Meaning
Distribution measure for assessing the consistency of a manufacturing process within a discrete temporal window. Within subgroup standard deviation represents the inherent noise that remains when external variables are held constant. This value is used to determine if the variation is due to common causes.
Short Term Variability
Samples are taken in close proximity to minimize the influence of external factors like temperature shifts or operator changes. The within subgroup standard deviation is used to build the control limits for an r-chart or an s-chart. If this value is high, it suggests that the process is fundamentally unstable or that the measurement system itself is unreliable.
A high level of noise within a single batch indicates that the equipment or the environment is not properly controlled.
Pooled Variance
Multiple subgroups are combined to create a more accurate estimate of the process noise. The within subgroup standard deviation is calculated as the average of the individual group deviations or through a pooled variance formula. This combined figure is the basis for calculating the cp and cpk indices, which quantify how well the process fits within its specification limits.
A reliable estimate requires at least twenty subgroups to be statistically valid.
Root Cause Separation
Comparison with the overall variation helps identify the source of quality problems. When the within subgroup standard deviation is much smaller than the total standard deviation, the problem is likely a shift in the process mean over time. Addressing these shifts requires a different strategy than reducing the inherent noise within the process itself.
Precise calculation of this value is required for any meaningful process improvement.
