Capability Analysis (Poisson) - Quick Tab
Select the Quick tab in the Capability analysis (Poisson distribution) dialog box to access options to select the variables for the analysis, optionally select a fixed (sample, unit) size and/or historical average rate (mu), and to specify optionally a Target and USL (upper specification limit) value for the analysis.
Input data as
In general, all analyses are performed on Poisson rates (such as failures per unit, such as numbers of defects per wafer).
Therefore, there are two options for specifying (interpreting) count or defect data:
Counts (divide measures by sample sizes to compute rates)
Select this option button if the data to be analyzed are counts (such as numbers of defects; numbers of calls to a call center that got a busy signal, etc.).
In this case, to perform the analyses (for example, to compute u charts), Statistica will automatically divide the respective counts (found in the data) by the respective sample sizes.
Rates (do not divide measures by sample size)
Select this option button if the data to be analyzed are proportions (0 ≤ u). In this case, the respective values will not be divided by sample sizes (but will be multiplied by sample sizes).
Note that, regardless of the type of input file (counts or proportions), the Historical mu values (if specified) always are interpreted in terms of rates.
Variables
Click this button to display a standard variable selection dialog box. You must specify a variable with the measurements (numbers or rates of defects).
Optionally, you can also specify a variable with unit sizes. If no sample size variable is selected, the fixed value from the Sample size field (see option description below) will be taken as the unit size and applied to all samples.
Sample size (if constant)
If no variable with sample (unit) sizes is specified via the Variables option, the constant sample size specified in this field will be applied to all samples.
Historical mu
Select this check box, and specify the respective value [0<mu or average rate (mean DPU or defects per unit)] to use a historical estimate of the average proportion or rate.
This value will be used:
- As the center line in the summary u chart
- In the evaluation of Runs tests, if they are specified
- In the Summary statistics spreadsheet (created from the Results dialog box), as the mean DPU (defects per unit), and to compute the respective confidence intervals for DPU.
Enter a USL value (number of defects per unit)
- Enter the upper specification limits for the respective process.
- Enter the value as the number of defects acceptable or permitted (inclusive) per unit (for example, do not enter a rate, but rather the actual count, regardless of the setting of the
Input data as options).
The USL value is shown in the Capability Summary histogram, and is used to compute capability indices (see Computational Details).
See also Example 10: Capability Analysis (Binomial) and Process Analysis - Capability Analysis Binomial and Poisson.