota float withinwt str14 District1
4 0 1.1162767 "D0309"
4 0 1.1162767 "D0309"
4 0 1.1162767 "D0309"
4 0 1.1162767 "D0309"
1 0 1.1162767 "D0308"
1 0 1.1162767 "D0309"
4 0 1.1162767 "D0308"
3 0 1.1162767 "D0308"
1 1  .9732189 "B0706"
4 1  .9732189 "B0706"
4 1  .9732189 "B0706"
4 1  .9732189 "B0706"
3 1  .9732189 "B0706"
3 1  .9732189 "B0706"
1 1  .9732189 "B0706"
4 1  .9732189 "B0706"
4 0  .6487038 "A0308"
2 0  .6487038 "A0308"
4 0  .6487038 "A0308"
4 0  .6487038 "A0308"
2 0  .6487038 "A0308"
4 0  .6487038 "A0308"
4 0  .6487038 "A0308"
2 0  .6487038 "A0308"
1 0  .9725117 "E1208"
1 0  .9725117 "E1208"
4 0  .9725117 "E1208"
3 0  .9725117 "E1208"
2 0  .9725117 "E1208"
3 0  .9725117 "E1209"
3 0  .9725117 "E1209"
2 0  .9725117 "E1209"
3 0  .9393414 "C1009"
4 0  .9393414 "C1009"
3 0  .9393414 "C1010"
4 0  .9393414 "C1010"
2 0  .9393414 "C1010"
4 0  .9393414 "C1010"
3 1  .6602322 "J1106"
. 1  .6602322 "J1106"
1 1  .6602322 "J1106"
4 1  .6602322 "J1106"
4 1  .6602322 "J1106"
1 1  .6602322 "J1106"
2 1  .6602322 "J1106"
3 1  .6602322 "J1106"
1 1 1.0847074 "C1505"
3 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
1 1 1.0847074 "C1505"
2 1 1.0847074 "C1505"
3 0   .895265 "G0609"
3 0   .895265 "G0609"
1 0 1.0417415 "E0206"
1 0 1.0417415 "E0206"
1 0 1.0417415 "E0206"
4 0 1.0417415 "E0206"
3 0 1.0417415 "E0206"
2 0 1.0417415 "E0206"
1 0 1.0417415 "E0206"
3 0 1.0417415 "E0206"
2 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
1 1 1.1030433 "F0406"
2 1 1.1030433 "F0406"
1 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
1 1  .6517281 "G0802"
1 1  .6517281 "G0802"
4 0   .674803 "F1307"
1 0   .674803 "F1307"
4 0   .674803 "F1307"
. 0   .674803 "F1307"
end

eststo diff: reg elections i.Quota [aweight=withinwt], vce(cluster District1)
eststo m1: margins 1.Quota, post
est restore diff
margins 0.Quota, post
*RENAME COEFFICIENT
mat b=e(b)
mat colname b= "1.Quota"
erepost b=b, rename
est sto m2

*OUTPUT
esttab m1 m2 diff using myfile.tex, replace keep(1.Quota) ///
mlab("Mean 1" "Mean 2" "Difference") nonumb ///
starlevels(* 0.1 ** 0.05 *** 0.01) coeflab(1.Quota "Some description") ///
noobs varwidth(20)
    Res.:

    Code:
    . esttab m1 m2 diff, replace keep(1.Quota) ///
> mlab("Mean 1" "Mean 2" "Difference") nonumb ///
> starlevels(* 0.1 ** 0.05 *** 0.01) coeflab(1.Quota "Some description") ///
> noobs varwidth(20)

--------------------------------------------------------------------
                           Mean 1          Mean 2      Difference   
--------------------------------------------------------------------
Some description            2.748***        2.776***      -0.0280   
                          (24.89)         (12.33)         (-0.11)   
--------------------------------------------------------------------
t statistics in parentheses
* p<0.1, ** p<0.05, *** p<0.01

    Comment

    • #3
      Hi Andrew,

      Thank you so much for your reply! I have been stuck at this for so long, so I really appreciate it. I see that margins do exactly what I need :-)


      I did not consider that the significance and t statistics would also be shown for the means. It is maybe a bit strange to show that the means are significantly different from 0, when the variable elections can only take values 1-4. Is there a way to only show significance and t statistics for the difference? Maybe that is "breaking the rules" a bit, so I am considering showing the p-value instead, making it more like the table in my original post. I tried to save it and then add it the table by adding a line to your code:

      Code:
      eststo diff: reg elections i.Quota [aweight=withinwt], vce(cluster District1)
scalar p1 = 2*ttail(e(df_r), abs(_b[1.Quota]/_se[1.Quota]))
di 2*ttail(e(df_r), abs(_b[1.Quota]/_se[1.Quota]))
eststo m1: margins 1.Quota, post
est restore diff
margins 0.Quota, post
*RENAME COEFFICIENT
mat b=e(b)
mat colname b= "1.Quota"
erepost b=b, rename
//
est sto m2



*OUTPUT
esttab m1 m2 diff p1 using myfile.tex, replace keep(1.Quota) ///
mlab("Mean 1" "Mean 2" "Difference") nonumb ///
starlevels(* 0.1 ** 0.05 *** 0.01) coeflab(1.Quota "Some description") ///
noobs varwidth(20)
      The display function shows the right p-value, but in the table the p1 column shows the same value as the difference, cf. below. Do you know what is wrong?

      Code:
      . esttab m1 m2 diff p1, replace keep(1.Quota) ///
> mlab("Mean 1" "Mean 2" "Difference") nonumb  ///
> starlevels(* 0.1 ** 0.05 *** 0.01) coeflab(1.Quota "Some description") ///
> noobs varwidth(20)

------------------------------------------------------------------------------------
                           Mean 1          Mean 2      Difference              p1   
------------------------------------------------------------------------------------
Some description            2.773***        2.955***       -0.182**        -0.182** 
                          (40.74)         (69.03)         (-2.26)         (-2.26)   
------------------------------------------------------------------------------------
t statistics in parentheses
* p<0.1, ** p<0.05, *** p<0.01
      Finally, I would like to add more than variable to the table, i.e. have another row for another variable (like in the table in #1), but I am uncertain how to add it, now that the first rows are entered as 3 (or 4) seperate entries. Is there a way to do that?
      Last edited by Agathe Engell; 31 Mar 2023, 04:08.

      Comment

      • #4
        This will give you what you want for one regression. Please provide an example that includes multiple regressions.

        Code:
        * Example generated by -dataex-. For more info, type help dataex
clear
input float elections byte Quota float withinwt str14 District1
4 0 1.1162767 "D0309"
4 0 1.1162767 "D0309"
4 0 1.1162767 "D0309"
4 0 1.1162767 "D0309"
1 0 1.1162767 "D0308"
1 0 1.1162767 "D0309"
4 0 1.1162767 "D0308"
3 0 1.1162767 "D0308"
1 1  .9732189 "B0706"
4 1  .9732189 "B0706"
4 1  .9732189 "B0706"
4 1  .9732189 "B0706"
3 1  .9732189 "B0706"
3 1  .9732189 "B0706"
1 1  .9732189 "B0706"
4 1  .9732189 "B0706"
4 0  .6487038 "A0308"
2 0  .6487038 "A0308"
4 0  .6487038 "A0308"
4 0  .6487038 "A0308"
2 0  .6487038 "A0308"
4 0  .6487038 "A0308"
4 0  .6487038 "A0308"
2 0  .6487038 "A0308"
1 0  .9725117 "E1208"
1 0  .9725117 "E1208"
4 0  .9725117 "E1208"
3 0  .9725117 "E1208"
2 0  .9725117 "E1208"
3 0  .9725117 "E1209"
3 0  .9725117 "E1209"
2 0  .9725117 "E1209"
3 0  .9393414 "C1009"
4 0  .9393414 "C1009"
3 0  .9393414 "C1010"
4 0  .9393414 "C1010"
2 0  .9393414 "C1010"
4 0  .9393414 "C1010"
3 1  .6602322 "J1106"
. 1  .6602322 "J1106"
1 1  .6602322 "J1106"
4 1  .6602322 "J1106"
4 1  .6602322 "J1106"
1 1  .6602322 "J1106"
2 1  .6602322 "J1106"
3 1  .6602322 "J1106"
1 1 1.0847074 "C1505"
3 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
4 1 1.0847074 "C1505"
1 1 1.0847074 "C1505"
2 1 1.0847074 "C1505"
3 0   .895265 "G0609"
3 0   .895265 "G0609"
1 0 1.0417415 "E0206"
1 0 1.0417415 "E0206"
1 0 1.0417415 "E0206"
4 0 1.0417415 "E0206"
3 0 1.0417415 "E0206"
2 0 1.0417415 "E0206"
1 0 1.0417415 "E0206"
3 0 1.0417415 "E0206"
2 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
1 1 1.1030433 "F0406"
2 1 1.1030433 "F0406"
1 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
4 1 1.1030433 "F0406"
1 1  .6517281 "G0802"
1 1  .6517281 "G0802"
4 0   .674803 "F1307"
1 0   .674803 "F1307"
4 0   .674803 "F1307"
. 0   .674803 "F1307"
end

eststo diff: reg elections i.Quota [aweight=withinwt], vce(cluster District1)
eststo m1: margins 1.Quota, post
est restore diff
margins 0.Quota, post
*RENAME COEFFICIENT
mat b=e(b)
mat colname b= "1.Quota"
erepost b=b, rename
est sto m2

*OUTPUT
esttab m1 m2 diff diff, cells("b(pattern(1 1 1 0)) p(pattern(0 0 0 1) star)") replace keep(1.Quota) ///
mlab("Mean 1" "Mean 2" "Difference" "p-value") nonumb ///
starlevels(* 0.1 ** 0.05 *** 0.01) coeflab(1.Quota "Some description") ///
noobs varwidth(20) collab(none)
        Res.:

        Code:
        . esttab m1 m2 diff diff, cells("b(pattern(1 1 1 0)) p(pattern(0 0 0 1) star)") replace keep(1.Quota) ///
> mlab("Mean 1" "Mean 2" "Difference" "p-value") nonumb ///
> starlevels(* 0.1 ** 0.05 *** 0.01) coeflab(1.Quota "Some description") ///
> noobs varwidth(20) collab(none)

---------------------------------------------------------------------------
                           Mean 1       Mean 2   Difference      p-value   
---------------------------------------------------------------------------
Some description         2.747591     2.775629    -.0280383     .9125291   
---------------------------------------------------------------------------

        Comment

        • #5
          Thank you once again!

          Apologies for not adding other variables. I have now added demo_sat and demo_fut to the data below, which also needs to be added to the table along with a couple of other variables.


          Code:
          * Example generated by -dataex-. For more info, type help dataex
clear
input float elections byte(demo_sat demo_fut Quota) float withinwt str14 District1
4 1  4 0 1.1162767 "D0309"
4 1 10 0 1.1162767 "D0309"
4 2 10 0 1.1162767 "D0309"
4 3 10 0 1.1162767 "D0309"
1 .  . 0 1.1162767 "D0309"
1 4  9 0 1.1162767 "D0310"
4 4 10 0 1.1162767 "D0310"
3 3  4 0 1.1162767 "D0310"
1 4  6 1  .9732189 "B0706"
4 1  . 1  .9732189 "B0706"
4 1  8 1  .9732189 "B0706"
4 4 10 1  .9732189 "B0706"
3 2  0 1  .9732189 "B0706"
3 .  . 1  .9732189 "B0706"
1 3  . 1  .9732189 "B0707"
4 4 10 1  .9732189 "B0707"
4 4 10 0  .6487038 "A0308"
2 3  . 0  .6487038 "A0308"
4 4 10 0  .6487038 "A0308"
4 1  . 0  .6487038 "A0308"
2 2  5 0  .6487038 "A0308"
4 4 10 0  .6487038 "A0308"
4 1 10 0  .6487038 "A0308"
2 3  7 0  .6487038 "A0308"
1 3  8 0  .9725117 "E1208"
1 1 10 0  .9725117 "E1208"
4 2 10 0  .9725117 "E1208"
3 3  . 0  .9725117 "E1208"
2 2  0 0  .9725117 "E1208"
3 2  8 0  .9725117 "E1209"
3 3 10 0  .9725117 "E1209"
2 3 10 0  .9725117 "E1209"
3 2 10 0  .9393414 "C1009"
4 2  . 0  .9393414 "C1009"
3 4  . 0  .9393414 "C1009"
4 3 10 0  .9393414 "C1009"
2 1  0 0  .9393414 "C1009"
4 .  . 0  .9393414 "C1009"
3 3 10 1  .6602322 "J1106"
. 4  . 1  .6602322 "J1106"
1 1  . 1  .6602322 "J1106"
4 3 10 1  .6602322 "J1106"
4 3 10 1  .6602322 "J1106"
1 0  . 1  .6602322 "J1106"
2 3 10 1  .6602322 "J1106"
3 1 10 1  .6602322 "J1106"
1 2  5 1 1.0847074 "C1505"
3 4  8 1 1.0847074 "C1505"
4 4 10 1 1.0847074 "C1505"
4 2 10 1 1.0847074 "C1505"
4 1 10 1 1.0847074 "C1505"
4 1  . 1 1.0847074 "C1505"
1 4 10 1 1.0847074 "C1505"
2 3  7 1 1.0847074 "C1505"
3 4  . 0   .895265 "G0609"
3 .  . 0   .895265 "G0609"
1 3 10 0 1.0417415 "E0206"
1 3  . 0 1.0417415 "E0206"
1 1 10 0 1.0417415 "E0206"
4 3 10 0 1.0417415 "E0206"
3 2  5 0 1.0417415 "E0206"
2 2  0 0 1.0417415 "E0206"
1 .  . 0 1.0417415 "E0206"
3 3  6 0 1.0417415 "E0206"
2 2  9 1 1.1030433 "F0406"
4 1  . 1 1.1030433 "F0406"
1 4  . 1 1.1030433 "F0406"
2 2  . 1 1.1030433 "F0406"
1 3  5 1 1.1030433 "F0406"
4 1 10 1 1.1030433 "F0406"
4 4 10 1 1.1030433 "F0406"
4 4 10 1 1.1030433 "F0406"
1 2  0 1  .6517281 "G0802"
1 3 10 1  .6517281 "G0802"
2 .  . 3   .674803 "F1307"
1 2 10 0   .674803 "F1307"
4 1 10 0   .674803 "F1308"
. 3 10 0   .674803 "F1308"
3 1  . 1 1.2975078 "A1203"
3 4 10 1 1.2975078 "A1203"
4 3  . 1 1.2975078 "A1203"
2 . 10 1 1.2975078 "A1203"
4 4 10 1  .9651449 "H0106"
. 2 10 1  .9651449 "H0106"
3 2 10 1  .9651449 "H0106"
4 3 10 1  .9651449 "H0106"
1 1 10 0   .674803 "F1307"
4 1  . 0   .674803 "F1307"
2 4 10 0   .674803 "F1307"
4 3  . 0   .674803 "F1307"
4 3 10 1 1.0614818 "B0408"
4 2  1 1 1.0614818 "B0408"
4 2 10 1 1.0614818 "B0408"
4 . 10 1 1.0614818 "B0408"
4 1  5 1 1.0614818 "B0409"
4 1  5 1 1.0614818 "B0409"
4 3 10 1 1.0614818 "B0408"
4 4 10 1 1.0614818 "B0408"
3 2 10 0  .7088118 "F0101"
1 .  . 0  .7088118 "F0101"
end
          Last edited by Agathe Engell; 01 Apr 2023, 05:34.

          Comment

          • #6
            Code:
            * Example generated by -dataex-. For more info, type help dataex
clear
input float elections byte(demo_sat demo_fut Quota) float withinwt str14 District1
4 1  4 0 1.1162767 "D0309"
4 1 10 0 1.1162767 "D0309"
4 2 10 0 1.1162767 "D0309"
4 3 10 0 1.1162767 "D0309"
1 .  . 0 1.1162767 "D0309"
1 4  9 0 1.1162767 "D0310"
4 4 10 0 1.1162767 "D0310"
3 3  4 0 1.1162767 "D0310"
1 4  6 1  .9732189 "B0706"
4 1  . 1  .9732189 "B0706"
4 1  8 1  .9732189 "B0706"
4 4 10 1  .9732189 "B0706"
3 2  0 1  .9732189 "B0706"
3 .  . 1  .9732189 "B0706"
1 3  . 1  .9732189 "B0707"
4 4 10 1  .9732189 "B0707"
4 4 10 0  .6487038 "A0308"
2 3  . 0  .6487038 "A0308"
4 4 10 0  .6487038 "A0308"
4 1  . 0  .6487038 "A0308"
2 2  5 0  .6487038 "A0308"
4 4 10 0  .6487038 "A0308"
4 1 10 0  .6487038 "A0308"
2 3  7 0  .6487038 "A0308"
1 3  8 0  .9725117 "E1208"
1 1 10 0  .9725117 "E1208"
4 2 10 0  .9725117 "E1208"
3 3  . 0  .9725117 "E1208"
2 2  0 0  .9725117 "E1208"
3 2  8 0  .9725117 "E1209"
3 3 10 0  .9725117 "E1209"
2 3 10 0  .9725117 "E1209"
3 2 10 0  .9393414 "C1009"
4 2  . 0  .9393414 "C1009"
3 4  . 0  .9393414 "C1009"
4 3 10 0  .9393414 "C1009"
2 1  0 0  .9393414 "C1009"
4 .  . 0  .9393414 "C1009"
3 3 10 1  .6602322 "J1106"
. 4  . 1  .6602322 "J1106"
1 1  . 1  .6602322 "J1106"
4 3 10 1  .6602322 "J1106"
4 3 10 1  .6602322 "J1106"
1 0  . 1  .6602322 "J1106"
2 3 10 1  .6602322 "J1106"
3 1 10 1  .6602322 "J1106"
1 2  5 1 1.0847074 "C1505"
3 4  8 1 1.0847074 "C1505"
4 4 10 1 1.0847074 "C1505"
4 2 10 1 1.0847074 "C1505"
4 1 10 1 1.0847074 "C1505"
4 1  . 1 1.0847074 "C1505"
1 4 10 1 1.0847074 "C1505"
2 3  7 1 1.0847074 "C1505"
3 4  . 0   .895265 "G0609"
3 .  . 0   .895265 "G0609"
1 3 10 0 1.0417415 "E0206"
1 3  . 0 1.0417415 "E0206"
1 1 10 0 1.0417415 "E0206"
4 3 10 0 1.0417415 "E0206"
3 2  5 0 1.0417415 "E0206"
2 2  0 0 1.0417415 "E0206"
1 .  . 0 1.0417415 "E0206"
3 3  6 0 1.0417415 "E0206"
2 2  9 1 1.1030433 "F0406"
4 1  . 1 1.1030433 "F0406"
1 4  . 1 1.1030433 "F0406"
2 2  . 1 1.1030433 "F0406"
1 3  5 1 1.1030433 "F0406"
4 1 10 1 1.1030433 "F0406"
4 4 10 1 1.1030433 "F0406"
4 4 10 1 1.1030433 "F0406"
1 2  0 1  .6517281 "G0802"
1 3 10 1  .6517281 "G0802"
2 .  . 3   .674803 "F1307"
1 2 10 0   .674803 "F1307"
4 1 10 0   .674803 "F1308"
. 3 10 0   .674803 "F1308"
3 1  . 1 1.2975078 "A1203"
3 4 10 1 1.2975078 "A1203"
4 3  . 1 1.2975078 "A1203"
2 . 10 1 1.2975078 "A1203"
4 4 10 1  .9651449 "H0106"
. 2 10 1  .9651449 "H0106"
3 2 10 1  .9651449 "H0106"
4 3 10 1  .9651449 "H0106"
1 1 10 0   .674803 "F1307"
4 1  . 0   .674803 "F1307"
2 4 10 0   .674803 "F1307"
4 3  . 0   .674803 "F1307"
4 3 10 1 1.0614818 "B0408"
4 2  1 1 1.0614818 "B0408"
4 2 10 1 1.0614818 "B0408"
4 . 10 1 1.0614818 "B0408"
4 1  5 1 1.0614818 "B0409"
4 1  5 1 1.0614818 "B0409"
4 3 10 1 1.0614818 "B0408"
4 4 10 1 1.0614818 "B0408"
3 2 10 0  .7088118 "F0101"
1 .  . 0  .7088118 "F0101"
end
replace Quota=1 if Quota==3

local depvars elections demo_sat demo_fut
local i 0
foreach depvar of local depvars{
    local ++i
    rename (`depvar' Quota) (Quota `depvar')
    eststo diff_`i': reg Quota i.`depvar' [aweight=withinwt], vce(cluster District1)
    eststo m1_`i': margins 1.`depvar', post
    est restore diff_`i'
    margins 0.`depvar', post
    *RENAME COEFFICIENT
    mat b=e(b)
    mat colname b= "1.`depvar'"
    erepost b=b, rename
    est sto m2_`i'
    rename (`depvar' Quota) (Quota `depvar')
}

*PROGRAM TO APPEND MODELS
capt prog drop appendmodels
*! version 1.0.0  14aug2007  Ben Jann
program appendmodels, eclass
    // using first equation of model
    version 8
    syntax namelist
    tempname b V tmp
    foreach name of local namelist {
        qui est restore `name'
        mat `tmp' = e(b)
        local eq1: coleq `tmp'
        gettoken eq1 : eq1
        mat `tmp' = `tmp'[1,"`eq1':"]
        local cons = colnumb(`tmp',"_cons")
        if `cons'<. & `cons'>1 {
            mat `tmp' = `tmp'[1,1..`cons'-1]
        }
        mat `b' = nullmat(`b') , `tmp'
        mat `tmp' = e(V)
        mat `tmp' = `tmp'["`eq1':","`eq1':"]
        if `cons'<. & `cons'>1 {
            mat `tmp' = `tmp'[1..`cons'-1,1..`cons'-1]
        }
        capt confirm matrix `V'
        if _rc {
            mat `V' = `tmp'
        }
        else {
            mat `V' = ///
            ( `V' , J(rowsof(`V'),colsof(`tmp'),0) ) \ ///
            ( J(rowsof(`tmp'),colsof(`V'),0) , `tmp' )
        }
    }
    local names: colfullnames `b'
    mat coln `V' = `names'
    mat rown `V' = `names'
    eret post `b' `V'
    eret local cmd "whatever"
end

*APPEND THE MODELS
local diff 
local m1 
local m2
forval i=1/`=wordcount("`depvars'")'{
   local diff "`diff' diff_`i'" 
   local m1 "`m1' m1_`i'"
   local m2 "`m2' m2_`i'"  
}
eststo diff: appendmodels `diff'
eststo m1: appendmodels `m1'
eststo m2: appendmodels `m2'

esttab m1 m2 diff diff, cells("b(pattern(1 1 1 0)) p(pattern(0 0 0 1) star)") replace keep(1.*) ///
mlab("Mean 1" "Mean 2" "Difference" "p-value", lhs(Variables)) nonumb ///
starlevels(* 0.1 ** 0.05 *** 0.01) substitute(1.e "e" 1.d "d") ///
noobs varwidth(20) collab(none)
            Res.:

            Code:
            . esttab m1 m2 diff diff, cells("b(pattern(1 1 1 0)) p(pattern(0 0 0 1) star)") replace keep(1.*) ///
> mlab("Mean 1" "Mean 2" "Difference" "p-value", lhs(Variables)) nonumb ///
> starlevels(* 0.1 ** 0.05 *** 0.01) substitute(1.e "e" 1.d "d") ///
> noobs varwidth(20) collab(none)

---------------------------------------------------------------------------
Variables                  Mean 1       Mean 2   Difference      p-value   
---------------------------------------------------------------------------
elections              3.038563     2.729372     .3091909     .2594785   
demo_sat               2.550968     2.479386     .0715826     .6979601   
demo_fut               8.408523     8.026123     .3824004     .5102858   
---------------------------------------------------------------------------

            Comment

            • #7
              Thank you once again!

              One finale question: When I look at outcomes of the regressions as they are run in the first loop, I get slightly different p-values than in the final table. It's down at the third decimal point, so it is very small differences, but is there an explanation for this? I get .0238954** in the table for elections, but 0.026 in the regression, and 0.1739278 for demo_fut in the table, but 0.176 in the regression.

              Comment

              • #8
                There is some issue with the calculations once we combine the results. I don't have time to figure it out, but you can save the p-values directly into matrices to output in the final step.

                Code:
                clear
input float elections byte(demo_sat demo_fut Quota) float withinwt str14 District1
4 1  4 0 1.1162767 "D0309"
4 1 10 0 1.1162767 "D0309"
4 2 10 0 1.1162767 "D0309"
4 3 10 0 1.1162767 "D0309"
1 .  . 0 1.1162767 "D0309"
1 4  9 0 1.1162767 "D0310"
4 4 10 0 1.1162767 "D0310"
3 3  4 0 1.1162767 "D0310"
1 4  6 1  .9732189 "B0706"
4 1  . 1  .9732189 "B0706"
4 1  8 1  .9732189 "B0706"
4 4 10 1  .9732189 "B0706"
3 2  0 1  .9732189 "B0706"
3 .  . 1  .9732189 "B0706"
1 3  . 1  .9732189 "B0707"
4 4 10 1  .9732189 "B0707"
4 4 10 0  .6487038 "A0308"
2 3  . 0  .6487038 "A0308"
4 4 10 0  .6487038 "A0308"
4 1  . 0  .6487038 "A0308"
2 2  5 0  .6487038 "A0308"
4 4 10 0  .6487038 "A0308"
4 1 10 0  .6487038 "A0308"
2 3  7 0  .6487038 "A0308"
1 3  8 0  .9725117 "E1208"
1 1 10 0  .9725117 "E1208"
4 2 10 0  .9725117 "E1208"
3 3  . 0  .9725117 "E1208"
2 2  0 0  .9725117 "E1208"
3 2  8 0  .9725117 "E1209"
3 3 10 0  .9725117 "E1209"
2 3 10 0  .9725117 "E1209"
3 2 10 0  .9393414 "C1009"
4 2  . 0  .9393414 "C1009"
3 4  . 0  .9393414 "C1009"
4 3 10 0  .9393414 "C1009"
2 1  0 0  .9393414 "C1009"
4 .  . 0  .9393414 "C1009"
3 3 10 1  .6602322 "J1106"
. 4  . 1  .6602322 "J1106"
1 1  . 1  .6602322 "J1106"
4 3 10 1  .6602322 "J1106"
4 3 10 1  .6602322 "J1106"
1 0  . 1  .6602322 "J1106"
2 3 10 1  .6602322 "J1106"
3 1 10 1  .6602322 "J1106"
1 2  5 1 1.0847074 "C1505"
3 4  8 1 1.0847074 "C1505"
4 4 10 1 1.0847074 "C1505"
4 2 10 1 1.0847074 "C1505"
4 1 10 1 1.0847074 "C1505"
4 1  . 1 1.0847074 "C1505"
1 4 10 1 1.0847074 "C1505"
2 3  7 1 1.0847074 "C1505"
3 4  . 0   .895265 "G0609"
3 .  . 0   .895265 "G0609"
1 3 10 0 1.0417415 "E0206"
1 3  . 0 1.0417415 "E0206"
1 1 10 0 1.0417415 "E0206"
4 3 10 0 1.0417415 "E0206"
3 2  5 0 1.0417415 "E0206"
2 2  0 0 1.0417415 "E0206"
1 .  . 0 1.0417415 "E0206"
3 3  6 0 1.0417415 "E0206"
2 2  9 1 1.1030433 "F0406"
4 1  . 1 1.1030433 "F0406"
1 4  . 1 1.1030433 "F0406"
2 2  . 1 1.1030433 "F0406"
1 3  5 1 1.1030433 "F0406"
4 1 10 1 1.1030433 "F0406"
4 4 10 1 1.1030433 "F0406"
4 4 10 1 1.1030433 "F0406"
1 2  0 1  .6517281 "G0802"
1 3 10 1  .6517281 "G0802"
2 .  . 3   .674803 "F1307"
1 2 10 0   .674803 "F1307"
4 1 10 0   .674803 "F1308"
. 3 10 0   .674803 "F1308"
3 1  . 1 1.2975078 "A1203"
3 4 10 1 1.2975078 "A1203"
4 3  . 1 1.2975078 "A1203"
2 . 10 1 1.2975078 "A1203"
4 4 10 1  .9651449 "H0106"
. 2 10 1  .9651449 "H0106"
3 2 10 1  .9651449 "H0106"
4 3 10 1  .9651449 "H0106"
1 1 10 0   .674803 "F1307"
4 1  . 0   .674803 "F1307"
2 4 10 0   .674803 "F1307"
4 3  . 0   .674803 "F1307"
4 3 10 1 1.0614818 "B0408"
4 2  1 1 1.0614818 "B0408"
4 2 10 1 1.0614818 "B0408"
4 . 10 1 1.0614818 "B0408"
4 1  5 1 1.0614818 "B0409"
4 1  5 1 1.0614818 "B0409"
4 3 10 1 1.0614818 "B0408"
4 4 10 1 1.0614818 "B0408"
3 2 10 0  .7088118 "F0101"
1 .  . 0  .7088118 "F0101"
end
replace Quota=1 if Quota==3

local depvars elections demo_sat demo_fut
local i 0
foreach depvar of local depvars{
    local ++i
    rename (`depvar' Quota) (Quota `depvar')
    eststo diff_`i': reg Quota i.`depvar' [aweight=withinwt], vce(cluster District1)
    mat pval_`i'= r(table)["pvalue", "0b.`depvar'".."1.`depvar'"]
    eststo m1_`i': margins 1.`depvar', post
    est restore diff_`i'
    margins 0.`depvar', post
    *RENAME COEFFICIENT
    mat b=e(b)
    mat colname b= "1.`depvar'"
    erepost b=b, rename
    est sto m2_`i'
    rename (`depvar' Quota) (Quota `depvar')
}

*PROGRAM TO APPEND MODELS
capt prog drop appendmodels
*! version 1.0.0  14aug2007  Ben Jann
program appendmodels, eclass
    // using first equation of model
    version 8
    syntax namelist
    tempname b V tmp
    foreach name of local namelist {
        qui est restore `name'
        mat `tmp' = e(b)
        local eq1: coleq `tmp'
        gettoken eq1 : eq1
        mat `tmp' = `tmp'[1,"`eq1':"]
        local cons = colnumb(`tmp',"_cons")
        if `cons'<. & `cons'>1 {
            mat `tmp' = `tmp'[1,1..`cons'-1]
        }
        mat `b' = nullmat(`b') , `tmp'
        mat `tmp' = e(V)
        mat `tmp' = `tmp'["`eq1':","`eq1':"]
        if `cons'<. & `cons'>1 {
            mat `tmp' = `tmp'[1..`cons'-1,1..`cons'-1]
        }
        capt confirm matrix `V'
        if _rc {
            mat `V' = `tmp'
        }
        else {
            mat `V' = ///
            ( `V' , J(rowsof(`V'),colsof(`tmp'),0) ) \ ///
            ( J(rowsof(`tmp'),colsof(`V'),0) , `tmp' )
        }
    }
    local names: colfullnames `b'
    mat coln `V' = `names'
    mat rown `V' = `names'
    eret post `b' `V'
    eret local cmd "whatever"
end

*APPEND THE MODELS
local diff 
local m1 
local m2
forval i=1/`=wordcount("`depvars'")'{
   local diff "`diff' diff_`i'" 
   local m1 "`m1' m1_`i'"
   local m2 "`m2' m2_`i'" 
   local pval "`pval' pval_`i',"
}
eststo diff: appendmodels `diff'
eststo m1: appendmodels `m1'
eststo m2: appendmodels `m2'
local pval= substr("`pval'", 1,length("`pval'")-1)
mat pval= `pval'
est restore diff
estadd matrix pval= pval

esttab m1 m2 diff diff, cells("b(pattern(1 1 1 0)) pval(pattern(0 0 0 1) star)") replace keep(1.*) ///
mlab("Mean 1" "Mean 2" "Difference" "p-value", lhs(Variables)) nonumb ///
starlevels(* 0.1 ** 0.05 *** 0.01) substitute(1.e "e" 1.d "d") ///
noobs varwidth(20) collab(none)
                Res.:

                Code:
                . esttab m1 m2 diff diff, cells("b(pattern(1 1 1 0)) pval(pattern(0 0 0 1) star)") replace keep(1.*) ///
> mlab("Mean 1" "Mean 2" "Difference" "p-value", lhs(Variables)) nonumb ///
> starlevels(* 0.1 ** 0.05 *** 0.01) substitute(1.e "e" 1.d "d") ///
> noobs varwidth(20) collab(none)

---------------------------------------------------------------------------
Variables                  Mean 1       Mean 2   Difference      p-value   
---------------------------------------------------------------------------
elections              3.038563     2.729372     .3091909     .2728253   
demo_sat               2.550968     2.479386     .0715826     .7020599   
demo_fut               8.408523     8.026123     .3824004      .518185   
---------------------------------------------------------------------------

                Comment

                • #9
                  Hi Andrew, thank you once again for taking the time to reply and clarifying how to get the right values into the table.

                  Comment

                  Previous Next
                  Working...
                  X