You are on page 1of 2

Book Reviews

doi: 10.1111/1753-6405.12166

SPSS survival manual:


a step by step guide to data
analysis using IBM SPSS
By Julie Pallant. 5th edition. Published by Allen and Unwin,
Australia, 2013; ISBN-978-1-74331-400-5; 354 pages; RRP
$45.00.

Reviewed by Dr Peter Dunn


Faculty of Science, Health, Education and Engineering,
University of the Sunshine Coast, Queensland

When this book appeared on my desk, I thought “Another SPSS


book!” I then went on to Amazon.com and searched their books
section for “SPSS”, which returned 3,767 results. In other words,
this text has a lot of competition. Given that the book is into its 5th
edition, the author must be doing something right.
This book has many good points – but first, some comments.
Unusually, the book includes very few screenshots of SPSS; instead
Pallant describes the steps necessary to complete given tasks in SPSS
(and also gives the SPSS syntax and provides SPSS output). This is
not a criticism, but an observation that may immediately lead some
readers to choose a different book.
Another observation: This book is not about learning SPSS; the
purpose of this book is about conducting research and using SPSS
as a tool to accomplish this goal. As examples, the first chapter is
called Designing a study, and Chapter 10 is called Choosing the right
statistic; SPSS barely rates a mention in either. In addition, the book
often points the reader to other software, other books, or to manual
calculations when necessary.
The book is divided into five parts: Getting started; Preparing
the data file; Preliminary analyses; Statistical techniques to explore
relationships among variables; and Statistical techniques to compare
groups. Although the author explicitly states that “this book is not
intended to cover all possible statistical procedures available in IBM

2013 vol. 37 no. 6 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 597
© 2013 The Authors. ANZJPH © 2013 Public Health Association of Australia
Book Reviews

SPSS…”, the most common statistical techniques are included: data files in Chapter 4, but not one screenshot appears in this chapter
descriptive statistics; graphics; correlation and regression (including (or elsewhere) showing data entered into SPSS. One scenario comes
logistic regression); factor analysis; t-tests, common non-parametric to mind that I have seen confuse students and researchers alike:
tests; chi-square tests of independence; one-way and two-way preparing data files for paired t-tests compared to two independent
ANOVA; MANOVA; and ANCOVA. Each technique is well explained samples’ t-tests. The lack of screenshots is obviously an intentional
– appropriately, and in context. For example, Pallant shows how to design feature of the book, but a few carefully chosen screenshots
communicate the SPSS output, and essential information about each would have been helpful.
analysis is given (an example research question; what type of data is A further criticism relates to the confusing discussion at the start
needed; what the technique does; assumptions; alternative analyses). of Chapter 15 (entitled Factor Analysis). The chapter introduction
This information is usually terse, but helpful. explains (correctly) that “the term ‘factor analysis’ encompasses
Despite the plaudits, problems are evident. Once of the more a variety of different… techniques. One of the main distinctions is
curious – and most important – is Pallant’s use of the Legacy Dialogs between principal components analysis (PCA) and factor analysis
for some procedures, rather than the newer approaches (available (FA).” On the next page Pallant then declares that she has “chosen to
since version 18, released in 2009). This often has important demonstrate principal components analysis in this chapter”, but later
ramifications. For example, Pallant describes the Mann-Whitney test on the same page says, “There are three main steps in conducting
(p. 235) as a test that “actually compares medians”, which is only factor analysis (I am using the term in a general sense to indicate
true under certain circumstances. The new procedures lead the user any of this family of techniques, including principal components
to either an explicit test of medians, or to the Mann-Whitney test for analysis)”. What is a reader expected to make of this?
comparing distributions. Like many books, this text has an associated webpage (www.
Furthermore, when discussing post hoc tests that may emerge allenandunwin.com/SPSS). This page includes the six data files used
from a Kruskal-Wallis test, Pallant instructs the reader to use a throughout the text (in SPSS format only), answers to all the exercises
series of pairwise Mann-Whitney tests, and to manually adjust the at the end of (some) chapters, a feedback link, some resources for
resulting p-values using the Bonferroni correction. However, the students, and an instructor’s guide.
new procedures automatically conduct post hoc tests and adjust for In summary, this book is easy to read, has a functional (though
multiple testing (using Dunn’s test). By remaining with the Legacy short) index, explains the techniques well in context, is spiral bound
Dialogs, Pallant potentially directs users away from using the correct so it lays flat when working at the computer, and explains SPSS as
tests, and actually makes analysis more difficult. The Legacy Dialogs part of the research process. However, a number of shortcomings
are also used when describing the production of graphics. Perhaps severely hinder its usefulness in practice, and may actually mislead.
this insistence on retaining the old ways reflects that this book is
showing signs of age.
Some other omissions are also curious. For example, only eight
lines of text (p. 23) are devoted to showing the user how to use the
extensive help system in SPSS. Another is the lack of direction in
preparing the data files. True, Pallant discusses the preparation of

598 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2013 vol. 37 no. 6
© 2013 The Authors. ANZJPH © 2013 Public Health Association of Australia

You might also like