add "positive sign" before estimate value | collect style

[Luis Pecht]

Hi ,

I was wondering if is it possible to add a "positive sign" before estimate value, same way, it occurs automatically with negative values.

webuse auto
(1978 automobile data)

. reg price mpg headroom weight length turn displacement

      Source |       SS           df       MS      Number of obs   =        74
-------------+----------------------------------   F(6, 67)        =      8.52
       Model |   274827242         6  45804540.3   Prob > F        =    0.0000
    Residual |   360238154        67  5376688.87   R-squared       =    0.4328
-------------+----------------------------------   Adj R-squared   =    0.3820
       Total |   635065396        73  8699525.97   Root MSE        =    2318.8

------------------------------------------------------------------------------
       price | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         mpg |  -98.60354   80.80491    -1.22   0.227    -259.8908    62.68374
    headroom |  -565.2827   378.3782    -1.49   0.140    -1320.529    189.9633
      weight |   4.584672    1.39686     3.28   0.002     1.796528    7.372815
      length |  -58.35298   41.27537    -1.41   0.162     -140.739    24.03301
        turn |  -338.0579   127.1015    -2.66   0.010    -591.7535   -84.36232
displacement |   3.657153   6.726976     0.54   0.588    -9.769948    17.08425
       _cons |   19762.45   6048.361     3.27   0.002     7689.874    31835.03
------------------------------------------------------------------------------

. qui etable, cstat(_r_b,nformat(%3.2f))

. collect style cell result[_r_b] , nformat("%3.1f") 

. collect layout (coleq#colname#result[_r_b] result[N]) (etable_depvar#stars) ()

Collection: ETable
      Rows: coleq#colname#result[_r_b] result[N]
   Columns: etable_depvar#stars
   Table 1: 8 x 2

---------------------------------
                          price  
---------------------------------
Mileage (mpg)            -98.6   
Headroom (in.)          -565.3   
Weight (lbs.)              4.6 **
Length (in.)             -58.4   
Turn circle (ft.)       -338.1 **
Displacement (cu. in.)     3.7   
Intercept              19762.5 **
Number of observations      74   
---------------------------------

would be:

---------------------------------
                          price  
---------------------------------
Mileage (mpg)            -98.6   
Headroom (in.)          -565.3   
Weight (lbs.)              +4.6 **
Length (in.)             -58.4   
Turn circle (ft.)       -338.1 **
Displacement (cu. in.)     3.7   
Intercept              19762.5 **
Number of observations      74   
---------------------------------

thks

[Andrew Musau]

There is no numeric format that displays positive numbers with a plus sign in front of them. Note that this is not a collect issue; there are no options for this in estout (from SSC) or other exporting commands that I know of. People generally accept this convention, as there is no ambiguity in understanding that an unsigned coefficient is positive. You could request a string format, but then how would you handle negative coefficients? If all coefficients are positive, appending a plus sign could work, but the output might look strange.

sysuse auto, clear
reg price mpg headroom weight length turn displacement
etable, keep(weight displacement _cons)
collect layout (coleq#colname[weight displacement _cons]#result[_r_b _r_se] result[N]) (etable_depvar#stars[value])
collect style cell result[_r_b] , sformat("+%s")
collect preview

Res.:

. collect style cell result[_r_b] , sformat("+%s")

. 
. collect preview

---------------------------------
                          price  
---------------------------------
Weight (lbs.)              +4.585
                          (1.397)
Displacement (cu. in.)     +3.657
                          (6.727)
Intercept              +19762.453
                       (6048.361)
Number of observations         74
---------------------------------

[Luis Pecht]

thks Andrew,

The project I am working aims a non-technical audience, and , in this particular context, signs , also, means a "favorable" and "negative" consequence for the business. Prefixing signs to the estimate would help to emphasize that.

My plan B is to move to coefplot (from SSC).

thks, again.