Process Report - Options
main topic     see also  

Six Sigma > Process Report > Options

Unbiasing constants

Specify whether to use unbiasing constants or not when estimating short-term and long-term standard deviations. Unbiasing constants vary with sample size and are equal to the ratio of the expected value of the standard deviation estimate for samples of a given sample size to the standard deviation of the population. The default is to use unbiasing constants when estimating short-term standard deviation, but not when estimating long-term standard deviation. See Calculations for more information.

Calculations when subgroup size = 1

Specify how to calculate short-term Z when the data are not collected in rational subgroups (subgroup size = 1). Choose to:

·    Calculate Z.Bench ST using the moving range technique to compute the short-term standard deviation. Because the individual subgroup size is 1, you cannot calculate the short-term standard deviation in the usual way.

·    Calculate Z.Bench ST using a long-term Z standard deviation estimate. Use when you can't assume that units produced consecutively are less variable than units produced nonconsecutively.

See Calculations for more information.

Box-Cox transformations

Use the Box-Cox power transformation when your data are very skewed or the within-subgroup variations are unstable. The transformation takes the original data to the power l, unless l = 0, in which case it takes the natural log. Minitab determines the optimal l, or you can choose any value between -5 and 5 if the data are positive.

Dialog box items

Use unbiasing constants when estimating

Short-term standard deviation: Check for Minitab to calculate short-term process standard deviation using unbiasing constants.

Long-term standard deviation: Check for Minitab to calculate long-term process standard deviation using unbiasing constants.

For subgroup size = 1, calculate short-term Z using

Moving range to estimate short-term standard deviation: Choose to use a short-term estimate of process standard deviation.

Long-term Z + a sigma shift of: Choose to use a long-term estimate of process standard deviation (default). You can specify the assumed sigma shift, a value between 0 and 6.

Use Box-Cox power transformation (W = Y**Lambda) with: Check to perform a Box-Cox transformation.

Optimal lambda: Choose to have Minitab estimate an optimal value of lambda (default).

Other [a value between -5 and 5]: Choose to specify a value for lambda. Enter a value between -5 and 5.