ADDITIONAL RESULTS 
Responses of Private and Public Schools to Voucher Funding: 
The Czech Experience 
by 
Randall Filer and Daniel Munich

List of variables
VERSION PRESENTED IN THE PAPER
VERSION WITH q1 INSTEAD OF q4 on RHS
VERSION WITH q2 INSTEAD OF q4  on RHS
VERSION WITH 2 ADDITIONAL RHS VARIABLES
VERSION WITH 4 ADDITIONAL RHS VARIABLES
VERSION WITH q1 INSTEAD OF q4 on RHS + 2 ADDITIONAL RHS VARIABLES
 
 

List of variables (by appearance)

newa Dummy = 1 if non-state school appeared, 0 otherwise
edu3 District share of district population with university education
kohortA District percentage decline in high school aged cohort between 1992 and 1994
dens District population density
rune District unemployment rate
q4 Index of district public schools’ quality:
ss District number of students in technical and vocational high schools in 1991
q1 Alternative index of district public schools’ quality:
q2 Alternative index of district public schools’ quality:
lwage log(district av. wage)
ragr District share of  employment in agriculture
rcoa District share of votes for government coalition
h School concentration index in district
*VERSION PRESENTED IN THE PAPER
. dprobit  newa edu3 kohortA dens rune q4 ss if ii==1&set==0 
Probit estimates                                        Number of obs =     76
                                                        LR chi2(6)    =  32.95
                                                        Prob > chi2   = 0.0000
Log likelihood = -34.050989                             Pseudo R2     = 0.291
------------------------------------------------------------------------------
    newa |      dF/dx   Std. Err.      z    P>|z|     x-bar  [    95% C.I.   ]
---------+--------------------------------------------------------------------
    edu3 |   21.18414   8.738791     2.46   0.014   .055139   4.05643  38.3119
 kohortA |   -1.94558   .9258328    -2.07   0.038   1.35548  -3.76018 -.130981
    dens |   .0018524   .0013834     1.25   0.211   210.304  -.000859  .004564
    rune |   10.24152   4.790854     2.19   0.029   .035283   .851621  19.6314
      q4 |  -1.124217   .5840996    -1.91   0.056   -.01738  -2.26903  .020597
      ss |   .0001209   .0001607     0.76   0.448   2036.59  -.000194  .000436
---------+--------------------------------------------------------------------
  obs. P |   .3815789
 pred. P |   .5714219  (at x-bar)
------------------------------------------------------------------------------
    z and P>|z| are the test of the underlying coefficient being 0
 
. *VERSION WITH q1 INSTEAD OF q4 on RHS
. dprobit  newa edu3 kohortA dens rune q1 ss if ii==1&set==0 
Probit estimates                                        Number of obs =     76
                                                        LR chi2(6)    =  33.55
                                                        Prob > chi2   = 0.0000
Log likelihood = -33.754522                             Pseudo R2     = 0.3320
------------------------------------------------------------------------------
    newa |      dF/dx   Std. Err.      z    P>|z|     x-bar  [    95% C.I.   ]
---------+--------------------------------------------------------------------
    edu3 |   21.81399   8.888626    2.52   0.012   .055139    4.3926  39.2354
 kohortA |  -1.897152   .8924231    -2.09   0.037   1.35548  -3.64627 -.148035
    dens |   .0020018   .0013245     1.36   0.175   210.304  -.000594  .004598
    rune |   9.308694   4.811895     2.00   0.045   .035283  -.122448  18.7398
      q1 |  -.5984756   .2974465    -1.98   0.048  -.013593  -1.18146 -.015491
      ss |   .0001196   .0001604     0.75   0.452   2036.59  -.000195  .000434
---------+--------------------------------------------------------------------
  obs. P |   .3815789
 pred. P |   .5997858  (at x-bar)
------------------------------------------------------------------------------
    z and P>|z| are the test of the underlying coefficient being 0
 
. *VERSION WITH q2 INSTEAD OF q4  on RHS
. dprobit  newa edu3 kohortA dens rune q4 ss if ii==1&set==0 
Probit estimates                                        Number of obs =     76
                                                        LR chi2(6)    =  32.95
                                                        Prob > chi2   = 0.0000
Log likelihood = -34.050989                             Pseudo R2     = 0.3261
------------------------------------------------------------------------------
    newa |      dF/dx   Std. Err.      z    P>|z|     x-bar  [    95% C.I.   ]
---------+--------------------------------------------------------------------
    edu3 |   21.18414   8.738791     2.46   0.014   .055139   4.05643  38.3119
 kohortA |   -1.94558   .9258328    -2.07   0.038   1.35548  -3.76018 -.130981
    dens |   .0018524   .0013834     1.25   0.211   210.304  -.000859  .004564
    rune |   10.24152   4.790854     2.19   0.029   .035283   .851621  19.6314
      q4 |  -1.124217   .5840996    -1.91   0.056   -.01738  -2.26903  .020597
      ss |   .0001209   .0001607     0.76   0.448   2036.59  -.000194  .000436
---------+--------------------------------------------------------------------
  obs. P |   .3815789
 pred. P |   .5714219  (at x-bar)
------------------------------------------------------------------------------
    z and P>|z| are the test of the underlying coefficient being 0
. *VERSION WITH 2 ADDITIONAL RHS VARIABLES
. dprobit  newa edu3 kohortA dens rune q4 ss lwage ragr  if ii==1&set==0 
Probit estimates                                        Number of obs =     76
                                                        LR chi2(8)    =  33.03
                                                        Prob > chi2   = 0.0001
Log likelihood = -34.010388                             Pseudo R2     = 0.3269
------------------------------------------------------------------------------
    newa |      dF/dx   Std. Err.      z    P>|z|     x-bar  [    95% C.I.   ]
---------+--------------------------------------------------------------------
    edu3 |   21.20668   8.776818     2.47   0.014   .055139   4.00444  38.4089
 kohortA |  -1.881006   .9534914    -1.96   0.050   1.35548  -3.74981 -.012197
    dens |   .0019199   .0014869     1.20   0.229   210.304  -.000994  .004834
    rune |   10.23278   4.836405     2.18   0.030   .035283     .7536   19.712
      q4 |  -1.143463   .5852012    -1.94   0.052   -.01738  -2.29044   .00351
      ss |   .0001146   .0001636     0.71   0.480   2036.59  -.000206  .000435
   lwage |  -.1753134   1.024789    -0.17   0.864   8.93469  -2.18386  1.83324
    ragr |   1.104964   5.362667     0.21   0.837   .025681  -9.40567  11.6156
---------+--------------------------------------------------------------------
  obs. P |   .3815789
 pred. P |   .5739565  (at x-bar)
------------------------------------------------------------------------------
    z and P>|z| are the test of the underlying coefficient being 0
 
. *VERSION WITH 4 ADDITIONAL RHS VARIABLES
. dprobit  newa edu3 kohortA dens rune q4 ss lwage ragr rcoa h if i==1&set==0 
Probit estimates                                        Number of obs =     76
                                                        LR chi2(10)   =  33.29
                                                        Prob > chi2   = 0.0002
Log likelihood = -33.880299                             Pseudo R2     = 0.3295
------------------------------------------------------------------------------
    newa |      dF/dx   Std. Err.      z    P>|z|     x-bar  [    95% C.I.   ]
---------+--------------------------------------------------------------------
    edu3 |   21.78139   9.216148     2.41   0.016   .055139   3.71807  39.8447
 kohortA |  -1.969958    .995114    -1.96   0.050   1.35548  -3.92035 -.019571
    dens |   .0019501   .0015124     1.20   0.230   210.304  -.001014  .004914
    rune |   8.727281   5.837846     1.52   0.128   .035283  -2.71469  20.1692
      q4 |  -1.079401   .6016125    -1.78   0.075   -.01738  -2.25854  .099738
      ss |   .0001115   .0001797     0.63   0.531   2036.59  -.000241  .000464
   lwage |  -.1908873   1.056327    -0.18   0.856   8.93469  -2.26125  1.87948
    ragr |   1.590491   5.780005     0.27   0.784   .025681  -9.73811  12.9191
    rcoa |  -.7813926    1.89893    -0.41   0.681     .4186  -4.50323  2.94044
       h |  -.1835991   .5235274    -0.35   0.726   .425607  -1.20969  .842496
---------+--------------------------------------------------------------------
  obs. P |   .3815789
 pred. P |   .5758879  (at x-bar)
------------------------------------------------------------------------------
    z and P>|z| are the test of the underlying coefficient being 0
 
. *VERSION WITH q1 INSTEAD OF q4 on RHS + 2 ADDITIONAL RHS VARIABLES
. dprobit  newa edu3 kohortA dens rune q1 ss lwage ragr if ii==1&set==0 
Probit estimates                                        Number of obs =     76
                                                        LR chi2(8)    =  33.65
                                                        Prob > chi2   = 0.0000
Log likelihood = -33.700459                             Pseudo R2     = 0.3330
------------------------------------------------------------------------------
    newa |      dF/dx   Std. Err.      z    P>|z|     x-bar  [    95% C.I.   ]
---------+--------------------------------------------------------------------
    edu3 |    21.9554   8.966852     2.53   0.012   .055139   4.38069  39.5301
 kohortA |  -1.856029    .934468    -1.97   0.049   1.35548  -3.68755 -.024505
    dens |   .0019827   .0014392     1.25   0.213   210.304  -.000838  .004803
    rune |   9.440861   4.932664     2.00   0.046   .035283  -.226984  19.1087
      q1 |  -.6102569   .2992684    -2.01   0.044  -.013593  -1.19681 -.023701
      ss |   .0001181   .0001641     0.73   0.466   2036.59  -.000204   .00044
   lwage |  -.3219118   1.042937    -0.31   0.756   8.93469  -2.36603  1.72221
    ragr |   .3767844   5.341528     0.07   0.944   .025681  -10.0924   10.846
---------+--------------------------------------------------------------------
  obs. P |   .3815789
 pred. P |   .5957184  (at x-bar)
------------------------------------------------------------------------------
    z and P>|z| are the test of the underlying coefficient being 0