Home 
1) There are many advantages of looking at the three factors together rather than one at a time. The cost of conducting a factorial study is lower than studying one variable at a time. Furthermore the factorial design requires fewer samples, thus the amount of labour and process downtime is kept low. In addition the factorial method gives us information on the interaction between different variables, whereas looking at the factors independently will not. For example, the factorial method will tell us if rotors made on different machines have poor cores positioning, but looking at the factors separately will not show any relation. 2) When selecting the two levels for each factor, we should look at our problem, and all the possible explanatory variates that might cause an unwanted response. The levels must be probable, reasonable values for the factors to take. They should also be values that could possibly yield a relationship with the response. 3) The tests should be run randomly. This will minimize any timedependent factors in the study. Since the time variability of the factor is unknown, and is not one of the selected factors of study, we can minimise its effects by randomising the timing of the studies.
From the interaction graphs, it can be seen that position has very little effect upon the imbalance weight of the brake rotor. Both thickness and the operating machine shows a large effects upon the imbalance weight. Larger thickness greatly increase the imbalance weight, and Machine B is also the source for large imbalance weight. There is a moderate interaction between thickness and machine type with respect to causing imbalance weight.

