Defining Variables for Attribute Charts - Quick Tab

Select the Quick tab of the Defining Variables for Attribute Charts dialog box to access the options described here.

Input data are counts or raw data (with piece info)

The settings of these option buttons determine whether Counts, Proportions or rates, or  Raw data are being used. For example, if the input data specify defect proportions (0<=p<=1), select the Proportions or rates option button. When creating an Np chart, the proportions will then be multiplied by the respective sample sizes. Note that for both options - Counts or Proportions - Statistica assumes that each observation in the input data represents one sample, and the counts or proportions specified for the sample will be used to construct the respective chart. The specification and interpretation of data when the Raw data format is specified does not make this assumption and is more flexible (see next paragraph).

Raw data reading for attribute charts (database format)

 When the Raw data format is specified, Statistica will count the number of pieces (computer chips identified in a piece identifier variable) with defects greater than 0 (computer chips with one or more defects as specified in a defect indicator variable); samples of pieces (batches of computer chips) are identified in a sample identifier variable. Note that in this method of data reading, the unit of analysis is the individual piece and not the number of defects: Each piece is counted as being defective only once if the defect indicator variable shows a value greater than 0 (and regardless of the actual value greater than 0). For example, a computer chip (identified with a unique piece ID) that has 100 defects is counted as a single defective observation, in the same way as a computer chip would be counted with a single defect. Hence, this method of reading and tabulating data is very flexible, and particularly well suited for processing automated data streams.

Variables

Click the Variables button to display a standard variable selection dialog box. If the Counts or Proportions or rates option button is selected, you can select Counts or proportions, Sample sizes (optional for U, P, and Np charts; not applicable to C charts) and Part identifiers (optional for short run charts). Note that the Sample sizes setting must be greater than or equal to the maximum value in the Counts variable. If the Raw data option button is selected, you can select Defect indicator, Piece ID's, Sample ID's, and Part identifiers (optional for short run charts).

Constant sample size

Select this check box to specify a fixed sample size (N) for all samples; this option is not applicable to C charts.

Constant number of samples per part

 If this check box is selected and no part identifier variable is selected, you can specify here a fixed number of consecutive samples for each part. For example, if your input data consist of batches with different specifications (different parts) from which a fixed number of samples is automatically drawn, then you can select this option so that on reading the data each chunk of k consecutive samples is automatically assigned to the same part.

Minimum number of observations per sample

 For U, P, and Np charts, and when reading Counts, Proportions or rates, or  Raw data, you can specify a minimum n per sample by typing it in or by adjusting the number with the microscrolls. Samples with fewer observations than the minimum specified will be ignored.