BIOL 373 Lecture Notes - Lecture 15: Biostatistics, Analysis Of Variance, Complement Factor B
Document Summary
Ssb=a ( x i x )2 ( x j x )2 i=1 b j=1. N 1 = ab 1 a 1. 1 (a 1)(b 1: total, factor a, factor b, remainder, interaction cannot be tested, as mserror is not separate from msinteraction, if (theoretically) interaction is not present: Remainder: if an interaction is actually present: *** increased chance of type ii error using msremainder to estimate. Mserror will include additional variation owning to the interaction. F ratio will be lower than it should be. Two other anova designs use the same procedures: Block: a group of related subjects: ex: siblings, animals reared in the same cage, plants grown in the same. Subjects are not independent within blocks, but randomly assigned to plot treatments. Of little interest & no independent estimate of mserror. Testing h0: 1 = 2 = 3: can also be tested with a single-factor anova; Group of subjects for each factor option.
