FOCUS AREAS

SAS/QC Software

Summary of Capabilities

Basic Quality Improvement

The PARETO procedure creates the following types of charts:

The ISHIKAWA procedure provides an interactive graphics environment for creating and modifying Ishikawa diagrams with annotation and drill-down to subdiagrams.

More

Statistical Process Control

The SHEWHART procedure creates a comprehensive set of Shewhart charts, including control charts for the following:

You can request Western Electric rules (tests for special causes and runs tests) for Shewhart charts. You can also create specialized control charts, including multivariate control charts, and control charts for autocorrelated data. The SHEWHART procedure provides extensive facilities for enhancing control charts graphically, including adding box-and-whiskers plots and stars to means charts. Distinct sets of control limits can be displayed for phases in historical control charts. Control limits can be saved and reused with new measurements.

The CUSUM procedure creates cumulative sum control charts for means and individual measurements. Control schemes can be based on average run length or V-mask parameters.

The MACONTROL procedure creates moving average and exponentially weighted moving average charts for means and individual measurements. Control chart parameters can be based on average run length.

More

Process Capability Analysis

The CAPABILITY procedure provides the following:

More

Reliability Analysis

The RELIABILITY procedure provides the following methods for reliability data analysis, survival data analysis, and recurrence data analysis:

The RELIABILITY procedure also creates a variety of probability plots for lifetime data (including Weibull plots), plots for accelerated life test data, and nonparametric plots of the mean cumulative function and associated confidence intervals.

There is also an experimental graphical interface for the RELIABILITY procedure.

More

Design of Experiments

The FACTEX procedure constructs orthogonal factorial experimental designs. These designs can be either full or fractional factorial designs, and they can be with or without blocks. You can also construct designs for experiments with multiple stages, such as split-plot and split-lot designs.

The OPTEX procedure searches for optimal experimental designs. You specify a set of candidate design points and a linear model, and the procedure chooses points so that the terms in the model can be estimated as efficiently as possible.

The SAS ADX Interface for Design of Experiments guides you through the steps of creating a design, recording responses, analyzing the data, and generating a report. It provides two-level screening designs, response surface designs, optimal designs, mixture designs, mixed-level designs, Taguchi designs, split-plot designs, and general factorial designs. Interactive graphics include main effects plots, interaction plots, cube plots, scatter plots, factorial plots, Lenth plots, Bayes plots, prediction profile plots, contour plots, and surface plots. Analysis methods include regression analysis, analysis of variance, residual analysis, outlier and influential observation analysis, optimal Box-Cox transformations, canonical analysis, and a variety of methods for selecting active effects in saturated two-level designs.


More

Analysis of Means

The ANOM procedure creates analysis of means (ANOM) charts for identifying group or treatment means that differ signicantly from the overall mean. ANOM differs from analysis of variance (ANOVA) in that ANOVA only determines whether there is a significant difference in treatment effects. The ANOM procedure can be applied both to count data with a Poisson or binomial distribution, and also to continuous data with a normal distribution. Exact decision limits are computed for both equal and unequal sample sizes. The ANOM procedure provides extensive features for enhancing and modifying ANOM charts.

More

Multivariate Process Monitoring

The MVPMODEL and MVPMONITOR procedures (experimental in SAS/QC 9.3) are used together to monitor multivariate process variation over time in order to determine whether a process is stable or to detect and diagnose changes in a stable process. The MVPMODEL procedure builds statistical models, which the MVPMONITOR procedure then uses to create multivariate control charts.

More