How many observations are required when calculating the standard deviation in the quality control context described?

• A. 10
• B. 20
• C. 30
• D. 40

Answer: (B)

Estimating how much a process varies hinges on having enough data so that the spread you compute isn’t driven by random noise. The standard deviation is based on deviations from the mean and uses n−1 degrees of freedom, so small samples produce a highly unstable estimate. In quality control, a reliable estimate of sigma is essential because it underpins control limits and our ability to detect true shifts in the process.

With around twenty observations, you get a reasonable balance between the effort of data collection and the precision of the spread estimate. More observations tighten the confidence interval around sigma (via the chi-square distribution), giving more trustworthy control limits. Therefore twenty observations provide enough information to calculate a stable standard deviation for QC purposes.