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

st: Post-estimation regression-survey


From   Randall Burd <burdrs@umdnj.edu>
To   statalist@hsphsun2.harvard.edu
Subject   st: Post-estimation regression-survey
Date   Mon, 08 Aug 2005 15:38:19 -0400

I am working with survey data. I am looking at the frequency of a procedure in each year from 1996 to 2003. Using basic survey commands I obtain:
 
svytotal dischgs, subpop(dischgs) by( year)
pweight:  discwt                                  Number of obs    =      8101
Strata:   fixemstr                                Number of strata =        53
PSU:      fixempsu                                Number of PSUs   =       450
Subpop.:  dischgs                                 Population size  = 39466.646
 
------------------------------------------------------------------------------
Total  Subpop. |   Estimate    Std. Err.   [95% Conf. Interval]        Deff
---------------+--------------------------------------------------------------
dischgs        |
    year==1996 |   3435.294    546.9666    2359.981    4510.607    19.57772
    year==1997 |   3829.646    603.5231    2643.146    5016.147    21.61788
    year==1998 |    3648.96    727.7667    2218.202    5079.718    32.82485
    year==1999 |    7143.03    1204.972    4774.107    9511.954    50.93736
    year==2000 |   4941.163    1032.757    2910.807     6971.52    50.64219
    year==2001 |   4907.347    1026.106    2890.067    6924.626     50.2872
    year==2002 |   3898.941    716.5405    2490.253    5307.629    29.98914
    year==2003 |   7662.265    1403.643    4902.762    10421.77     65.4869
------------------------------------------------------------------------------
I know the census population from which the samples were drawn from in each year:
 
1996 100 million
1997 102 million
1998 103 million
...
2003 107 million
 
I have gotten advice from Stata before about know how to compare the population frequency of discharges in a pairwise fashion:
 
test ([dischgs]1996)/100000000=([dischgs]1997)/102000000
test ([dischgs]1996)/100000000=([dischgs]1998)/103000000
 
etc.
 
I would now like to ask the question: has there been an increase in the population frequency of 'dischgs' from 1996 to 2003? Perhaps this can be stated as 'was there a linear increase in the population frequency of 'dischgs'.
 
Thanks.
 
Randy
 
Randall S. Burd, M.D., Ph.D.
Division of Pediatric Surgery, Department of Surgery
Box 19, 1 RWJ Place
Robert Wood Johnson Medical School
New Brunswick, NJ  08903
732-235-7821
 



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