Scribd
Upload
38 views

Mcnutt 1996

This book provides an accessible introduction to programming with SAS. The book removes the mystery around SAS programming and shows how it can provide a flexible approach to data analysis beyond standard procedures. It covers key programming concepts like input/output, operations, logic, branching, loops and arrays in an easy to understand way. While recommending expansion of the chapter on standard procedures, the review praises the book as a valuable resource for researchers to gain more power and flexibility in data analysis.

Uploaded by

chemistpl420
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
38 views

Mcnutt 1996

This book provides an accessible introduction to programming with SAS. The book removes the mystery around SAS programming and shows how it can provide a flexible approach to data analysis beyond standard procedures. It covers key programming concepts like input/output, operations, logic, branching, loops and arrays in an easy to understand way. While recommending expansion of the chapter on standard procedures, the review praises the book as a valuable resource for researchers to gain more power and flexibility in data analysis.

Uploaded by

chemistpl420
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Book Reviews

SAS® Programming for Researchers and Social


Scientists, by Paul E. Spector. Newbury Park,
CA: Sage Publications, 1993, 169 pp.

Reviewed by: JOHN G. McNUTT

This is book I wish I had had in graduate school. In less than 200 pages, Paul Spector
a

removes the veil of mystery that has surrounded programming with statistical packages. No
longer will users of SAS@ be restricted to those procedures offered in the familiar &dquo;PROC&dquo;
commands. The SAS@ programming language offers a flexible approach to data analysis and
Spector’s work makes it accessible.
Programming has become almost a lost art among data analysts. A look through any of
the older manuals will reveal that in the early days of statistical packages, some understanding
of programming was needed for most applications. As the packages improved, less program-
ming was needed. Since the advent of sophisticated graphic user interfaces, such as Windows
and Macintosh systems, using the packages has become consistently easier and little or no
programming is required. As a consequence, users no longer have the need to develop skills
in using the programming languages available in these packages. This means that these users
are limited to the procedures available in the standard programs. When faced with difficult or
unusual situations where the standard solutions do not apply, they are at a disadvantage. Spec-
tor’s book will be a welcome help to those in this predicament.
Spector uses what he calls a problem solving approach to programming. He demon-
strates how programming can address a research problem that may not be covered by the stan-
dard programs.
He begins by discussing the SAS@ system in detail. In the initial chapter, he relates the
task of programming to other aspects of the SAS@ package. Based on this discussion, he pro-
vides a strategy for programming. He does an excellent job of demonstrating that program-
ming is an integral part of the SAS@ approach to data analysis.
The next chapter looks at input and output functions with SAS@. Spector shows how
the user can go beyond the usual conventions in this phase. Chapter 3 looks at algebraic
and mathematical operations. This discussion demonstrates the power of programming in

Dr. John G. McNutt •Graduate School of Social Work, Boston College, Chestnut Hill, MA 02167.

283

Downloaded from aje.sagepub.com at NORTH DAKOTA STATE UNIV LIB on June 20, 2015
284

the SAS@ environment. Discussions of random number generators and probability func-
tions are included.
Logical statements are discussed in chapter 4. As a bonus, one section shows the reader,
step by step, how to develop a program to address a common research need-detecting error
in a data file. Branching is discussed in chapter 5 and loops are discussed in chapter 6. These
are standard programming issues and Spector shows clearly how to perform these tasks effec-

tively and avoid the traps. Chapter 6 also shows the reader how to set up a simulation for eval-
uating the effects on alpha levels of violating t-test assumptions. Arrays are covered in chapter
7, including a discussion of matrix operations with SAS@.
Chapter 8 discusses ways to combine data files within SAS@. This is valuable informa-
tion that is often difficult to find, and this discussion alone is worth the price of the book. A
brief examination of the familiar SAS@ &dquo;PROC&dquo; procedures is presented in Chapter 9. Only
the basic procedures are discussed.
The concluding chapter presents final advice for SAS@ programmers, and information
on debugging, preplanning and alternative problem solving. This chapter promises a great
deal but delivers less. I found the discussion superficial, but still worth reading.
This is a valuable book. It is well written and organized in an easy to follow manner. The
author not only explains the fundamentals of programming, he hints at the possibilities of fur-
ther exploration into the core of statistical analysis. While he doesn’t make programming look
easy, he does make it look possible and important. The reader will gain a new respect for the
flexibility and precision that the SAS@ programming language provides. Resorting to stan-
dard solutions will never again seem so attractive.
If I had to offer a major criticism, it would be that chapter 9, on &dquo;PROC&dquo; commands,
should have been either expanded or deleted. The &dquo;PROC&dquo; commands are an important part
of SAS@, and the cursory discussion in the book is not adequate. I hope that the author will
consider greatly expanding this discussion in a future edition.
SAS@ Programming for Researchers and Social Scientists will be useful to seasoned
evaluators and researchers as well as to students in graduate social science or evaluation meth-
odology programs. By and large, much of what Spector says can be used with SPSS@ and
other statistical packages. The logic will apply to other programming situations as well. The
evaluator or researcher who takes the time to learn what this book has to teach will not only
become a better analyst, he or she will also gain a deeper understanding of statistics and data.

Downloaded from aje.sagepub.com at NORTH DAKOTA STATE UNIV LIB on June 20, 2015

You might also like