[R] Wilcoxon paired signed rank test and continuity correction

[michael watson (IAH-C)]

Hi

I am applying the wilcoxon paired signed rank test to many sub-groups of
a data frame using by() and wilcox.test, and sometimes the test uses
continuity correction and sometimes it doesn't.  What I want to know is:

1) what is continuity correction?  The docs say:

 correct: a logical indicating whether to apply continuity correction
          in the normal approximation for the p-value.

But come on, I'm a biologist....

2) what makes wilcox.test sometimes decide to implement continuity
correction and sometimes not?

3) should I be letting wilcox.test use continuity correction when it
wants to?

I know I can turn this on/off using the correct and exact options, but I
want to know what is best to do.  Presently I have it turned off because
as I am doing ~200 tests, I want the tests to be consistent (ie none of
them use continuity correction rather than some of them do, some of them
don't)

Many thanks in advance for your replies :-)

Mick

Michael Watson
Head of Informatics
Institute for Animal Health,
Compton Laboratory,
Compton,
Newbury,
Berkshire RG20 7NN
UK

Phone : +44 (0)1635 578411 ext. 2535
Mobile: +44 (0)7990 827831
E-mail: michael.watson at bbsrc.ac.uk

"To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to
say what the experiment died of." R.A. Fisher, 1938.