inaccuracy of SPSS logrank statistics?

From: Markus Scholz (markus.scholz_at_imise.uni-leipzig.de)
Date: 11/25/04 

Date: Thu, 25 Nov 2004 14:03:03 +0000 (UTC)

I found an example of logrank-test in literature, calculated it again
by hand and by a self-written computer programme. All three results
were exactly the same.
After that I calculated the test with SPSS logrank option and obtained
a slightly deviation from the true value (second decimal). But there
are other examples with deviations already at the first decimal. Is it
possible that SPSS calculates with (heavily) rounded values? What
other possibilities can be true?

Sincerely

Markus