Stats Final Exam

STAT 3504 Final Review Problems II

1.

a) b) c) d) e)

What type of study would use to investigate the effect of 2 factors on a response? What are the advantages of such a study? What are the disadvantages? How do randomized block designs increase the precision in an experiment? Explain the difference between a CRD and a RBD. When is a RBD more advantageous? Why? When is a CRD more advantageous? Why? Why is replication used in an experiment? Explain how a GRBD differs from a RBD. What is the advantage of using a GRBD instead of an RBD? What is meant by an "important interaction"? Give two circumstances under which you might use a transformation of the response variable in an analysis of variance. Why do you have to be able to assume no block-treatment interaction in a RBD?

2.

a) b) c)

3.

An experiment was conducted to study the effect of machine type (A or B) and cardboard grade (Y or Z) on the number of defective cardboard boxes produced in an hour. The number of defectives produced was counted during 3 randomly sampled hours for each combination of machine type and cardboard grade. The following data and partial SAS output were obtained. Cardboard Grade Machine Type A B Y 9 10 12 3 5 7 Z 8 11 12 4 5 5

Analysis of Variance Procedure Dependent Variable: DEFECTS Source Model Error Corrected Total R-Square 0.805166 Source MACHINE GRADE MACHINE*GRADE DF 112.9967 C.V. 21.86786 Anova SS Mean Square F Value 33.00 0.03 0.03 Root MSE DEFECTS Mean 7.5833333 Pr > F DF Sum of Squares 90.9167 Mean Square F Value 11.02 Pr > F

a) b) c)

Complete the ANOVA table and include the EMS's. Carry out all relevant ANOVA F-tests with α = .05. Use the Kimball inequality to put an upper bound on the family significance level for the set of tests.

d)

Carry out all pairwise comparisons for the RELEVANT factor effects, using the most efficient method.

4.

Three dye formulas for a certain synthetic fiber are under consideration by a textile manufacturer...