Certified Six Sigma Black Belt Design of Experiments (DOE) Questions and Answers — Questions and Answers
Question 1: A project team is conducting an experiment to optimize a chemical process. To prevent unforeseen variables, such as changes in ambient temperature or raw material batch variations throughout the day, from systematically biasing the results, the team decides to run the experimental trials in a non-sequential, randomly generated order. This practice is known as:
- Replication
- Blocking
- Randomization (Correct answer)
- Confounding
Correct answer: Randomization
Randomization is the practice of running experimental trials in a random order. Its purpose is to average out the effects of 'lurking' or nuisance variables that are not controlled in the experiment, ensuring they do not systematically influence the results and lead to incorrect conclusions.
Question 2: In a fractional factorial experiment, the design resolution describes the degree to which effects are aliased with one another. If an experiment has a Resolution IV design, what does this imply?
- Main effects are aliased with other main effects.
- Main effects are aliased with two-factor interactions.
- No main effects are aliased with any two-factor interactions, but two-factor interactions are aliased with each other. (Correct answer)
- All main effects and two-factor interactions are aliased with three-factor interactions.
Correct answer: No main effects are aliased with any two-factor interactions, but two-factor interactions are aliased with each other.
A Resolution IV design is one where no main effects are confounded (aliased) with any two-factor interactions. However, two-factor interactions are aliased with other two-factor interactions. This is a significant improvement over a Resolution III design, where main effects are aliased with two-factor interactions.
Question 3: A Black Belt is analyzing the results of a 2-level, 3-factor (A, B, C) full factorial experiment. When reviewing the interaction plot for factors A and B, the lines representing the low and high levels of factor B are not parallel and cross each other. What is the most appropriate interpretation of this plot?
- There is a strong main effect for factor A, but not for factor B.
- Neither factor A nor factor B has a significant effect on the response.
- There is likely a significant interaction between factors A and B. (Correct answer)
- The experimental data contains a significant number of outliers.
Correct answer: There is likely a significant interaction between factors A and B.
A key visual indicator in an interaction plot is the parallelism of the lines. When lines are parallel, it suggests no interaction. When they are not parallel, and especially when they cross, it indicates a strong interaction effect. This means the effect of factor A on the response depends on the level of factor B, and vice-versa.
Question 4: A research and development team needs to investigate the effects of 8 different factors on product yield. A full factorial experiment would require 2^8 = 256 runs, which is too expensive and time-consuming. The team's primary goal is to identify the few factors that have the most significant main effects, and they are willing to accept that some two-factor interactions might be confounded. Which type of experimental design is most suitable for this situation?
- Response Surface Methodology (RSM)
- A fractional factorial screening design (Correct answer)
- A full factorial design with blocking
- One-Factor-at-a-Time (OFAT) experiment
Correct answer: A fractional factorial screening design
A fractional factorial design is ideal for screening a large number of factors efficiently. It uses a fraction of the runs of a full factorial design, saving time and resources. Its main purpose is to identify the vital few factors (main effects) that have the largest impact on the response, making it the perfect choice for this scenario.
Question 5: Which of the following describes a situation where the effect of one factor is indistinguishable from the effect of another factor or interaction?
- Replication
- Confounding (Correct answer)
- Collinearity
- Interaction
Correct answer: Confounding
Confounding (or aliasing) is the term used in DOE when the effects of two or more factors or interactions cannot be separated from one another. This is a deliberate trade-off in fractional factorial designs to reduce the number of required experimental runs.
Question 6: After using a fractional factorial experiment to identify the two most critical factors (Temperature and Pressure) affecting a product's strength, a process engineer now wants to determine the precise settings for these two factors that will maximize the strength. Which of the following is the most appropriate next step?
- Conduct another screening experiment with more factors.
- Perform a Gage R&R study on the strength measurement system.
- Use Response Surface Methodology (RSM) to model and optimize the response. (Correct answer)
- Implement a control chart for Temperature and Pressure.
Correct answer: Use Response Surface Methodology (RSM) to model and optimize the response.
Response Surface Methodology (RSM) is an optimization technique used after the vital few factors have been identified. It involves running experiments (like a Central Composite Design) to create a mathematical model of the relationship between the factors and the response, allowing the team to find the optimal settings (the 'peak' of the response surface) that maximize or minimize the output.
Certified Six Sigma Black Belt Design of Experiments (DOE) Questions and Answers
A project team is conducting an experiment to optimize a chemical process.
To prevent unforeseen variables, such as changes in ambient temperature or raw material batch variations throughout the day, from systematically biasing the results, the team decides to run the experimental trials in a non-sequential, randomly generated order.
This practice is known as: