> -----Original Message-----
> From: A UK-based worldwide e-mail broadcast system mailing list
> [mailto:[log in to unmask]] On Behalf Of Bilal
> Sent: Wednesday, October 14, 2009 7:35 AM
> To: [log in to unmask]
> Subject: Large Correlation Coefficients and No Significance
>
> Hi All
>
> Is it possible to have large correlation coefficients (say .62) yet it
> might not be flagged as significant at 0.05 or 0.01 level by SPSS..
> What could be the possible reasons.
Low sample size!
A correlation of 0.62 with an N of 10 is n.s. p ~ 0.06 two-tail.
Likely caused by perhaps using casewise deletion of missing cases when
computing correlation coefficients within a matrix of variables
containing missing data spaced "differentially within each variable.
Regards .. Paul
-----------------------------------
Pbarrett.net: Advanced Projects R&D
-----------------------------------
W: www.pbarrett.net
E: [log in to unmask]
M: +64-(0)21-415625
|