Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Odd Error Using svy jackknife


From   "Pham, Richard H. VHAREN" <Richard.Pham@va.gov>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Odd Error Using svy jackknife
Date   Wed, 22 Apr 2009 21:25:25 -0700

Here are my Stata Settings for Stata SE 10.1 x64-bit:
version 10.1
set more off, permanently
set matsize 5000, permanently
set mem 8000m, permanently
set maxvar 5000, permanently

------------------------
I am getting an odd error running svy jackknife.  I set up the survey
and the svyset command:

. use C:\Asthma\preparation\96RXasthmalongidentified.dta, clear

. svyset varpsu [pweight=wtdper], strata(varstr) vce(linearized)
singleunit(missing)

      pweight: wtdper
          VCE: linearized
  Single unit: missing
     Strata 1: varstr
         SU 1: varpsu
        FPC 1: <zero>

When I run a Taylor-Series estimate on a mean, I get a straightforward
(and CENVAR-verified standard error):

. svy linearized, subpop(if albuterol==1) : mean rxxpx

(running mean on estimation sample)

Survey: Mean estimation

Number of strata =     134      Number of obs    =      143555
Number of PSUs   =     849      Population size  =  1816655934
                                Subpop. no. obs  =        2996
                                Subpop. size     = 35436769.95
                                Design df        =         715

--------------------------------------------------------------
             |             Linearized
             |       Mean   Std. Err.     [95% Conf. Interval]
-------------+------------------------------------------------
       rxxpx |   26.02374    .836471      24.38151    27.66597
--------------------------------------------------------------
Note: 6 strata omitted because they contain no subpopulation
      members. 


**************************************************************

When I try running this with svy jackknife, I receive an error 2000
insufficient observations to compute jackknife standard errors
no results will be saved
r(2000);


:
. svy jackknife, subpop(if albuterol==1) : mean rxxpx
(running mean on estimation sample)

Jackknife replications (867)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5 
nnnnnnnnnnnnnnnnnnnnnnnnssnnnnnnnnnnnnnnnnnnnnnnnn    50
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   100
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   150
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   200
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   250
nnnnnnnnnnnnnssnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   300
nnnnnnnnnnnnnnnnnnnnnnnnnnssssssssnnnnnnnnnnnnnnnn   350
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnssnnnnnnnn   400
nnnnnnnnnnnnnnnnnnnnssnnnnnnnnnnnnnnnnnnnnnnnnnnnn   450
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   500
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   550
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   600
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   650
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   700
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   750
nnnnnnnnnnnnnnnssnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   800
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn   850
nnnnnnnnnnnnnnnnn



************************************************************

. svy jackknife, trace subpop(if albuterol==1) : mean rxxpx
. svy jackknife, noisily subpop(if albuterol==1) : mean rxxpx

(Results omitted.)

I tried it noisily and traced, and it didn't seem to error out at any
time during the calculation, but stopped with an error for the final
answer.  I have tried playing with the n-option (n(albuterol==1)), but
that didn't seem to help either.

Is the error a programming issue on my part, or is it an insufficient
observations within the strata problem that would leave the jackknife
with a 0 for a calculation?  I have the noisy and the trace file if
anyone is interested.

Warmest regards,
Richard

---
Richard Pham, PharmD, PhD (candidate)
VISN 21 PBM Program Manager
Department of Veterans Affairs
Richard.Pham@va.gov
(775)-328-1484
Office: Building 1A, A105A
 

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index