Date:         Wed, 18 Oct 2006 19:58:58 -0400
Reply-To:     Arthur Tabachneck <art297@NETSCAPE.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Arthur Tabachneck <art297@NETSCAPE.NET>
Subject:      Re: How can I do this type of analysis
Comments: To: sasncsu@GMAIL.COM

Xu,

I'm surprised that no one has responded yet.  While I'm sure that not
everyone would agree, I don't think that you can use statistics to do what
you want.

There are numerous tests to discover whether differences might likely be
due only to chance, but none (that I am aware of) to test equivalence
(i.e., no difference).

Art
---------
On Wed, 18 Oct 2006 18:13:21 -0400, Xu Yuan <sasncsu@GMAIL.COM> wrote:

>hello friends,
>
>I am a rookie in statistics, so please don't laught at me if my question
is
>silly.
>
>I have two columns of data (X and Y), which come from the same group of
>experiment units but using two different methods. My objective is to see
if
>the two methods are (statistically) equivalent. The ideal situation is
Y=X,
>but this world is never ideal. I tried to fit linear regression in MS
Excel
>and the default model is Y=a*X+b with a R2. My question is what technique
>should I use to achieve my goal (compare the two methods). Should I force
>the intercept to zero? Or should I specify the slope to 1? (if so, please
>tell me how). Please advise me. I can work in either MS Excel or SAS or
R. I
>am considering using paired t-test, but I am not if this is the right
>procedure.
>
>Thanks a lot.
>
>Xu