Question

1. Use the data on serum T3 concentrations from the experiments with chickens in Exercise 2.3.

Treatment	Serum T3, (ng/dl) x 10(-1)
Premolt	90.09	90.45	99.38	73.56	74.39
Fasting	98.81	103.55	115.23	129.06	117.61
60g bran	197.18	207.31	177.5	226.05	222.74
80 g bran	102.93	117.51	119.92	112.01	101.1
Laying mash	83.14	89.59	87.76	96.43	82.94

Contrasts of interest were the serum T3 concentration differences between successive stages: (1) premolt versus fasting, (2) fasting versus 60 grams of bran, (3) 60 grams of bran versus 80 grams of bran, and (4) 80 grams of bran versus laying mash.

1. Estimate each of the contrasts and their standard errors.
2. Test the null hypothesis for one of the contrasts with the F test.
3. Suppose you were to test the four contrasts each with a comparisonwise error rate of .05. Compute the maximum experimentwise error rate for this family of four tests.

Answer 1

Give data a) The linecơ model is aivonaj ไม่,2, 3,4,5 h]heYe ץ'' i's yield from ith -treatment Mo General mean Ppect ๙.. ERpect due to it, t eatment assum tion Data is and em-815 Y DTMALL y distributed b)varianes are Equau e Hypothesis test Hn : au tveatment means are equa HI At least one treatment ma s di fPeven G01 25D S4 8912 Su 3quares due treatpments MSST: 6263.7243 mean Sum o P squares dueto treatmets HSS :108. 32 il But He a n sum 아 Squares P-0 000066 0D b Since p Conctude thut these is sinf Picant diPPervence bleo the treatments. That at least one-treatome dinkeremt mean is