[18955] in s-news-athena
[S] Kolmogorov-Smirnov p-value
daemon@ATHENA.MIT.EDU (Walker, Randal {USA~Palo Alto})
Tue Aug 17 17:49:11 1999
Content-Return: allowed
Date: Tue, 17 Aug 1999 14:42:42 -0700
From: "Walker, Randal {USA~Palo Alto}" <RANDAL.WALKER@ROCHE.COM>
To: "'s-news@wubios.wustl.edu'" <s-news@wubios.wustl.edu>
Message-Id: <B9E51B9719BDD111AB640000F80197CD0146ED05@rplmsem2.pal.roche.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
The one-sample Kolmogorov-Smirnov Test on pg. 83 of the S-PLUS 4 manual
lists a p-value of 0.0914. Can anyone identify how this was derived?
Based on the "Numerical Recipies in C" (Press, Veterring, Teukolsky and
Flannnery) Second Edition, pg. 627, I am using Monte Carlo simulation to
derive the cdf of the test statistic using random N(0,1) deviates. The
answer I obtain from such an approach is close to the approximation given by
14.3.9 (pg. 624 of the NRC text). However, neither of these derived p-values
is anywhere near the published value above.
Thanks.
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
