Question 1)
In an experimental program to investigate the properties of a thin film plastic coating on ceramic-based resistors, six factors were identified.
X1 - Supplier of a basic component of the plastic (there were 2)
X2 - Viscosity of the coating at time of application
X3 - Thickness of coating
X4 - Temperature of the first bake (dry) cycle
X5 - Temperature of the second bake (cure) cycle
X6 - Speed of the conveyer
(a) List the experiments and a random list of run orders for a full factorial design with these six factors using symbolic - and + signs for the levels. Show the complete confounding table.
(b) List the experiments and random run orders if we are to run a ½ fractional design.
(c) Describe how you would analyse the data in each of the situations defined in (a) and (b) above.
Question 2)
You are trying to help an ice cream vendor to be able to predict how many ice cream cones he will sell in a given afternoon (so that he knows how much ice cream to buy). He decides, with your help, to run a half fraction of a 25 to study the influence of several factors on sales. The data are given on the next page.
The factors studied and their levels are:
From the data given, is a prediction equation justified? If so, what is the equation? Justify your answers.
Question 3)
Consider the 25b2 fractional factorial design with generators: 4 = 12 and 5 = 123. Specify the factor level combinations for this design and determine the complete confounding pattern.