Getting the number of the groups and overall R-squared from RE in the result table

[Guest]

Dear All,

I want to export random effect regression result into Word table. I use estout from SSC. I run regression for 3 models m1 m2 m3 and stored it using eststo. I know how to get the number of observations. However, I don't know how to tell Stata to export number of groups and overall R-squared. I use the code:

esttab m1 m2 m3 using results.rtf, ///
cells(b(star fmt(%9.3f)) se(par) ///
stats( N, fmt(%9.3f, %9.0fc) labels("Observations")) ///
​​​legend collabels(none) varlabels(_cons Constant) ///
note(Standard errors are in parentheses.) ///
title({\b Table1.}) replace

I know when I put N after stats Stata will take the number of observations. I didn't find the abbrevation for the overall R-squared and number of groups.

[Carlo Lazzaro]

Guest,

have you already taken a look at the new -table- suite?

[Guest]

Thank you for your answer Carlo. I'm sorry but new -table- suite doesn't sound familiar to me. I googled it but found nothing. If you could provide me with the link or a hint where to look I would be grateful

[Carlo Lazzaro]

Emila:
assuming that you're working with Stata 17, just type -help table- and take a look at the related entries in Stata .pdfr manual.

[Mead Over]

After estimating any of your three random effects regression, type -ereturn list- to see the results that Stata has stored. For example, using the random effects example from Stata's help for -xtreg, re- :

clear
webuse nlswork 
xtset idcode 
xtreg ln_w grade age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure c.tenure#c.tenure 2.race not_smsa south, re theta
ereturn list

I see in the results window:

. ereturn list

scalars:
               e(rank) =  11
               e(df_m) =  10
               e(chi2) =  9244.735826836899
                  e(p) =  0
            e(sigma_u) =  .2579052592523687
            e(sigma_e) =  .2906892282370427
              e(sigma) =  .3886069352997693
                e(rho) =  .440452726681115
               e(rmse) =  .2904426239253425
                  e(N) =  28091
               e(Tbar) =  3.292459474371592
               e(Tcon) =  0
                e(N_g) =  4697
              e(g_min) =  1
              e(g_avg) =  5.980625931445603
              e(g_max) =  15
           e(thta_min) =  .2519710745439819
             e(thta_5) =  .2519710745439819
            e(thta_50) =  .5498871394578555
            e(thta_95) =  .7016329908849215
               e(r2_w) =  .171484938061179
               e(r2_b) =  .4784030188110969
               e(r2_o) =  .3707694091066819
           e(thta_max) =  .7205721155454422

These returned results are likely to include the statistics you would like to report in your table. The -table- suite of commands that Carlo Lazzaro references rely on these saved results to produce tables.

But long before Stata added the -tables- suite, Ben Jann contributed -esttab- that you are using and his companion utility -estout-. The command -esttab- can report any of these returned statistics using its option -stats()-. This behavior is documented in the help file for Jann's -estout-. To go directly to the relevant section, in Stata type:

help estout##sum

For example, the overall R^2 is saved as e(r2_o). So the -esttab- syntax is:

. esttab, stats(r2_o)

----------------------------
                      (1)   
                  ln_wage   
----------------------------
grade              0.0646***
                  (36.30)   

age                0.0368***
                  (11.80)   

c.age#c.age     -0.000713***
                 (-14.27)   

ttl_exp            0.0290***
                  (11.98)   

c.ttl_exp#~p     0.000305** 
                   (2.62)   

tenure             0.0393***
                  (22.36)   

c.tenure#c~e     -0.00200***
                 (-16.80)   

2.race            -0.0531***
                  (-5.31)   

not_smsa           -0.131***
                 (-18.23)   

south             -0.0869***
                 (-11.90)   

_cons               0.239***
                   (4.83)   
----------------------------
r2_o                0.371   
----------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

Even better, you can replace the cryptic abbreviation -r2_o- with a more comprehensible label like this:

. esttab, stats(r2_o, labels("Overall R^2"))

----------------------------
                      (1)   
                  ln_wage   
----------------------------
grade              0.0646***
                  (36.30)   

age                0.0368***
                  (11.80)   

c.age#c.age     -0.000713***
                 (-14.27)   

ttl_exp            0.0290***
                  (11.98)   

c.ttl_exp#~p     0.000305** 
                   (2.62)   

tenure             0.0393***
                  (22.36)   

c.tenure#c~e     -0.00200***
                 (-16.80)   

2.race            -0.0531***
                  (-5.31)   

not_smsa           -0.131***
                 (-18.23)   

south             -0.0869***
                 (-11.90)   

_cons               0.239***
                   (4.83)   
----------------------------
Overall R^2         0.371   
----------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

[Guest]

Thank you both for your kind help