Dear stata users:
I have two questions related to running fuzzy rd in stata and hope someone can help.
  1. Is it possible to do pooled subgroup analysis with the command rdrobust (instead of running one regression for each gender for example)?
  2. I am having difficulties getting the same fuzzy rd estimates with rdrobust and ivregress. Here is the code and output (with stata 14). Maybe I am missing something obvious here.
Z is the running variable in terms of GPA
ac_num is the number of college a student got accepted to.

Code:

gen above =(z>0)
gen uw=1/2*(abs(z)<=1) //uniform weight
gen zabove=z*above

rdrobust ac_num z, fuzzy(treated) h(0.208) kernel(uniform)
ivreg2 ac_num (treated=above) z zabove [pw=uw] if z>=-0.208 & z<=0.208


Output:
Code:
. rdrobust ac_num z, fuzzy(treated) h(0.208) kernel(uniform)

Fuzzy RD estimates using local polynomial regression.
 
      Cutoff c = 0 | Left of c  Right of c            Number of obs =     115413
-------------------+----------------------            BW type       =     Manual
     Number of obs |     68603       46810            Kernel        =    Uniform
Eff. Number of obs |     11247       13008            VCE method    =         NN
    Order est. (p) |         1           1
    Order bias (q) |         2           2
       BW est. (h) |     0.208       0.208
       BW bias (b) |     0.208       0.208
         rho (h/b) |     1.000       1.000
 
First-stage estimates. Outcome: treated. Running variable: z.
--------------------------------------------------------------------------------
            Method |   Coef.    Std. Err.    z     P>|z|    [95% Conf. Interval]
-------------------+------------------------------------------------------------
      Conventional |  .12829     .03517   3.6477   0.000     .05936       .19723
            Robust |     -          -     2.3896   0.017    .023406      .236954
--------------------------------------------------------------------------------
 
Treatment effect estimates. Outcome: ac_num. Running variable: z. Treatment Status: treated.
--------------------------------------------------------------------------------
            Method |   Coef.    Std. Err.    z     P>|z|    [95% Conf. Interval]
-------------------+------------------------------------------------------------
      Conventional |  .06976      .1392   0.5012   0.616   -.203061      .342579
            Robust |     -          -     0.1367   0.891     -.4409      .507025
--------------------------------------------------------------------------------


. ivreg2 ac_num (treated=above) z zabove [pw=uw] if z>=-0.208 & z<=0.208
(sum of wgt is     1.2128e+04)
 
IV (2SLS) estimation
--------------------
 
Estimates efficient for homoskedasticity only
Statistics robust to heteroskedasticity
 
                                                      Number of obs =    24255
                                                      F(  3, 24251) =   115.04
                                                      Prob > F      =   0.0000
Total (centered) SS     =  11836.76809                Centered R2   =  -0.0999
Total (uncentered) SS   =        14482                Uncentered R2 =   0.1010
Residual SS             =  13019.64892                Root MSE      =    .7327
 
------------------------------------------------------------------------------
             |               Robust
      ac_num |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     treated |  -.1120709   .1809334    -0.62   0.536    -.4666938     .242552
           z |   .8990076   .3405309     2.64   0.008     .2315793    1.566436
      zabove |   .1783963   .2341688     0.76   0.446     -.280566    .6373587
       _cons |   .3866149   .0952684     4.06   0.000     .1998921    .5733376
------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic):              6.887
                                                   Chi-sq(1) P-val =    0.0087
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic):                8.053
                         (Kleibergen-Paap rk Wald F statistic):          6.891
Stock-Yogo weak ID test critical values: 10% maximal IV size             16.38
                                         15% maximal IV size              8.96
                                         20% maximal IV size              6.66
                                         25% maximal IV size              5.53
Source: Stock-Yogo (2005).  Reproduced by permission.
NB: Critical values are for Cragg-Donald F statistic and i.i.d. errors.
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments):         0.000
                                                 (equation exactly identified)
------------------------------------------------------------------------------
Instrumented:         treated
Included instruments: z zabove
Excluded instruments: above
------------------------------------------------------------------------------