This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Stat 500 – Homework 12 answers – 2009 Notes are marked by • 1. Bottling machines (a) Yes, there are no blocks in the design. However, this is not a reasonable analysis. Days are subsamples; there are only 12 e.u.’s. A more appropriate analysis would be to either: average the subsamples to get one average per machine, then use ANOVA or include the e.u., machine(model), as a random effect in the ANOVA. (b) My SAS code, for the subsampling analysis: data bottle; infile ’bottle.txt’ firstobs=2 missover; input model machine @; do day = 1 to 5; input y @; output; end; proc mixed method=type3; class model machine; model y = model; random machine(model); run; I got: estimated difference (model 1 - 2): -9.8 s.e.: 5.0 p-value: 0.084 2. Barley fungus experiment design All four studies have the same treatment structure (2 way factorial) but differ in their experimental design. (a) This is a CRD. Source d.f. barley 2 fungus 1 b*f 2 error 12 total 17 use MSE to test each effect 1 (b) This is an RCB, with repetition of the study being the block....
View Full Document
- Fall '08
- e.u., main plot error, split plot error, proc mixed method=type3