Date:         Fri, 26 Sep 2003 14:53:02 -0700
Reply-To:     cassell.david@EPAMAIL.EPA.GOV
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "David L. Cassell" <cassell.david@EPAMAIL.EPA.GOV>
Subject:      Re: TIME Series Help
Content-type: text/plain; charset=US-ASCII

kende jan <kendejan@YAHOO.FR> wrote:
>  I have data relating to the temperatures and the deaths day per day
> and age group for one year. I will like to compare the rates of deaths
> for two particular periods. What can bring the methods of the time
> series to me?

> If somebody can give me indications in order to treat these data this
> would help me enormously.

Hmmm.  First, it is not clear to me that you are really analyzing the
data as a time series.  If you only want to compare two rates taken at
different times, then I would say that you do *not* have much of a time
series problem.  If you want to look at temporal trends for different
groups, then that looks much more like a time series problem to me.
either way, I think that you will want to consult with a statistician
near you and get his/her opinions after he/she has a chance to study
your data.

If you want to try to analyze the data yourself using time series
analysis, then I would recommend a book like:

"Applied time series analysis for the social sciences" by McCleary and
Hay

But be wanred that a lot of time series analysis is studying the data,
performing diagnostic plots, evaluating analyses, model-building and
model-testing, and more hands-on work which is not strictly the sort of
"plug in the numbers and come up with a magic formula" work that plenty
of
forecasting software packages try to tell you it is.

HTH,
David
--
David Cassell, CSC
Cassell.David@epa.gov
Senior computing specialist
mathematical statistician