Statistics for Six Sigma Made Easy

13.2. WHAT WE HAVE LEARNED IN CHAPTER 13

1. Valid sampling and analysis of proportional data may be needed for all the steps in the DMAIC process.

2. Proportional data are generated from attribute inputs such as yes/no and go/no-go.

3. Testing for statistically significant change with proportional data involves comparing proportions stated as decimals.

4. Proportional data require much larger sample sizes than variables data.

5. Sample size is usually a compromise between cost and desire for accuracy.

6. We normally work to a 95% confidence level.

7. We generally want to be able to sense a change of 10% of the difference between the proportions or 10% of the tolerance.

8. When checking for a change, we can compare a sample with earlier population data or compare two samples ...