Liz's Stata Guide

xtreg

Estiamtes fixed effects models, random effects models and others. It cannot be abbreviated.

Typical Use:
xtreg depvar indvar1 indvar2, fe i(groupvar)   - or -
xtreg depvar indvar1 indvar2, re i(groupvar)

Example

• Fixed effects:

  .  xtreg hhsize age, fe i(religion)
  
  Fixed-effects (within) regression               Number of obs      =      1081
  Group variable (i): religion                    Number of groups   =         3
  
  R-sq:  within  = 0.0134                         Obs per group: min =         6
         between = 0.4530                                        avg =     360.3
         overall = 0.0141                                        max =       822
  
                                                  F(1,1077)          =     14.64
  corr(u_i, Xb)  = 0.0531                         Prob > F           =    0.0001
  
  ------------------------------------------------------------------------------
        hhsize |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
  -------------+----------------------------------------------------------------
           age |   .0350813   .0091684     3.83   0.000     .0170913    .0530713
         _cons |   6.044863   .4038994    14.97   0.000     5.252345    6.837382
  -------------+----------------------------------------------------------------
       sigma_u |  .73678801
       sigma_e |  4.6363588
           rho |  .02463199   (fraction of variance due to u_i)
  ------------------------------------------------------------------------------
  F test that all u_i=0:     F(2, 1077) =     1.63             Prob > F = 0.1961
  

• Random effects:

  .  xtreg hhsize age, re i(religion)
  
  Random-effects GLS regression                   Number of obs      =      1081
  Group variable (i): religion                    Number of groups   =         3
  
  R-sq:  within  = 0.0134                         Obs per group: min =         6
         between = 0.4530                                        avg =     360.3
         overall = 0.0141                                        max =       822
  
  Random effects u_i ~ Gaussian                   Wald chi2(1)       =     15.48
  corr(u_i, X)       = 0 (assumed)                Prob > chi2        =    0.0001
  
  ------------------------------------------------------------------------------
        hhsize |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
  -------------+----------------------------------------------------------------
           age |    .035959   .0091407     3.93   0.000     .0180435    .0538744
         _cons |   6.008633   .4028566    14.92   0.000     5.219048    6.798217
  -------------+----------------------------------------------------------------
       sigma_u |          0
       sigma_e |  4.6363588
           rho |          0   (fraction of variance due to u_i)
  ------------------------------------------------------------------------------
  

Options

• The option , r makes Stata calculate heteroskedastic robust standard errors.

Notes

• xtreg can also estimate between-effects models, ML random-effects models and population-averaged models.
• See the official Stata help for xtreg.