How to run the regression if the frquency of a variable is higher than a specific number?

[Phuc Nguyen]

Let say I have a dataset

SIC3        Freq.    Percent        Cum.
                        
    100        5    0.00        0.00
    101        278    0.11        0.11
    102        515    0.20        0.30
    103        444    0.17        0.47
    104        938    0.36        0.83
    106        6    0.00        0.83
    108        189    0.07        0.90
    109        547    0.21        1.11

I want to run a normal OLS regress

reg y x

I am wondering what I should add to run the regression if frequency of SIC3 is higher than 10. For example above, I will not run with SIC3=100 and 106

Thanks in advance.

[Carlo Lazzaro]

Phuc:
the first line of the following code may be helpful (mutatis mutandis):

. use "https://www.stata-press.com/data/r16/nlswork.dta"
(National Longitudinal Survey.  Young Women 14-26 years of age in 1968)

. bysort idcode: egen wanted=count(year)

. xtreg ln_wage i.msp if wanted>=10, fe vce(cluster idcode)

Fixed-effects (within) regression               Number of obs     =     11,681
Group variable: idcode                          Number of groups  =        982

R-sq:                                           Obs per group:
     within  = 0.0027                                         min =         10
     between = 0.0002                                         avg =       11.9
     overall = 0.0011                                         max =         15

                                                F(1,981)          =      11.16
corr(u_i, Xb)  = -0.0217                        Prob > F          =     0.0009

                               (Std. Err. adjusted for 982 clusters in idcode)
------------------------------------------------------------------------------
             |               Robust
     ln_wage |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       1.msp |   .0451795   .0135219     3.34   0.001     .0186443    .0717147
       _cons |   1.677765   .0083301   201.41   0.000     1.661419    1.694112
-------------+----------------------------------------------------------------
     sigma_u |  .33266075
     sigma_e |  .31187398
         rho |   .5322173   (fraction of variance due to u_i)
------------------------------------------------------------------------------

.

Put differently:

bysort SIC3: egen wanted=count(SIC3)
regress y x if wanted>10