Correlation table for regression

[Mohammad Haider]

I am a new user of statistical software. I have a sample of 500 companies, however, for data missing regression was conducted on 350 samples. Now how I get the correlation table of variables for only these 350 samples. Thank you very much.

[Marcos Almeida]

Welcome to the Stata Forum/ Statalust.

Please read the FAQ, particularly the topic about sharing data/command/output.

That said, I fear a table of correlations between 350 samples would be barely readable.

[Nick Cox]

I think Mohammad means a sample of 350, not 350 samples. (In many substantive fields, a sample is a chunk of something solid or fluid or gaseous, whether such chunks are enchanting or prosaic or disgusting.)

After regress you just ask for

correlate ... if e(sample)

where the dots are for your variable names.

[Basharat Hussain]

For insight:
The purpose of 'if e(sample) restriction' please?
The results do not seem to differ.

. sysuse auto
(1978 Automobile Data)
. regress mpg price weight length i.foreign

      Source |       SS           df       MS      Number of obs   =        74
-------------+----------------------------------   F(4, 69)        =     35.58
       Model |   1645.6886         4  411.422151   Prob > F        =    0.0000
    Residual |  797.770856        69  11.5618965   R-squared       =    0.6735
-------------+----------------------------------   Adj R-squared   =    0.6546
       Total |  2443.45946        73  33.4720474   Root MSE        =    3.4003

------------------------------------------------------------------------------
         mpg |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       price |  -.0000374   .0002009    -0.19   0.853    -.0004381    .0003634
      weight |  -.0041499   .0019865    -2.09   0.040    -.0081128   -.0001871
      length |   -.086156    .058149    -1.48   0.143    -.2021601     .029848
             |
     foreign |
    Foreign  |  -1.574443   1.292234    -1.22   0.227     -4.15238    1.003494
       _cons |   50.71772   6.364007     7.97   0.000     38.02187    63.41357
------------------------------------------------------------------------------

. correlate price weight length if e(sample)
(obs=74)

             |    price   weight   length
-------------+---------------------------
       price |   1.0000
      weight |   0.5386   1.0000
      length |   0.4318   0.9460   1.0000


. correlate price weight length
(obs=74)

             |    price   weight   length
-------------+---------------------------
       price |   1.0000
      weight |   0.5386   1.0000
      length |   0.4318   0.9460   1.0000

[Nick Cox]

Naturally they don't differ when there are no missing values in the variables used, as in the case you cite. Try using i.rep78 as a predictor instead.

See also

SJ-7-2 dm0030 . . . . . . . . . . Stata tip 44: Get a handle on your sample
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Jann
Q2/07 SJ 7(2):266--267 (no commands)
tip on how to keep track of the observations that make
up your estimation sample

[Mohammad Haider]

Thank you. I got it.

[Basharat Hussain]

Got it.
Much obliged.