Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: multicollinearity VIF vs condition number


From   "Yujin Jeong" <yujin@gwmail.gwu.edu>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: multicollinearity VIF vs condition number
Date   Thu, 8 Jul 2010 09:38:28 -0400

I would appreciate to learn any potential reasons of why condition number
indicates serious multicollinearity (over 30), while individual and mean
VIFs indicate no evidence.  

Thank you, 
Yujin

		SQRT				Eigenval	Cond.Index
Var	VIF	VIF	Tolerance	1	5.618		1.000
v1	1.04	1.02	0.965		2	1.701		1.818
v2	1.21	1.10	0.828		3	1.627		1.858
v3	1.15	1.07	0.871		4	1.136		2.224
v4	1.48	1.22	0.675		5	1.068		2.294
v5	1.80	1.34	0.556		6	1.009		2.360
v6	7.02	2.65	0.142		7	0.907		2.489
v7	2.19	1.48	0.457		8	0.612		3.031
v8	1.53	1.24	0.652		9	0.584		3.103
v9	1.70	1.30	0.589		10	0.538		3.231
v10	1.83	1.35	0.548		11	0.377		3.861
v11	4.29	2.07	0.233		12	0.329		4.136
v12	4.50	2.12	0.222		13	0.260		4.650
v13	1.62	1.27	0.616		14	0.198		5.332
v14	1.61	1.27	0.623		15	0.033		13.039
v15	1.29	1.14	0.775		16	0.006		31.578
Mean VIF	2.28			Condition Number	31.578

Attachment: oledata.mso
Description: Binary data



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index