The second printing in early 2004 includes corrections of everything here, so we have separated the list into recent and “fixed” errors”

For convenience, we’ve divided the errata into those that may cause incorrect analyses or computations, and those that are typographical errors in the text or bibliographic errors.

We have acknowledged the colleagues who drew our attention to these errors, and very much appreciate their efforts

Since second printing

None reported yet

In first printing

Location Problem
Equation 2.15 Should read  Courtesy of Carsten Dormann
Equation 3.4 We provide the separate variance version of the t-test. The pooled version is more appropriate at this stage, and this equation should be replaced by 
Section 7.2.1, Page 158, second column Penultimate sentence in top paragraph should read “Note that replication does not guarantee protection from confounding…”. Courtesy of Michael Moeykens.
p. 183 Table 8.1, m  has no subscript. Courtesty of Jean-Pierre Airoldi
Table 8.8, p. 203 Last contrast coefficient has the wrong sign; the correct set is +2 -1 -2 -1 +2. Courtesy of Staffan Karlsson.
Table 9.6 (p. 217) Expected mean squares for A and B(A) have an incorrect multiplier for the  term; it should be . The calculations of variance components are correct.
Box 10.5, page 279, at the end. SSnon-add = 0.026 with 1 df. Then we replace it with 0.025 in the next line – should still be 0.026. Then SSRemainder on next line (once 0.026 is used instead of 0.025) should equal 191.197. Rest of table is OK. Courtesy of Staffan Karlsson
Table 11.2 on page 314 last column [Test(A,C fixed, B random)], second line: 
Box 14.2 on page 383 Odds ratio should be ; in the book, the numerator is 25*6, although the odds ration given is correct. Courtesy of Ralph Mac Nally.
p. 403 left column, line 5, covariance can be obtained … Y’Y and not Y.Y-1. Courtesy of Jean-Pierre Airoldi