Testing equality of coefficients across different equations after xtreg

Maria Franco

Join Date: Jul 2014

Posts: 8
#1

Testing equality of coefficients across different equations after xtreg

19 Sep 2025, 10:04

Hi everyone,

I am working with a two-wave household panel dataset. I want to estimate the following regressions:
migration_it = A₀ + A₁subjective_it + A₂controls_it + e1
migration_it = B₀ + B₁objective_it + B₂controls_it + e2
where subjective, objective, and controls are vectors that contain 3 continuous variables each, called subjective1, subjective2, subjective3, objective1, objective2, etc.

I would like to compare the coefficients in A₁ and B₁. My code is as follows:

Code:

xtset household_id wave, yearly xtreg y1 subjective1 subjective2 subjective3 controls, fe vce(robust) est store betas1 xtreg y1 objective1 objective2 objective3 controls, fe vce(robust) est store betas2 suest betas1 betas2, vce(robust) test [betas1_mean]subjective1 = [betas2_mean]objective1 test [betas1_mean]subjective2 = [betas2_mean]objective2 test [betas1_mean]subjective3 = [betas2_mean]objective3

However, I get error r(322) "xtreg is not supported by suest". I have also read about this error in previous Stata Forum threads, but I think they don't apply to my case because I am not looking at different subgroups, but at different variables in the same sample. I'd be grateful if anyone could help. I am using Stata 18 on Windows 11.

Many thanks.

Maria
Tags: None
Clyde Schechter

Join Date: Apr 2014

Posts: 30365
#2

19 Sep 2025, 12:09

...but I think they don't apply to my case because I am not looking at different subgroups, but at different variables in the same sample.

No, that doesn't matter. -suest- simply doesn't support -xtreg-, regardless of why you are using it. Be aware also that -sureg- does not permit the user of robust variance estimation in the regressions that you provide it with: you must use ordinary variance estimation in those, and then apply robust variance estimation in -suest- itself.

You can get around the problem that -suest- doesn't support -xtreg- by emulating it with -regress ... i.panel_var-. Like this:

Code:

xtreg y1 subjective1 subjective2 subjective3 controls i.household_id est store betas1 xtreg y1 objective1 objective2 objective3 controls i.household_id est store betas2 suest betas1 betas2, vce(robust) test [betas1_mean]subjective1 = [betas2_mean]objective1 test [betas1_mean]subjective2 = [betas2_mean]objective2 test [betas1_mean]subjective3 = [betas2_mean]objective3
Comment

Andrew Musau

Join Date: Oct 2014
Posts: 10503

20 Sep 2025, 00:18

Originally posted by Maria Franco View Post

However, I get error r(322) "xtreg is not supported by suest"

In Clyde's code in #2, you should change xtreg to regress. Additionally, you need to rename the panel identifier (the FE variable) differently across estimations; otherwise, you'll get a factor-variable base-level conflict error. An alternative is to restructure your data so that the variables subjective and objective are combined into a single variable, and then use xtreg by interacting the identifier of that variable with all RHS variables. In previous versions of Stata (<18), this would have been complicated to implement, but now you can use the combination of xtset and -absorb()- to achieve this (see below). Finally, note that in recent versions of Stata, sureg does allow robust and cluster-robust variance estimation, so you could use that command directly.

Code:

webuse nlswork, clear
keep in 1/300
xtreg ln_wage hours tenure ind_code, fe robust
xtreg ln_wage hours tenure occ_code, fe robust

expand 2, g(new)
gen code= cond(new, occ_code, ind_code)
lab def _merge 1 "occ_code" 0 "ind_code", modify
gen absorb_var= 0.new#c.idcode
gen fe_var= 1.new#c.idcode
xtset fe_var
xtreg ln_wage i.new#(c.hours c.tenure c.code), fe absorb(absorb_var) robust

Res.:

Code:

. xtreg ln_wage hours tenure ind_code, fe robust

Fixed-effects (within) regression               Number of obs     =        297
Group variable: idcode                          Number of groups  =         38

R-squared:                                      Obs per group:
     Within  = 0.1507                                         min =          1
     Between = 0.0032                                         avg =        7.8
     Overall = 0.0007                                         max =         15

                                                F(3, 37)          =      12.64
corr(u_i, Xb) = -0.3891                         Prob > F          =     0.0000

                                (Std. err. adjusted for 38 clusters in idcode)
------------------------------------------------------------------------------
             |               Robust
     ln_wage | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       hours |   .0025956   .0041353     0.63   0.534    -.0057834    .0109746
      tenure |   .0350126   .0062749     5.58   0.000     .0222985    .0477267
    ind_code |  -.0217159   .0118752    -1.83   0.076    -.0457773    .0023456
       _cons |   1.911541   .2119714     9.02   0.000     1.482046    2.341036
-------------+----------------------------------------------------------------
     sigma_u |  .40117964
     sigma_e |  .25579615
         rho |  .71096121   (fraction of variance due to u_i)
------------------------------------------------------------------------------

.
. xtreg ln_wage hours tenure occ_code, fe robust

Fixed-effects (within) regression               Number of obs     =        298
Group variable: idcode                          Number of groups  =         38

R-squared:                                      Obs per group:
     Within  = 0.1413                                         min =          1
     Between = 0.1006                                         avg =        7.8
     Overall = 0.0397                                         max =         15

                                                F(3, 37)          =      10.61
corr(u_i, Xb) = -0.1451                         Prob > F          =     0.0000

                                (Std. err. adjusted for 38 clusters in idcode)
------------------------------------------------------------------------------
             |               Robust
     ln_wage | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       hours |   .0033687   .0040063     0.84   0.406    -.0047489    .0114863
      tenure |   .0345895   .0062077     5.57   0.000     .0220115    .0471675
    occ_code |   .0085758   .0142943     0.60   0.552    -.0203871    .0375388
       _cons |   1.677546   .1457919    11.51   0.000     1.382143    1.972948
-------------+----------------------------------------------------------------
     sigma_u |  .37166919
     sigma_e |  .25671007
         rho |  .67702073   (fraction of variance due to u_i)
------------------------------------------------------------------------------

.
.
. xtreg ln_wage i.new#(c.hours c.tenure c.code), fe absorb(absorb_var) robust

Halperin APM for regression coefficients:

Dependent variable:
Iteration 1:  Maximum absolute difference = 2.194e-15

Independent variables:
Iteration 1:  Maximum absolute difference = 3.375e-14

Halperin APM for panel effects:
Iteration 1:  Maximum absolute difference =     1.115
Iteration 2:  Maximum absolute difference = 7.401e-17

Fixed-effects (within) regression               Number of obs     =        595
Group variable: fe_var                          Number of groups  =         39

R-squared:                                      Obs per group:
     Within  = 0.5230                                         min =          1
     Between = 0.0855                                         avg =       15.3
     Overall = 0.0064                                         max =        297

                                                F(3, 38)          =          .
corr(u_i, Xb) = -0.3458                         Prob > F          =          .

--------------------------
Absorbed variable | Levels
------------------+-------
       absorb_var |     39
--------------------------
                                (Std. err. adjusted for 39 clusters in fe_var)
------------------------------------------------------------------------------
             |               Robust
     ln_wage | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
 new#c.hours |
   ind_code  |   .0025956   1.03e-16  2.5e+13   0.000     .0025956    .0025956
   occ_code  |   .0033687    .004141     0.81   0.421    -.0050143    .0117517
             |
new#c.tenure |
   ind_code  |   .0350126  1.09e-16  3.2e+14   0.000     .0350126    .0350126
   occ_code  |   .0345895   .0064163     5.39   0.000     .0216003    .0475787
             |
  new#c.code |
   ind_code  |  -.0217159   3.85e-16 -5.6e+13   0.000    -.0217159   -.0217159
   occ_code  |   .0085758   .0147747     0.58   0.565     -.021334    .0384857
             |
       _cons |   1.794347   .0754727    23.77   0.000      1.64156    1.947133
-------------+----------------------------------------------------------------
     sigma_u |  .36910439
     sigma_e |  .25650453
         rho |  .67433687   (fraction of variance due to u_i)
------------------------------------------------------------------------------

.

Last edited by Andrew Musau; 20 Sep 2025, 00:54.

Comment

Andrew Musau

Join Date: Oct 2014
Posts: 10503

20 Sep 2025, 01:15

Actually, in #3, we should cluster on idcode to replicate the robust standard errors of the original estimations. However, xtreg does not allow non-nested clusters and the absorb() option simultaneously. You could instead install reghdfe from https://github.com/sergiocorreia/reghdfe, which handles this. Therefore, using xtreg as above is not a viable solution.

Code:

webuse nlswork, clear
keep in 1/300
xtreg ln_wage hours tenure ind_code, fe robust
xtreg ln_wage hours tenure occ_code, fe robust

expand 2, g(new)
gen code= cond(new, occ_code, ind_code)
gen absorb_var= 0.new#c.idcode
gen fe_var= 1.new#c.idcode
lab def _merge 0 "occ_code" 1 "ind_code", modify
reghdfe ln_wage i.new#(c.hours c.tenure c.code), absorb(fe_var absorb_var) cluster(idcode)

Res.:

Code:

. reghdfe ln_wage i.new#(c.hours c.tenure c.code), absorb(fe_var absorb_var) cluster(idcode)
(dropped 6 singleton observations)
(MWFE estimator converged in 2 iterations)

HDFE Linear regression                            Number of obs   =        589
Absorbing 2 HDFE groups                           F(   6,     34) =       7.02
Statistics robust to heteroskedasticity           Prob > F        =     0.0001
                                                  R-squared       =     0.6649
                                                  Adj R-squared   =     0.6159
                                                  Within R-sq.    =     0.1460
Number of clusters (idcode)  =         35         Root MSE        =     0.2563

                                (Std. err. adjusted for 35 clusters in idcode)
------------------------------------------------------------------------------
             |               Robust
     ln_wage | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
 new#c.hours |
   occ_code  |   .0025956     .00441     0.59   0.560    -.0063665    .0115577
   ind_code  |   .0033687   .0042724     0.79   0.436    -.0053139    .0120513
             |
new#c.tenure |
   occ_code  |   .0350126   .0066916     5.23   0.000     .0214137    .0486115
   ind_code  |   .0345895     .00662     5.22   0.000      .021136     .048043
             |
  new#c.code |
   occ_code  |  -.0217159   .0126638    -1.71   0.095    -.0474517      .00402
   ind_code  |   .0085758   .0152437     0.56   0.577    -.0224032    .0395548
             |
       _cons |   1.797984   .1816269     9.90   0.000     1.428874    2.167094
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
      fe_var |        36           0          36     |
  absorb_var |        36           2          34     |
-----------------------------------------------------+

Comment

Carlo Lazzaro

Join Date: Apr 2014
Posts: 17859

20 Sep 2025, 01:56

Maria:
as an aside to previous helpful replies, depending on the number of your observations, you can go -regress- (with -i.panelid- and -i.time- among predictors) instead of -xtreg,fe-:

Code:

use "https://www.stata-press.com/data/r19/nlswork.dta"

. reg ln_wage i.year i.idcode if union==0 & idcode<=10

      Source |       SS           df       MS      Number of obs   =        42
-------------+----------------------------------   F(18, 23)       =      3.08
       Model |   2.7914684        18  .155081578   Prob > F        =    0.0061
    Residual |  1.15837794        23  .050364258   R-squared       =    0.7067
-------------+----------------------------------   Adj R-squared   =    0.4772
       Total |  3.94984634        41  .096337716   Root MSE        =    .22442

------------------------------------------------------------------------------
     ln_wage | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
        year |
         71  |  -.0983753    .167345    -0.59   0.562    -.4445548    .2478042
         72  |   -.113345   .1781259    -0.64   0.531    -.4818266    .2551366
         73  |   -.106495   .1944856    -0.55   0.589    -.5088192    .2958291
         77  |  -.0730668   .1783079    -0.41   0.686    -.4419248    .2957912
         78  |  -.1768894    .278609    -0.63   0.532    -.7532361    .3994572
         80  |  -.1665012   .1739248    -0.96   0.348    -.5262921    .1932897
         82  |  -.2953811   .1739248    -1.70   0.103     -.655172    .0644098
         83  |  -.1486247   .2304064    -0.65   0.525    -.6252567    .3280073
         85  |   .0784635    .203777     0.39   0.704    -.3430813    .5000082
         87  |  -.3596358   .1935054    -1.86   0.076    -.7599322    .0406606
         88  |    .140386   .2153137     0.65   0.521    -.3050244    .5857963
             |
      idcode |
          2  |  -.3930239   .3521312    -1.12   0.276    -1.121463    .3354151
          3  |   -.108223   .2610486    -0.41   0.682    -.6482433    .4317972
          4  |   -.167938   .2857371    -0.59   0.562    -.7590302    .4231543
          5  |   .1650127   .2620287     0.63   0.535     -.377035    .7070604
          6  |   .1048737   .2609581     0.40   0.691    -.4349592    .6447066
          7  |  -.4668099   .2793818    -1.67   0.108    -1.044755    .1111355
         10  |  -.4316578   .2658454    -1.62   0.118    -.9816008    .1182853
             |
       _cons |   1.851747   .2866321     6.46   0.000     1.258804    2.444691
------------------------------------------------------------------------------

. estimate store A

. reg ln_wage i.year i.idcode if union==1 & idcode<=10

      Source |       SS           df       MS      Number of obs   =        33
-------------+----------------------------------   F(15, 17)       =      2.45
       Model |  2.26062956        15  .150708638   Prob > F        =    0.0395
    Residual |  1.04726378        17  .061603752   R-squared       =    0.6834
-------------+----------------------------------   Adj R-squared   =    0.4041
       Total |  3.30789334        32  .103371667   Root MSE        =     .2482

------------------------------------------------------------------------------
     ln_wage | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
        year |
         71  |   .1659322    .248201     0.67   0.513    -.3577262    .6895906
         72  |    .012552   .2308652     0.05   0.957    -.4745309     .499635
         73  |   .0687019   .3127218     0.22   0.829    -.5910834    .7284873
         77  |   .3861009   .2620758     1.47   0.159    -.1668306    .9390325
         78  |   .4193511    .251016     1.67   0.113    -.1102464    .9489486
         80  |   .4646076   .2418379     1.92   0.072    -.0456258     .974841
         82  |   .5374705   .2620758     2.05   0.056    -.0154611    1.090402
         83  |   .5176834    .236394     2.19   0.043     .0189356    1.016431
         85  |    .592798   .2418379     2.45   0.025     .0825646    1.103031
         87  |   .4106394   .2216673     1.85   0.081    -.0570377    .8783164
         88  |   .4606505   .2216673     2.08   0.053    -.0070265    .9283275
             |
      idcode |
          2  |  -.6379563    .142936    -4.46   0.000    -.9395249   -.3363878
          4  |   .0768979   .1678046     0.46   0.653    -.2771389    .4309347
          6  |  -.4699437   .2324597    -2.02   0.059    -.9603909    .0205034
          9  |  -.2216597   .1339446    -1.65   0.116    -.5042581    .0609388
             |
       _cons |   1.952749   .2201417     8.87   0.000     1.488291    2.417208
------------------------------------------------------------------------------

. estimate store B

. suest A B, vce(cluster idcode)

Simultaneous results for A, B                               Number of obs = 75

                                 (Std. err. adjusted for 9 clusters in idcode)
------------------------------------------------------------------------------
             |               Robust
             | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
A_mean       |
        year |
         71  |  -.0983753    .107276    -0.92   0.359    -.3086324    .1118819
         72  |   -.113345   .1193757    -0.95   0.342    -.3473171    .1206272
         73  |   -.106495   .1357488    -0.78   0.433    -.3725577    .1595676
         77  |  -.0730668    .108068    -0.68   0.499    -.2848762    .1387426
         78  |  -.1768894   .1240099    -1.43   0.154    -.4199443    .0661655
         80  |  -.1665012   .1049822    -1.59   0.113    -.3722625    .0392601
         82  |  -.2953811   .1680191    -1.76   0.079    -.6246925    .0339303
         83  |  -.1486247   .1245632    -1.19   0.233    -.3927642    .0955147
         85  |   .0784635   .1450176     0.54   0.588    -.2057658    .3626927
         87  |  -.3596358   .3300229    -1.09   0.276    -1.006469    .2871972
         88  |    .140386   .1085514     1.29   0.196    -.0723709    .3531428
             |
      idcode |
          2  |  -.3930239   .1255395    -3.13   0.002    -.6390768   -.1469709
          3  |   -.108223    .030607    -3.54   0.000    -.1682116   -.0482344
          4  |   -.167938   .0319708    -5.25   0.000    -.2305995   -.1052764
          5  |   .1650127   .0401191     4.11   0.000     .0863808    .2436446
          6  |   .1048737   .0306692     3.42   0.001     .0447632    .1649842
          7  |  -.4668099   .0456375   -10.23   0.000    -.5562577   -.3773621
         10  |  -.4316578   .0361566   -11.94   0.000    -.5025234   -.3607921
             |
       _cons |   1.851747    .108068    17.14   0.000     1.639938    2.063557
-------------+----------------------------------------------------------------
A_lnvar      |
       _cons |  -2.988474   .2892987   -10.33   0.000    -3.555489   -2.421458
-------------+----------------------------------------------------------------
B_mean       |
        year |
         71  |   .1659322   .0591406     2.81   0.005     .0500188    .2818457
         72  |    .012552   .2029576     0.06   0.951    -.3852376    .4103416
         73  |   .0687019   .2149014     0.32   0.749     -.352497    .4899009
         77  |   .3861009    .235082     1.64   0.101    -.0746514    .8468533
         78  |   .4193511   .2419553     1.73   0.083    -.0548725    .8935748
         80  |   .4646076   .2405487     1.93   0.053    -.0068592    .9360745
         82  |   .5374705   .2627602     2.05   0.041     .0224699    1.052471
         83  |   .5176834   .2702686     1.92   0.055    -.0120333      1.0474
         85  |    .592798   .2697792     2.20   0.028     .0640405    1.121555
         87  |   .4106394   .3804763     1.08   0.280    -.3350805    1.156359
         88  |   .4606505   .3169904     1.45   0.146    -.1606392     1.08194
             |
      idcode |
          2  |  -.6379563   .0287222   -22.21   0.000    -.6942508   -.5816619
          4  |   .0768979     .09117     0.84   0.399    -.1017921    .2555879
          6  |  -.4699437   .0149967   -31.34   0.000    -.4993366   -.4405508
          9  |  -.2216597   .0474788    -4.67   0.000    -.3147165   -.1286029
             |
       _cons |   1.952749   .2618481     7.46   0.000     1.439537    2.465962
-------------+----------------------------------------------------------------
B_lnvar      |
       _cons |  -2.787033   .4095288    -6.81   0.000    -3.589694   -1.984371
------------------------------------------------------------------------------

. test [A_mean = B_mean], common

 ( 1)  [A_mean]70b.year - [B_mean]70b.year = 0
 ( 2)  [A_mean]71.year - [B_mean]71.year = 0
 ( 3)  [A_mean]72.year - [B_mean]72.year = 0
 ( 4)  [A_mean]73.year - [B_mean]73.year = 0
 ( 5)  [A_mean]77.year - [B_mean]77.year = 0
 ( 6)  [A_mean]78.year - [B_mean]78.year = 0
 ( 7)  [A_mean]80.year - [B_mean]80.year = 0
 ( 8)  [A_mean]82.year - [B_mean]82.year = 0
 ( 9)  [A_mean]83.year - [B_mean]83.year = 0
 (10)  [A_mean]85.year - [B_mean]85.year = 0
 (11)  [A_mean]87.year - [B_mean]87.year = 0
 (12)  [A_mean]88.year - [B_mean]88.year = 0
 (13)  [A_mean]1b.idcode - [B_mean]1b.idcode = 0
 (14)  [A_mean]2.idcode - [B_mean]2.idcode = 0
 (15)  [A_mean]4.idcode - [B_mean]4.idcode = 0
 (16)  [A_mean]6.idcode - [B_mean]6.idcode = 0
       Constraint 1 dropped
       Constraint 4 dropped
       Constraint 5 dropped
       Constraint 6 dropped
       Constraint 7 dropped
       Constraint 9 dropped
       Constraint 13 dropped
       Constraint 16 dropped

           chi2(  8) =  184.42
         Prob > chi2 =    0.0000

.

Kind regards,
Carlo
(Stata 19.0)

Comment

Clyde Schechter

Join Date: Apr 2014

Posts: 30365
#6

20 Sep 2025, 11:12

In Clyde's code in #2, you should change xtreg to regress.

Yes, of course. Sorry about that. I don't know why I made that mistake. I actually tested out my code with some toy data, using -regress-, and I don't recall how it got changed back to -xtreg-. But Andrew Musau is absolutely correct about that. That's even what I meant when I spoke of emulating -xtreg- with -regress i.panel_var-. The code I intended to post is:

Code:

regress y1 subjective1 subjective2 subjective3 controls i.household_id est store betas1 regress y1 objective1 objective2 objective3 controls i.household_id est store betas2 suest betas1 betas2, vce(robust) test [betas1_mean]subjective1 = [betas2_mean]objective1 test [betas1_mean]subjective2 = [betas2_mean]objective2 test [betas1_mean]subjective3 = [betas2_mean]objective3

Additionally, you need to rename the panel identifier (the FE variable) differently across estimations; otherwise, you'll get a factor-variable base-level conflict error.

I don't see why this would be the case. Unless missing values of the variables subjective and objective lead to estimation sets with different values of the household_id variable,* I don't think you would get a base value conflict. I didn't encounter that problem when I tested the code on toy data.

All of that said, I think that the -reghdfe- based solution he gave in #4 is better.

*Added: And in that case, the analysis is not giving you a bona fide comparison of the subjective and objective variables: the estimation of both models needs to be carried out on the same estimation set. So, if necessary, you would need to restrict the regression to the subset of observations that are retained in the estimation sample of both models, which would eliminate any base level conflict for household_id as well.

Last edited by Clyde Schechter; 20 Sep 2025, 11:16.
1 like
Comment

Andrew Musau

Join Date: Oct 2014
Posts: 10503

20 Sep 2025, 13:51

Originally posted by Clyde Schechter View Post

I don't think you would get a base value conflict. I didn't encounter that problem when I tested the code on toy data.

You are correct, Clyde—that does not apply in this case. For some reason, I was thinking of cases where the estimation sample differs across regressions (e.g., as shown below). The OP should disregard my earlier comment.

Code:

webuse nlswork, clear
keep in 1/100
reg ln_wage hours tenure ind_code i.id in 1/50
est sto est1
reg ln_wage hours tenure occ_code i.id in 50/l
suest est1 .

Res.:

Code:

. reg ln_wage hours tenure ind_code i.id in 1/50

      Source |       SS           df       MS      Number of obs   =        50
-------------+----------------------------------   F(6, 43)        =      4.19
       Model |  2.68739387         6  .447898979   Prob > F        =    0.0021
    Residual |  4.59270086        43  .106806997   R-squared       =    0.3691
-------------+----------------------------------   Adj R-squared   =    0.2811
       Total |  7.28009473        49  .148573362   Root MSE        =    .32681

------------------------------------------------------------------------------
     ln_wage | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       hours |   .0102175   .0076697     1.33   0.190    -.0052498    .0256849
      tenure |   .0411967   .0191277     2.15   0.037      .002622    .0797713
    ind_code |  -.0008585   .0471828    -0.02   0.986    -.0960116    .0942946
             |
      idcode |
          2  |  -.4442373   .1544554    -2.88   0.006    -.7557264   -.1327483
          3  |  -.5408019   .1468373    -3.68   0.001    -.8369276   -.2446762
          4  |  -.1043564   .2991615    -0.35   0.729     -.707673    .4989602
             |
       _cons |    1.59052   .4461646     3.56   0.001     .6907435    2.490297
------------------------------------------------------------------------------

. 
. est sto est1

. 
. reg ln_wage hours tenure occ_code i.id in 50/l

      Source |       SS           df       MS      Number of obs   =        50
-------------+----------------------------------   F(8, 41)        =     17.57
       Model |  4.38869461         8  .548586826   Prob > F        =    0.0000
    Residual |  1.28030985        41  .031227069   R-squared       =    0.7742
-------------+----------------------------------   Adj R-squared   =    0.7301
       Total |  5.66900446        49  .115693969   Root MSE        =    .17671

------------------------------------------------------------------------------
     ln_wage | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       hours |   .0201751   .0113163     1.78   0.082    -.0026786    .0430288
      tenure |   .0314927   .0066575     4.73   0.000     .0180475    .0449379
    occ_code |     -.1081   .1120734    -0.96   0.340    -.3344369    .1182369
             |
      idcode |
          5  |  -1.382497   1.131884    -1.22   0.229    -3.668383      .90339
          6  |  -1.343679   .7972854    -1.69   0.100    -2.953829    .2664718
          7  |  -1.856268   1.122798    -1.65   0.106    -4.123806    .4112696
          9  |  -1.435336   1.134154    -1.27   0.213    -3.725806    .8551338
         10  |  -1.428577   .8068323    -1.77   0.084    -3.058008    .2008535
             |
       _cons |   2.703807   1.438797     1.88   0.067    -.2019038    5.609517
------------------------------------------------------------------------------

. 
. suest est1 .
idcode: factor variable base category conflict
r(198);

Comment

Maria Franco

Join Date: Jul 2014

Posts: 8
#8

22 Sep 2025, 03:37

Thank you so much, Clyde, Andrew, and Carlo, for your extremely helpful responses.
Clyde's code in #6 worked wonders.
Many thanks again.
Maria
Comment

Announcement