Date:         Fri, 23 Sep 2005 10:12:47 -0300
Reply-To:     Hector Maletta <hmaletta@fibertel.com.ar>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Hector Maletta <hmaletta@fibertel.com.ar>
Subject:      Re: standardising weights in Poisson models?
Comments: To: kylie.lange@flinders.edu.au
In-Reply-To:  <4333A3A1.9020105@flinders.edu.au>
Content-Type: text/plain; charset="iso-8859-1"

Kylie,
Standardising the age distribution to conform to the national demographic
data is a common practice. There may be some objections to this practice in
general, but if the practice is accepted in general I do not see any reason
why you shouldn't use weights specifically in your Poisson model. You simply
incorporate a variable reflecting those weights and use it as your weighting
variable. For this the general procedure is as follows:

1. Create a variable with the proportion of population in the respective age
group of each case.
Example:
IF (AGE < 15)POP=0.23.
IF (AGE GE 15 AND AGE LE 24)POP=0.14.
.....................[CONTINUE FOR OTHER GROUPS; VALUES SUGGESTED ARE ONLY
EXAMPLES, USE ACTUAL VALUES FROM CENSUS.]

2. Create a variable with the proportion of cases in each age group WITHIN
YOUR SAMPLE:
IF (AGE < 15)sample=0.32.
IF (AGE GE 15 AND AGE LE 24)sample=0.18.
.....................[CONTINUE FOR OTHER GROUPS; VALUES SUGGESTED ARE ONLY
EXAMPLES, USE ACTUAL VALUES FROM YOUR SAMPLE].

3. COMPUTE THE RATIO OF THE ABOVE TWO VARIABLES:

COMPUTE AGEWGT=POP/SAMPLE.

4. Adopt this variable for weighting:

WEIGHT BY AGEWGT.

5. PROCEED WITH YOUR POISSON MODEL.

Hector

> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]
> On Behalf Of Kylie Lange
> Sent: Friday, September 23, 2005 3:42 AM
> To: SPSSX-L@LISTSERV.UGA.EDU
> Subject: standardising weights in Poisson models?
>
> Hi all,
>
> I have had a colleague explain the following problem to me. I
> hope it's clear and that someone can offer some suggestions.
> If anyone has any questions then I'll do my best to find out
> more info and clarify.
>
> They are looking at incidence rates of certain kinds of
> injury, broken down by age categories. They obviously have
> data on the number of injuries (#inj) and the population size
> (#popn) for each age group, for each of 9 years. In addition
> they have 'age standardised population weights' (wght). These
> weights account for the fact that the age distribution in the
> population changes over time (and I believe are based on data
> from the last national census).
>
> The usual way they use this information is to calculate #inj
> / #popn * wght for each age-by-year combination. These
> weighted rates are then all summed together to give an
> overall incidence rate.
>
> They now would like to use this data in a Poisson regression
> model so they can look at the rates over time and investigate
> possible trends that may be present. The question is how to
> incorporate the age-standardised weights into the Poisson model?
>
> I'm afraid I don't know enough about how the weights are
> calculated to know what is appropriate so I'm hoping that
> this will sound familiar to someone who deals with this kind
> of data. I also may be using confusing terminology but am
> happy to try clarify anything if needed.
>
> Thanks for any suggestions or references to articles where
> this kind of analysis has been carried out.
>
> Kylie.
>
> __________ Información de NOD32 1.1230 (20050922) __________
>
> Este mensaje ha sido analizado con  NOD32 Antivirus System
> http://www.nod32.com
>
>