Process Capability Analysis

Creates process capability indices for grouped (multiple sample) or ungrouped data. Process performance and capability indices are computed for grouped data. You can also fit various specific non-normal distributions (e.g., Weibull, log-normal, Beta, Gamma, etc.) as well as general non-normal distributions by moments, and compute capability indices using the percentile method.

Process Specifications

Element Name Description
Detail of computed results reported Level of detail in the reported results. If minimal detail is requested, only the basic process capability statistics are reported. If Comprehensive detail is requested, Statistica also reports the number of observations beyond specifications, frequency distribution and normality tests, and normal probability plots. If All results is requested, Statistica also reports various descriptive statistics.
Type of specification limits Specifies which method (see below) you want to use to enter the specification range. There are four options: Nominal +/- delta; Lower, Nominal, Upper; Lower, Nominal (one-sided limits); and Nominal, Upper (one-sided limits).
Nominal value (for spec. limits) Nominal value, for process capability computations.
Delta value (for spec. limits) Value of delta, for defining specification limits as nominal plus/minus delta.
Lower specification limit Lower specification limit; not applicable if Type of specification limits is equal to Nominal Upper (one-sided).
Upper specification limit Upper specification limit; not applicable if Type of specification limits is equal to Lower Nominal (one-sided).
Sigma limits Specifies constant for sigma limits; these limits will determine the process range that is used for the computation of the process (machine) capability indices.
Lower threshold equals LSL When computing the potential capability indices Cp and its inverse Cr (capability ratio), it is often desirable to treat the natural lower and upper bounds of a given distribution (such as Weibull) as the limits for the engineering tolerances.

Grouping (Multiple Samples)

Element Name Description
Grouping Select No grouping when you want to disregard any sample information and estimate sigma from the overall standard deviation. Select the Constant sample size or Grouping variable option if you want to consider sample information in the computations (estimate sigma from the within-sample variability) and compute process capability as well as process performance indices.
Constant sample size Enter a value n for the constant sample size (if the Constant sample size option is selected). When processing the data, Statistica will assign each n consecutive observations to the same sample, and estimate the sigma from the within-sample variability.
Estimate sigma from Select a method for estimating sigma for process capability computations; this option is not applicable if No grouping is selected (in which case sigma will always be estimated from the standard deviation for the single sample).

Distribution Fitting

Element Name Description
Distribution Select the type of distribution to use for computing percentile-based capability indices.
Goodness of fit Creates and reports a spreadsheet containing the frequency distribution for the variable of interest, along with the fitted (expected) frequencies.
Distribution offset Some distributions have particular valid value ranges (e.g., for Weibull the observed values must be greater than zero). For those distributions, you can specify a lower threshold (location) value.
Distribution scale Specifies a scale parameter for the Beta distribution; the observed values are rescaled as: (x-Location)/Scale.
Quantile-quantile plot Displays the quantile-quantile plot.
Probability-probability plot Displays the probability-probability plot.
Fit all distributions Creates spreadsheet(s) (one for each selected variable) with the results of fitting all available distributions, including the general non-normal distribution.