Class Notes for AB – Wednesday, 8^th February

· Correct typo in Ch. 3 p. 8, remove “Least Squares” from heading of Output 3.6a

· Note: we assume the Y’s (after removal of the “model”) that has the normal distribution, not the Y’s themselves

· We never look for an interaction in a RCBD – assumed zero

· Study IBD’s on your own (pp. 13-15)

· COD’s today

· 3.5 next Monday

· 3.6 LOF test – ex. 3.14 from D&S, ex. 3.15: LOF from line in the direction of what? LOF test is a Full-and-Reduced F test

· ex. 3.16: how can we do a LOF test in the presence of unequal variances? We must extend the results, and use the -2DLL c² test referenced in Chap. 2; output on top of p.31 is wrong, but SAS program in Appendix 3.7.1 (pp. 33-4) is right

· same idea for section 3.6.2 – heterogeneous one-way ANOVA