Estimating Some Metrics in Six Sigma Through Confidence Intervals

Methodology/Approach: This approach assumes that the process follows a normal distribution with a constant variance. The mean of the process is shifted from the target value to the right or left by 1.5 standard deviations. The estimates are based on a random sample of size n taken during a time when the process is stable.


Findings: The paper describes how to create confidence intervals for the number of defects per unit, the probability that a unit will be free of defects, and the rolled throughput yield.


Research Limitation/implication: We assume a discrete process in which n units of the product are selected during a time when the process is stable.


Originality/Value of paper: By applying the proposed estimation procedures, process performance evaluations can be improved, facilitating decision-making for Six Sigma projects.