Date: Thu, 5 Aug 1999 17:08:20 +0200
Reply-To: Hanna Zaremba <h.zaremba@CBOS.PL>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Hanna Zaremba <h.zaremba@CBOS.PL>
Subject: discriminant analisis-assumption
Content-Type: multipart/alternative;
Dear list membrs,
I was wondering if someone could answer my questin.
I perform discriminant analisis first time and i am not sure
if my data met discriminant assumptin.
How to interpret Box'M test?
Does 59.89% of "grouped" cases correctly classified: is enought or not?
I obtain such results:
Test of equality of group covariance matrices using Box's M
The ranks and natural logarithms of determinants printed are those
of the group covariance matrices.
Group Label Rank Log Determinant
1 10 12.390219
5 10 8.829367
8 10 10.435928
11 10 9.483442
Pooled Within-Groups
Covariance Matrix 10 11.639970
Box's M Approximate F Degrees of freedom Significance
386.77 2.2096 165, 98413.3 .0000
Classification Results -
No. of Predicted Group Membership
Actual Group Cases 1 5 8 11
-------------------- ------ -------- -------- -------- --------
Group 1 136 87 1 32 16
64.0% 1.1% 23.4% 11.5%
Group 5 48 12 4 31 1
25.6% 7.6% 64.1% 2.6%
Group 8 188 28 1 156 4
14.7% .4% 82.7% 2.1%
Group 11 61 22 0 26 13
35.6% .0% 42.6% 21.9%
Percent of "grouped" cases correctly classified: 59.89%
Thanks,
Hanna
Hanna Zaremba, CBOS
[text/html]
