At random, 20 different rater are assigned to each cookie.
Design contrasts to compare chewy vs crispy, and expensive with inexpensive.
Are the contrasts orthogonal? Show.
For the example in class with 3 levels: control, trt1 and trt2, we used contrasts w1=(1,-0.5,-0.5) and w2=(0,1,-1). Now assume the sample sizes were n1=10, n2=7, n3=8. Rescale w1 and/or w2 so that they are still orthogonal and still make the same relative comparisons, i.e. compare control to the two trts and compare trt1 to trt2.
Preview of Design-contrasts-to.docx