Date: Tue, 17 Jan 2012 11:18:07 -0800
Reply-To: Fareeza Khurshed <fkhurshed@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Fareeza Khurshed <fkhurshed@GMAIL.COM>
Subject: Re: What statistical method/proc would work for this "trend"?
Content-Type: text/plain; charset=ISO-8859-1
Are you looking for a 'trend' or just a difference, from the hypothesis it
sounds like a difference.
What I'd do, which may or may not be correct, is define my time periods
over the 7 day week into some 6 or 8 hour intervals that make sense ie
Period 1 - Sunday 8:00 am to 3:59PM, Period 2=Sunday 4:00PM-11:59PM Period
3=Monday 12:00AM to 7:59AM etc using the date and time of my scheduled
Then I'd just run a proc freq with a chisquare or fisher test to see if
there was a difference.
At least, that's where I'd start off.
On Tue, Jan 17, 2012 at 11:07 AM, Joe Matise <firstname.lastname@example.org> wrote:
> Don't know about the stat parts, but there is a weekday format to
> determine Fridays (then check time for night according to your rules).
> Sent from my iPhone
> On Jan 17, 2012, at 11:51 AM, Irin later <irinfigvam@YAHOO.COM> wrote:
> > I am working on analysis based on hypothesis that certain medical
> surgery is mostly scheduled
> > on fridays nights rather than on others week days.
> > The extract for year 2011 was 1300 unique medical claims for the
> certain surgery.
> > Other years extracts are comparable in terms of # of claims for years
> 2008, 2009, 2010.
> > Not having an experience , I am really confused....
> > How can I do it? Could you please give me a hand?
> > I actually wonder:
> > 1. How to extract fridays nights?
> > 2. Do I need randomly select a month for the certain year?
> > 3. what would be an appropriate kind of statistical method and procedure
> for this scenario?
> > Thank you in advance,
> > Irin