30.3 Steering Wheel – 2
In a later study on the hardness of the steering wheels, it was discovered that the previous experiment was not properly randomized. Initially, the first 8 replicates were done, and then the other 8, with the aggravating circumstance that two weeks passed between both sets of experiments; it is known that the environmental conditions (temperature, humidity, …) affect the characteristics of the components of polyurethane.
Because of the circumstances under which the data were obtained, it was decided to re-analyze the experiment considering it a 24 design where a new factor W (level -1 for the 8 experiments of the first replicate, and level + for the 8 of the second one) represents the differences (environmental or of other type) occurred during the two weeks passed between the first and the second experiment.
It would be interesting to draw conclusions considering this new point of view.
Stat > DOE > Factorial > Create Factorial Design
We have 4 factors; choose the complete design without replicates, and disable the option to randomize the order of the runs. Since we do not have replicates now, we activate the options of drawing a probability plot and a Pareto diagram of the effects to analyze its significance.
Stat > DOE > Factorial > Analyze Factorial Design > Graphs > Effects Plots: Normal
It is clear from both graphs that the significant effects ...