Flow Chart Analysis of Non-Normal Data

Flowchart

When the retrieval setting for Force distribution is set to Automatic (the default), GainSeeker follows these steps to determine the data distribution and method for statistical calculations:

Uniform distribution and Non-Normal distribution (modeled)

• d(U) is the point above the median where probability X > d(U) = 0.00135

• d(L) is the point below the median where probability X < d(L) = 0.00135

• Because 6 SD should represent 99.73% of the data spread and 8 SD should represent 99.994% of the data spread, GainSeeker will calculate these related statistics as follows:

Mean + 3SD  is the point on the x-axis where 0.135% of the data is above the value

Mean − 3SD  is the point on the x-axis where 0.135% of the data is below the value

Mean + 4SD  is the point on the x-axis where 0.003% of the data is above the value

Mean − 4SD  is the point on the x-axis where 0.003% of the data is below the value

Six SD  = (Mean + 3SD) − (Mean − 3SD)

• GainSeeker will calculate Cpk using either the Median method (the default) or the Probability method, based on the retrieval setting "Use the median method to calculate Z values instead of probability method".

Median method for calculating Cpk

The calculations for d(U) and d(L) are described above.

Probability method for calculating Cpk

Zu = z value (from the normal distribution table) corresponding to the predicted percent above the high spec

Zl = z value (from the normal distribution table) corresponding to the predicted percent below the low spec

Non-Normal (not modeled)

• The retrieval setting "Method for calculating standard deviation" determines which SD value is used to calculate the following statistics.

Mean + 3SD = Mean + (3 ✕ SD)

Mean − 3SD = Mean − (3 ✕ SD)

Mean + 4SD = Mean + (4 ✕ SD)

Mean − 4SD = Mean − (4 ✕ SD)

Six SD = 6 ✕ SD