Date: Tue, 29 Aug 2006 15:14:28 -0400
Reply-To: Wensui Liu <liuwensui@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Wensui Liu <liuwensui@GMAIL.COM>
Subject: Re: Macro for regression outputs and scoring
In-Reply-To: <44F456F5.B875.00C9.0@ndri.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Peter,
Though the macro like this should be very simple, I am wondering if a
macro is an over-kill here. Could you just just a BY statement in proc
reg and a SCORE statement to impute the missings.
On 8/29/06, Peter Flom <Flom@ndri.org> wrote:
> A colleague and I are working on the following, but it seems like a
> perfect question for SAS L
>
> We are estimating the number of drug injectors in each of 95 cities,
> for each of 11 years. We have 4 data sources. There is some missing
> data for each source.
>
> We've decided to impute the missing data using year by year regression
> equations based on the other cities. So, for instance, say that data
> source A is missing for Pittsburgh in 1999. We would then run a
> regression along the lines of
>
> proc reg;
> model A = B C D;
> where year = 1999;
> run;
>
> and use the parameter estimates to impute a value for A for
> Pittsburgh.
> (If people want more detail of why we are doing this, I'll provide it)
>
> If there were only a couple missing points, but we have a lot.
>
> So....I thought.....time for a macro.
>
> But how?
>
> My macro experience is pretty minimal, this is way beyond me
>
> All help, as always, is appreciated
>
> Peter
>
> (and, if you remember, please reply all so that my colleague gets these
> too)
>
--
WenSui Liu
(http://spaces.msn.com/statcompute/blog)
Senior Decision Support Analyst
Health Policy and Clinical Effectiveness
Cincinnati Children Hospital Medical Center
| 
