Errata for Second (Feb. 2002), Third (Dec. 2003), and Fourth (Dec. 2004) Printings of Regression Modeling Strategies

Pg 32 L3b	X1 should be X2
Pg 32 L1b	X2 should be X1
Pg 79 L6	degrees of should be degrees of freedom
Pg 88 L13	whole whole should be whole
Pg 127 Legend	chi-squared should be rho-squared
Pg 158 S code	transcan( ) should have data=prostate for R
Pg 161 S code	impute() should have data=prostate for R
Pg 172 L10	uncensored should be censored
Pg 193 L1-3	Delete these lines; last b needs I(b) updated
Pg 270 S code	impute( ) should have data=prostate[subset,] added for R
Pg 401 Eq 16.27	Should be less than or equal to t
	In line above Eq number at risk should use greater than or equal to t
Pg 402 Eq 16.29	Should be less than or equal to t in two places
Pg 403 Eq 16.31	Should be less than or equal to t
Pg 437 bottom	Add change dist to "loglogistic" for S-Plus 6 or R
Pg 473 Eq 19.13	Should be less than or equal to t
Pg 522 Problem 1	Strike omit that predictor from consideration for now and Decide whether ... should be retained.
Problem 3	Replace Test for and include ... stratification factor. with Include an interaction between two predictors of your choosing.
	Change t=100 days to t=730 days
Pg 530 Ref 52	Volume 75 should be 78

indicates change confirmed in master LaTeX file.

For the change on Pg 161 check that the transformed pf does not need to be stored in the prostate data frame.

Thanks to Eric Leifer (leifere@nhlbi.nih.gov) for pointing out the errors on Pg 401. Thanks to Bruce McCullough (bdm25@drexel.edu) for pointing out the error on Pg 193.

Thanks to Christoper Tong for pointing out the error in reference 52.