You are on page 1of 4

FACT SHEET

SAS/STAT® Software

What does SAS/STAT® do?


From analysis of variance and linear regression to Bayesian inference and high-per-
formance modeling tools for massive data, SAS/STAT software provides tools for both
specialized and enterprisewide statistical needs.

Why is SAS/STAT® important?


Organizations of all types and sizes depend on statistical analysis to guide critical deci-
sions. Modern statistical methods provide trustworthy evidence for creating effective treat-
ments for disease, improving manufacturing processes, predicting customer behavior and
making policy decisions. SAS/STAT provides a comprehensive set of up-to-date tools that
can meet the data analysis needs of your entire organization.

For whom is SAS/STAT® designed?


It is designed for use by business analysts, statisticians, data scientists, researchers and
engineers.

Organizations in every field depend on data


Benefits • Use proven and validated methods.
and analysis to provide new insights, gain SAS has decades of experience devel-
• Apply the latest statistical techniques.
competitive advantage and make informed oping advanced statistical analysis
With an accelerated release schedule,
decisions. For many organizations, the software designed for superior, reliable
SAS/STAT keeps up with new methods
complexity and volume of their data has results. A rigorous software testing and
emerging from the rapidly expanding
outgrown the capabilities of other statistical quality assurance program means you
field of statistics.
software. can count on the quality of each release.
• Analyze any kind and size of data. With SAS/STAT, you can produce code
SAS/STAT software provides extensive statis- SAS/STAT includes exact techniques for that is documented and verified for
tical capabilities that scale to meet the small data sets, high-performance statis- corporate and governmental compli-
expanding requirements of specialized and tical modeling tools for large-data tasks ance requirements.
enterprisewide analysis. and modern methods for analyzing data
with missing values.
Whether you want to perform a customer • Readily understand results with a wealth
preference study, analyze the results of of graphs. SAS/STAT output provides
clinical trials, predict credit card usage hundreds of built-in, customizable
patterns, model air pollution patterns or graphs that are designed for a consistent
design a health behavior survey, SAS/STAT look across analyses.
is the software of choice. • Take advantage of our technical support
and web user communities. Backed by
And because SAS remains committed to its industry-leading statistical technical
long tradition of constantly enriching its support, SAS/STAT software is the
statistical offerings, you’ll continue to have complete answer to a broad spectrum
the most up-to-date statistical techniques. of statistical needs.
Overview
Statistics is a rapidly expanding discipline, and
SAS/STAT is keeping pace. An accelerated
release schedule delivers more state-of-the-
art statistical methods and high-performance
computational tools, along with user-
requested enhancements. SAS/STAT soft-
ware provides the foundation for many
SAS Analytic offerings. It delivers a complete,
comprehensive set of statistical tools that
can meet the data analysis requirements of
your entire organization.

Expansive library of ready-to-use


statistical procedures
With SAS/STAT, you get more than 90
prewritten procedures for statistical analysis.
These procedures encapsulate and deliver
significant functionality that can be executed
with just a few simple commands, enabling
programmers to be more efficient and
Figure 1: Plot of component densities for a finite mixture model created with the
productive. This wide range of robust statis-
FMM procedure.
tical methods can help you solve your most
complicated business and scientific
problems, such as uncovering new informa-
tion for improving processes, driving devel-
opment and revenues and retaining valued
and satisfied customers.

Highly interpretable statistical output


Clarity and consistency of statistical output,
including a wealth of built-in graphs, enable
users to readily understand analysis results.

Comprehensive documentation
and training
Extensive online documentation, including
a rich set of introductory examples, allows
users to get up and running with the software
quickly and effectively. Free “how to” videos,
tutorials and demos help you build your
knowledge of statistical methods and learn
how to apply SAS/STAT in your work.

Cross-platform support and scalability Figure 2: A model-building progress plot created with the GLMSELECT procedure.
SAS runs on all major computing platforms
and can access nearly any data source. The
technology easily integrates into any organi-
zation’s computing environment and can
scale as you face larger or more complex
analytical problems.
Key Features
Analysis of Variance • Stepwise discriminant analysis. Missing Data Analysis
• Balanced and unbalanced designs. • More discriminant analysis capabilities. • Multiple imputation.
• Multivariate analysis of variance and • Weighted generalized estimating
repeated measurements. equations.
Distribution Analysis
• Linear models. • Imputation for survey data.
• Univariate and bivariate kernel
• More analysis of variance capabilities. • More missing data analysis capabilities.
density estimation.
• More distribution analysis
Bayesian Analysis capabilities. Mixed Models
• Built-in Bayesian modeling and infer- • Linear and nonlinear mixed models.
ence for generalized linear models, • Generalized linear mixed models.
Exact Inference
accelerated failure time models, Cox • Nested models.
• Exact p-values and confidence
regression models and finite mixture • More mixed models capabilities.
intervals for many test statistics and
models.
measures based on one-way and
• Wide range of Bayesian models
n-way frequency tables. Model Selection
available via general-purpose
• Exact tests for the parameters of • Linear models.
MCMC simulation procedure.
a logistic regression model. • Generalized linear models.
• Bayesian discrete choice modeling.
• Exact tests for the parameters of a • Quantile regression models.
• More Bayesian analysis capabilities.
Poisson regression model. • More model selection capabilities.
• More exact methods capabilities.
Categorical Data Analysis
Multivariate Analysis
• Contingency tables and measures of
Finite Mixture Models • Exploratory and confirmatory factor
association.
• Modeling of component distribu- analysis.
• Bioassay analysis.
tions and mixing probabilities. • Principal components analysis.
• Generalized linear models.
• Maximum likelihood and Bayesian • Canonical correlation and partial
• More categorical data analysis
methods. canonical correlation.
capabilities.
• More finite mixture capabilities. • More multivariate analysis capabilities.

Causal Inference
Group Sequential Design and Nonlinear Regression
• Propensity score analysis.
Analysis • Automatic derivatives.
• Estimation of causal treatment effects.
• Design of interim analyses. • Bootstrapped confidence intervals.
• Perform interim analyses. • More nonlinear regression
Cluster Analysis • More on group sequential design capabilities.
• Hierarchical clustering of multivariate and analysis capabilities.
data or distance data.
Nonparametric Analysis
• Disjoint clustering of large data sets.
Longitudinal Data Analysis • Kruskal-Wallis, Wilcoxon-Mann-
• Nonparametric clustering with
• Marginal and mixed models. Whitney and Friedman tests.
hypothesis tests for the number of
• Continuous and categorical • Other rank tests for balanced or
clusters.
outcomes. unbalanced one-way or two-way
• More cluster analysis capabilities.
• More longitudinal data analysis designs.
capabilities. • Exact probabilities for many
Descriptive Statistics nonparametric statistics.
• Box-and-whisker plots. • More nonparametric analysis
Market Research
• Compute directly and indirectly capabilities.
• Simple and multiple correspondence
standardized rates and risks for study
analysis.
populations.
• Two-way and three-way metric and Nonparametric Regression
• More descriptive statistics capabilities.
nonmetric multidimensional scaling • Multivariate adaptive regression
models. splines.
Discriminant Analysis • Discrete choice models. • Generalized additive models.
• Canonical discriminant analysis. • More market research capabilities. • Local regression.
Key Features (continued)
• Thin-plate smoothing splines. • Model selection for linear regression • Missing value imputation.
• More nonparametric regression models. • More survey sampling and analysis
capabilities. • More quantile regression capabilities. capabilities.

Power and Sample Size Regression Survival Analysis


• Computations for linear models • Least squares regression. • Nonparametric survival function
including MANOVA repeated • Principal components regression. estimates.
measurements. • Quadratic response surface models. • Competing-risk models.
• Computations for many hypothesis • Accurate estimation for ill-conditioned • Accelerated failure time models.
tests, equivalence tests and correla- data. • Proportional hazards models.
tion analysis. • More regression capabilities. • Interval-censored data analysis.
• Computations for binary logistic • More survival analysis capabilities.
regression and survival analysis.
• More power and sample size Robust Regression
capabilities. • M estimation and high-breakdown High Performance
methods. • 14 SAS/STAT procedures are
• Outlier diagnostics. multithreaded.
Post Processing • More robust regression capabilities. • 12 SAS® High-Performance Statistics
• Hypothesis tests. procedures are available with
• Prediction plots. SAS/STAT for single machine use.
• Scoring. Spatial Analysis
• More post-processing capabilities. • Ordinary kriging in two dimensions.
• Spatial point pattern analysis. Statistical Graphics
• Variogram diagnostics. • Hundreds of statistical graphs
Predictive Modeling • More spatial analysis capabilities. available with analyses.
• Classification and regression trees. • Customization provided.
• Partitioning of data into training, vali-
• Base SAS “SG” procedures create
dation and testing roles. Standardization user-specified statistical graphics.
• Modern model selection methods • 18 standardization methods.
• More statistical graphics capabilities.
such as elastic net and group LASSO. • More standardization capabilities.
• More predictive modeling
capabilities.
Structural Equations   TO LEARN MORE  » 
• Structural equation models specified
Psychometric Analysis with popular modeling languages. To learn more about SAS/STAT, download
• Multidimensional scaling. • Parameter estimation and hypothesis white papers, view screenshots and see
• Conjoint analysis with variable testing for constrained and uncon- other related material, please visit
transformations. strained problems. sas.com/stat.
• Item response theory (IRT) models. • More structural equation
• More psychometric analysis capabilities. SAS/STAT takes advantage of the SAS®9
capabilities.
engine, part of the SAS Platform. Many SAS
Survey Sampling and Analysis procedures have been enhanced so code
Quantile Regression • Sample selection. launched from SAS 9 can run in SAS Viya,
• Simplex, interior point and • Descriptive statistics. the SAS Platform’s new distributed,
smoothing algorithms. • Linear and logistic regression. in-memory engine. For more information,
• Analysis of censored data. • Proportional hazards regression. visit sas.com/platform.

To contact your local SAS office, please visit: sas.com/offices


SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute
Inc. in the USA and other countries. ® indicates USA registration. Other brand and product names are trademarks of
their respective companies. Copyright © 2017, SAS Institute Inc. All rights reserved. 101372_G60016.0917

You might also like