[R] The meaning of 'z value' and 'Pr(>|z|)' in the result of glm

[Bill.Venables at csiro.au]

Stephen Lau asks:
> 
> I used the function 'glm' to perform logistic regression, and
> then use 'summary' to display the fitting result. There are 'z
> value' and 'Pr(>|z|)'

The z value is the Wald statistic for testing the hypothesis that the
corresponding parameter (regression coefficient) is zero.  Under the
null hypothesis it has an approximately N(0,1) distribution.  P(>|z|)
is the tail area in a 2-tail test, i.e. the test within a 2-sided
outer hypothesis.

> 
> Do they simpily represent the result of 'z test' ? 

Yes, in the sense above.

> as I perform the 'z test' for a particular variable following the
> definition, it yields a different result.

Have you found your mistake yet?

___

Bill Venables
CSIRO Laboratories
PO Box 120, Cleveland, 4163
AUSTRALIA
Office Phone (email preferred): +61 7 3826 7251
Fax (if absolutely necessary):  +61 7 3826 7304
Mobile:                         +61 4 8819 4402
Home Phone:                     +61 7 3286 7700
mailto:Bill.Venables at csiro.au
http://www.cmis.csiro.au/bill.venables/