Date: Thu, 20 Mar 2003 08:37:47 -0500
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: "Arthur J. Kendall" <Art@DrKendall.org>
Organization: Social Research Consultants
Subject: Re: I need help!!!
Content-Type: text/plain; charset=us-ascii; format=flowed
If you look at the formula for the two sample t-test, you'll see that
that it has the two variances (SD squared), one for each group in the
denominator. The one sample t-test against a constant is the two-sample
t-test with the error term for the second group resolving to zero.
a kludge: If you cannot get Exner's SD, use yours for the second mean
but turn it into a SE using 600 df.
You can use matrix input in ONEWAY with 2 groups to learn about this.
With one numerator df, F = t**2.
If you can get the SD from Exner, try this all 3 ways.
1) with Exner's SD.
2) with a zero for Exner's SD
3) with your sum of squares scaled as if you had 600 df.
The two means remain the same in each of the 3 runs.
Use one of the "post hocs" designed for unequal variances.
Hope this helps.
Social Research Consultants
University Park, MD USA
> I am working on my dissertation and am stuck on figuring out which
> statistical test to use. I'll try to explain what I am doing, I hope
> you can help me.
> I am hoping to compare a group of individual's test scores (N=43) (my
> sample) to a given norm group (Exner's norms on the Rorschach) (my
> control group). Exner has 600 subjects in his sample. I tried to do
> a one sample t-test on this, comparing my numbers to a "known number"
> (i.e., the mean that Exner got in his normative sample). However, my
> supervisor told me that he did not believe that this was the right
> statistical test since the groups should have different degrees of
> freedom. Please help me select a statistical test. I would really
> appreciate it. Thank you!!! Marlena