Suggest simpler is better? Especially as you seem to have a lot of
data, with few potential variables.
On those already retired, regression of age at retirement and years
service at retirement. Going to be terrible confounding of these two,
but... Maybe add position in company (numerical job grade if such is
used), skill level of job, some measure of authority (nah, skip that for
job grade:), and like that. If correlation works (valid), and
confounding is bearable, then use the regression model to predict
retirement dates for currently employed population.
If your company history includes retirement buy-out offers, this will be
a factor to include - perhaps as a dummy variable, 0 for retired with no
offer, 1 for retired using a public, general offer of any type.
I predict right now that the correlation will not be wonderful, and that
confounding of independent variables will be a real issue. Inasmuch as
any prediction is better than none, it may work. Just be very careful
about the confidence intervals of the predictions.
Jay
Alfred Barron wrote:
> Hi everyone,
>
> The event I'm interested in is an estimate of time
> to retirement (or attrition) of our staff.
>
> A brief subset of my data looks like,
>
> SUB AGE HIRED RETIRED ACTIVE
>
> 1 50 09/15/70 06/15/90 0
> 2 45 09/12/75 06/15/92 0
> 3 52 09/13/72 . 1
> 4 55 09/15/70 . 1
> 5 35 09/17/90 . 1
> ...etc...
>
> The variables are SUB (subject), AGE, HIRED (date
> hired), RETIRED (date retired or quit), and ACTIVE
> (0=retired, 1=employed). This appears to be fairly
> standard data.
>
> Could anyone suggest a working model in SAS ? I'm
> considering PROC LIFEREG with Allison's PREDICT
> macro (as described in his 1995 book). Maybe use
> age as a covariate...
>
> Any suggestions on this ? Is this in the published
> literature ? Survival analysis is new to me. For
> the curious, I'm trying to get a rough estimate of
> the number of people taking retirement this year
> for our budget planning. I'm on the local Board of
> Education.
>
> thanks,
> Al Barron
> Metuchen, NJ
>
>
>
>
> __________________________________________________
> Do You Yahoo!?
> Get email at your own domain with Yahoo! Mail.
> http://personal.mail.yahoo.com/
>
>
> =================================================================
> Instructions for joining and leaving this list and remarks about
> the problem of INAPPROPRIATE MESSAGES are available at
> http://jse.stat.ncsu.edu/
> =================================================================
>
>
>
--
Jay Warner
Principal Scientist
Warner Consulting, Inc.
4444 North Green Bay Road
Racine, WI 53404-1216
USA
Ph: (262) 634-9100
FAX: (262) 681-1133
email: [EMAIL PROTECTED]
web: http://www.a2q.com
The A2Q Method (tm) -- What do you want to improve today?
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
