I'm using Stata 8. I think there is a problem with -dstdize-, or I'm
overlooking something. In short, if you use an external reference
population to directly standardise rates by age and sex (as in the
example on page 291, vol 1 of the manual), the resulting standardised
rates are half of what they should be.
The sum of the weights for the standard population distribution (the
column of output headed P) should total to 1, but when you adjust for
age and sex it totals only to 0.5 (or the proportion for the sex being
tabulated) Here is an example:
use http://www.stata-press.com/data/r8/hbp
gen pop=1
dstdize hbp pop age_group sex, by(race sex) using (test_stan) print
[first part of output . . ]
-----------Standard Population-----------
Stratum Pop. Dist.
----------------------------------------
15 - 19 Female 11000 0.106
15 - 19 Male 10000 0.096
20 - 24 Female 16000 0.154
20 - 24 Male 15000 0.144
25 - 29 Female 16000 0.154
25 - 29 Male 15000 0.144
30 - 34 Female 11000 0.106
30 - 34 Male 10000 0.096
----------------------------------------
Total: 104000
(7 observations excluded due to missing values)
-------------------------------------------------------------------
-> race sex= 1 Female
-----Unadjusted----- Std.
Pop. Stratum Pop.
Stratum Pop. Cases Dist. Rate[s] Dst[P] s*P
-------------------------------------------------------------------
15 - 19 Female 14 0 0.233 0.0000 0.106 0.0000
20 - 24 Female 13 0 0.217 0.0000 0.154 0.0000
25 - 29 Female 19 1 0.317 0.0526 0.154 0.0081
30 - 34 Female 14 0 0.233 0.0000 0.106 0.0000
-------------------------------------------------------------------
Totals: 60 1 Adjusted Cases: 0.5
Crude Rate: 0.0167
Adjusted Rate: 0.0081
95% Conf. Interval: [0.0000, 0.0235]
The proportion of females in the 25-29 age group in the reference
population is 0.30 (not 0.154) and the adjusted number of cases should
be 0.95 (not 0.5) and the adjusted rate 0.0158, not 0.0081. Is this a
bug or my mistake?
Dr Colin Fischbacher
Edinburgh
*
* 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/
