Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Do dropped weights invalidate the DuMouchel-Duncan test?


From   "Clive Nicholas" <[email protected]>
To   [email protected]
Subject   st: Do dropped weights invalidate the DuMouchel-Duncan test?
Date   Tue, 13 Jul 2004 02:53:07 +0100 (BST)

All,

I've just run an OLS test in Stata 8.2 with the aid of Michael Mitchell
and Phil Ender's -xi3- package, downloadable from SSC (so that I could
automatically produce weight*region dummy interactions):

. xi3: reg edlabpc ledlabpc ed2-ed13 edpollch lagconch laglabch lagldmch
clmargin ldmargin labplace edenp c
> lass rurnorth restnw rurwales indswles perwmids eastmids eanglia
corndevn wessex roseland wt_ed2 wt_ed3 w
> t_ed4 wt_ed5 wt_ed6 wt_ed7 wt_ed8 wt_ed9 wt_ed10 wt_ed11 wt_ed12 wt_ed13
wt_edpollch wt_lagconch wt_lagla
> bch wt_lagldmch wt_clmargin wt_ldmargin wt_labplace wt_edenp wt_class
weight*i.region if edmarker==1
i.region          _Iregion_4-22       (naturally coded; _Iregion_4 omitted)

      Source |       SS       df       MS              Number of obs =   
1842
-------------+------------------------------           F( 49,  1792) =  
99.60
       Model |  306063.379    49   6246.1914           Prob > F      = 
0.0000
    Residual |  112384.594  1792  62.7146174           R-squared     = 
0.7314
-------------+------------------------------           Adj R-squared = 
0.7241
       Total |  418447.973  1841  227.293847           Root MSE      = 
7.9193

------------------------------------------------------------------------------
     edlabpc |      Coef.   Std. Err.      t    P>|t|     [95% Conf.
Interval]
-------------+----------------------------------------------------------------
    ledlabpc |   .3836418   .0189998    20.19   0.000     .3463777    
.420906
         ed2 |  (dropped)
         ed3 |   9.824244   61.86381     0.16   0.874    -111.5085    
131.157
         ed4 |  (dropped)
         ed5 |   42.29502   40.81002     1.04   0.300     -37.7452   
122.3352
         ed6 |   14.82855   34.32255     0.43   0.666    -52.48787   
82.14497

[...]

    wt_edenp |    10.2346   12.59054     0.81   0.416    -14.45909   
34.92829
    wt_class |   1.713143   .5285719     3.24   0.001     .6764606   
2.749825
      weight |  (dropped)
  _Iregion_8 |  -5.153087   1.875245    -2.75   0.006    -8.830983  
-1.475191
 _Iregion_11 |  (dropped)
 _Iregion_12 |  -10.85289   2.475908    -4.38   0.000    -15.70886  
-5.996917

[...]

   _IweXre16 |  -18.31036   7.794188    -2.35   0.019    -33.59702  
-3.023711
   _IweXre17 |  (dropped)
   _IweXre18 |  (dropped)
   _IweXre21 |  (dropped)
   _IweXre22 |  -14.36693   5.297401    -2.71   0.007    -24.75666  
-3.977195
       _cons |  -35.20043   39.90924    -0.88   0.378     -113.474   
43.07311
------------------------------------------------------------------------------

As you might be able to spot, Stata dropped the weight from the model. I
don't know why. Does this mean that the DuMouchel-Duncan F-test that I ran
afterwards is invalid:

. test weight wt_ed2 wt_ed3 wt_ed4 wt_ed5 wt_ed6 wt_ed7 wt_ed8 wt_ed9
wt_ed10 wt_ed11 wt_ed12 wt_ed13 wt_ed
> pollch wt_lagconch wt_laglabch wt_lagldmch wt_clmargin wt_ldmargin
wt_labplace wt_edenp wt_class _Iregion
> _8 _Iregion_11 _Iregion_12 _Iregion_14 _Iregion_15 _Iregion_16
_Iregion_17 _Iregion_18 _Iregion_21 _Iregi
> on_22 _IweXre8 _IweXre11 _IweXre12 _IweXre14 _IweXre15 _IweXre16
_IweXre17 _IweXre18 _IweXre21 _IweXre22
> if e(sample)

 ( 1)  weight = 0
 ( 2)  wt_ed2 = 0
 ( 3)  wt_ed3 = 0
 ( 4)  wt_ed4 = 0
 ( 5)  wt_ed5 = 0
 ( 6)  wt_ed6 = 0
 ( 7)  wt_ed7 = 0
 ( 8)  wt_ed8 = 0
 ( 9)  wt_ed9 = 0
 (10)  wt_ed10 = 0
 (11)  wt_ed11 = 0
 (12)  wt_ed12 = 0
 (13)  wt_ed13 = 0
 (14)  wt_edpollch = 0
 (15)  wt_lagconch = 0
 (16)  wt_laglabch = 0
 (17)  wt_lagldmch = 0
 (18)  wt_clmargin = 0
 (19)  wt_ldmargin = 0
 (20)  wt_labplace = 0
 (21)  wt_edenp = 0
 (22)  wt_class = 0
 (23)  _Iregion_8 = 0
 (24)  _Iregion_11 = 0
 (25)  _Iregion_12 = 0
 (26)  _Iregion_14 = 0
 (27)  _Iregion_15 = 0
 (28)  _Iregion_16 = 0
 (29)  _Iregion_17 = 0
 (30)  _Iregion_18 = 0
 (31)  _Iregion_21 = 0
 (32)  _Iregion_22 = 0
 (33)  _IweXre8 = 0
 (34)  _IweXre11 = 0
 (35)  _IweXre12 = 0
 (36)  _IweXre14 = 0
 (37)  _IweXre15 = 0
 (38)  _IweXre16 = 0
 (39)  _IweXre17 = 0
 (40)  _IweXre18 = 0
 (41)  _IweXre21 = 0
 (42)  _IweXre22 = 0
       Constraint 1 dropped
       Constraint 2 dropped
       Constraint 13 dropped
       Constraint 24 dropped
       Constraint 26 dropped
       Constraint 27 dropped
       Constraint 28 dropped
       Constraint 29 dropped
       Constraint 30 dropped
       Constraint 32 dropped
       Constraint 33 dropped
       Constraint 35 dropped
       Constraint 37 dropped
       Constraint 39 dropped
       Constraint 40 dropped
       Constraint 41 dropped

       F( 26,  1792) =    3.07
            Prob > F =    0.0000

Apologies for the spew of Stata output. Thanks in advance.

CLIVE NICHOLAS        |t: 0(044)191 222 5969
Politics              |e: [email protected]
Newcastle University  |http://www.ncl.ac.uk/geps
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index