Setting up matrix

[Grace Heymsfield]

Hello! I have the following code, which I'd like to put out into Excel instead of laboriously typing- any advice? All binary variables

*Overlap between WHZ and MUAC
*<125 mm & WHZ <-2
svy: tabulate GAM GAMMUAC, ci obs deff
*<135 mm & WHZ <-2
svy: tabulate GAM MUAC135, ci obs deff
*<130 mm & WHZ <-2
svy: tab GAM MUAC130, ci obs deff
*<120 mm & WHZ<-2
svy: tab GAM MUAC120, ci obs deff
*<115 mm & WHZ<-2

svy: tab GAM MUAC115, ci obs deff

[Grace Heymsfield]

Realizing now that I could use the source function!

*Overlap between WHZ and MUAC
syvset CLUSTER

*<125 mm & WHZ <-2
svy: tabulate GAM GAMMUAC, ci obs deff
*<135 mm & WHZ <-2
svy: tabulate GAM MUAC135, ci obs deff
*<130 mm & WHZ <-2
svy: tab GAM MUAC130, ci obs deff
*<120 mm & WHZ<-2
svy: tab GAM MUAC120, ci obs deff


Advice on how to put this into a matrix, and then to Excel? I need to retain the confidence interval.

[Andrew Musau]

The means and CIs are easy to export via esttab (SSC), but you will need some work to exactly replicate the table. Here is a reproducible example including 3 variables.

ssc install estout, replace
webuse nhanes2b, clear
eststo clear
foreach var in diabetes heartatk fhtatk{
    estpost svy: tabulate race `var', ci deff obs
    eststo
}
esttab est* using myfile.xls, ci unstack collabels("") nonumber noobs replace

Res.: (Note: Output for 2 variables shown to enhance readability)

. esttab est1 est2, ci unstack collabels("") nonumber noobs compress

----------------------------------------------------------------------------------------------------------------------------------
                       Mean                                                        Mean                                          
                          0                   1               Total                   0                   1               Total  
                                                                                                                                  
----------------------------------------------------------------------------------------------------------------------------------
White                 0.851***           0.0281***            0.879***            0.852***           0.0270***            0.879***
              [0.819,0.883]     [0.0241,0.0320]       [0.845,0.913]       [0.819,0.885]     [0.0236,0.0305]       [0.845,0.913]  

Black                0.0899***          0.00564***           0.0955***           0.0932***          0.00237***           0.0955***
             [0.0651,0.115]    [0.00392,0.00736]      [0.0695,0.122]      [0.0678,0.118]    [0.00124,0.00349]      [0.0695,0.122]  

Other                0.0248*           0.000515              0.0253*             0.0250*           0.000353              0.0253*  
           [0.00408,0.0456]    [-0.000274,0.00130]    [0.00384,0.0469]    [0.00414,0.0458]    [-0.000305,0.00101]    [0.00384,0.0469]  

Total                 0.966***           0.0342***                1               0.970***           0.0297***                1  
              [0.962,0.969]     [0.0305,0.0379]               [.,.]       [0.966,0.974]     [0.0260,0.0335]               [.,.]  
----------------------------------------------------------------------------------------------------------------------------------
95% confidence intervals in brackets
* p<0.05, ** p<0.01, *** p<0.001

[Grace Heymsfield]

Thank you so much. What would be the next step for exporting this to Excel?

[Andrew Musau]

The command already does this.

esttab est* using myfile.xls, ci unstack collabels("") nonumber noobs replace

or you could use the file extension

esttab est* using myfile.csv, ci unstack collabels("") nonumber noobs replace

and open the csv file in Excel. You get a clickable link to the file, or type

cd

and within this folder lies your file.

[Grace Heymsfield]

Thanks!!!