book

Statistical Programming in SAS®

by A. John Bailer

July 2010

Intermediate to advanced

460 pages

8h 24m

English

SAS Institute

Read now

Unlock full access

PurposeIntended AudienceScope of This BookUse in a Statistical Programming ClassBelief and StyleSoftware Used for This BookCode Used in This BookSAS Author PageAdditional ResourcesComments or Questions?SAS Publishing News
1.1 Statistical Computing versus Programming versus Managing Data1.2 Good Programming Practice1.2.1 Document your programs!1.2.2 Use meaningful variable names1.2.3 DON'T USE ONLY CAPITALS IN PROGRAM STATEMENTS (although some judicious use is reasonable)1.2.4 Indent program statements that naturally go together1.3 SAS Program Structure1.4 What Is a SAS Data Set?1.5 Internally Documenting SAS Programs1.6 Summary
1.8 Exercises
2.1 Temporary versus Permanent Status of Data Sets2.2 Reading Data into a SAS Data Set2.2.1 Reading data directly as part of a program–anyone for datalines?2.2.2 Reading data sets saved as text–INFILE can be your friend2.2.3 Sometimes variables are in particular columns or in particular formats2.2.4 Reading comma-separated values–text files with comma delimiters2.2.5 Reading Excel spreadsheets directly2.2.6 Reading SPSS data files–the little SPSS engine that could2.3 Writing Out a File or Making a Simple Report2.4 Concatenating Data Sets and Adding Observations2.5 Merging Data Sets and Adding Variables2.6 Database Processing with PROC SQL2.7 Summary
2.9 Exercises2.10 Self-Study Laboratory Explorations
3.1 SAS System Options3.2 Statements That Can Modify the Output of Most Procedures3.3 Defining Your Own Formats for Variable Values3.4 Selecting or Stratifying an Analysis by Values of a Variable3.5 Displaying Data Set Properties and Observations3.6 Using PROC PRINT to List the Observations in a Data Set3.7 Basic Graphical Displays3.8 Using Scatter Plots to Display Relationships between Numeric Variables3.8.1 Comparing distributions of responses between multiple groups3.9 Summarizing Categorical Variables3.10 Summarizing Numeric Variables3.10.1 PROC MEANS for descriptive statistics3.11 Selecting a Simple Random Sample3.12 Randomly Assigning Treatments to Observations3.13 Summary
3.15 Exercises

4.1 Introducing PROC TABULATE4.2 Building from Simple Specifications4.3 Enhancing PROC TABULATE Output4.4 Using the Output Delivery System4.4.1 Basic ideas4.4.2 Destinations—RTF, HTML, PDF, and more!4.4.3 What's produced and how to select it4.4.4 Another destination that stat programmers should visit—OUTPUT4.5 Summary
4.7 Exercises
5.1 Overview of Modeling5.2 Linear Regression Models5.2.1 Motorboats and manatees—a look at simple linear regression5.2.2 Big brains and big bodies—specifying and fitting a multiple regression model5.3 ANOVA Models—PROC GLM for a One-Way ANOVA5.3.1 Comparing bacterial growth under different packaging conditions using a one-way ANOVA model5.4 ANOVA Models—PROC GLM for an ANOVA Model with Two or More Factors5.5 Summary
5.7 Exercises
6.1 Graphics in SAS6.2 ODS Statistical Graphics6.3 Modifying Graphics Using the ODS Graphics Editor6.4 Graphing with Styles and Templates6.5 Statistical Graphics—Entering the Land of SG Procedures6.6 Case Study: Using the SG Procedures6.7 Summary
6.9 Exercises
7.1 Traditional SAS Graphics7.2 Customizing Graphics7.3 Why You Need to Learn about Annotate Data Sets7.4 Case Study: Comparing Distributions of Responses7.5 Descriptive Displays of Spatial Data7.6 Summary
7.8 Exercises7.9 Appendix: Complete Ohio County Population Data Set
8.1 Internal Representations and Output Displays8.1.1 Defining your own formats and informats8.2 Character, Numeric, Time, and Date Formats8.3 Recoding and Transforming Variables in a DATA Step8.4 Ordering How Tasks Are Done8.5 What Goes and What Stays in a Data Set8.6 Structured Thinking about Writing Programs8.7 Case Study 1: Is the Two-Sample t-Test Robust Enough for Heterogeneous Variances?8.7.1 Case Study 1: Is the Two-Sample t-Test Robust Enough for Heterogeneous Variances? (Revisited Using More DATA Step Programming)8.8 Case Study 2: Monte Carlo Integration to Estimate an Integral8.9 Case Study 3: Simple Percentile-Based Bootstrap8.10 Throw Out Your Tables of Statistical Distributions8.11 Generating Variables Using Random Number Generators8.12 Summary
8.14 Exercises
9.1 Storage Bins for Collections of Values9.1.1 Defining values in the variable list9.1.2 Inputting values in the variable list9.1.3 Reassign missing value codes for numeric variables "."9.1.4 Recoding missing values for all numeric and character variables9.1.5 Creating multiple observations from a single record9.2 Case Study 1: Monte Carlo p-Value for Test of Spatial Randomness9.3 Remembering Variable Values across Observations9.3.1 Processing multiple observations for a single observation9.4 Case Study 2: Randomization Test for the Equality of Two Populations9.5 Summary
9.7 Exercises
10.1 What Is a Macro and Why Would You Use It?10.2 Motivation for Macros: Numerical Integration to Determine P(0<Z<1.645)10.3 Processing Macros10.4 Macro Variables, Parameters, and Functions10.5 Conditional Execution, Looping, and Macros10.6 Debugging Macro Code and Programs10.7 Saving Macros10.8 Functions and Routines for Macros10.9 Bonus Material: Processing Multiple Data Sets10.10 Summary
10.12 Exercises
11.1 Defining a Matrix and Subscripting11.2 Using Diagonal Matrices and Stacking Matrices11.3 Using Elementwise Operations, Repeating, and Multiplying Matrices11.4 Importing a Data Set into SAS/IML and Exporting Matrices from SAS/IML to a Data Set11.4.1 Creating matrices from SAS data sets and vice versa11.5 Case Study 1: Monte Carlo Integration to Estimate π11.6 Case Study 2: Bisection Root Finder11.7 Case Study 3: Randomization Test Using Matrices Imported from PROC PLAN11.8 Case Study 4: SAS/IML Module to Implement Monte Carlo Integration to Estimate π11.8.1 Storing and loading SAS/IML modules11.9 Introducing SAS/IML Studio11.9.1 Case Study 1: Dynamic and interactive analysis of the SMSA data set11.9.2 Case Study 2: Multiple-linked graphics windows11.9.3 Case Study 3: SAS/IML matrix manipulations and invocations of SAS/STAT procedures11.9.4 Case Study 4: Generating bootstrap CIs for mean AGE using R11.10 Summary
11.12 Exercises

Overview

In Statistical Programming in SAS, author A. John Bailer integrates SAS tools with interesting statistical applications and uses SAS 9.2 as a platform to introduce programming ideas for statistical analysis, data management, and data display and simulation. Written using a reader-friendly and narrative style, the book includes extensive examples and case studies to present a well-structured introduction to programming issues.

This book has two parts. The first part addresses the nuts and bolts of programming, including fostering good programming habits, getting external data sets into SAS to construct an analysis data set, generating basic descriptive statistical summaries, producing customized tables, generating more attractive output, and producing high-quality graphical displays. The second part emphasizes programming in the context of a DATA step, in macros, and in SAS/IML software.

Examples of statistical methods and concepts not always encountered in basic statistics courses (for example, bootstrapping, randomization tests, and jittering) are used to illustrate programming ideas. This book provides extensive illustrations of the new ODS Statistical Graphics procedures in SAS, a description of the new ODS Graphics Editor, and a brief introduction to some of the capabilities of SAS/IML Studio, such as producing dynamically linked data displays and invoking R from SAS.