«

»

Book review – Julie Pallant (2013). SPSS survival manual (5th edition). Sydney: Allen and Unwin

Print Friendly, PDF & Email

Return to Journal Articles


Felicity Allen, Monash University

 

It’s fair to say that not everyone would greet the appearance of a book with the word ‘SPSS’ in the title with a glad cry. There are probably two reasons for my reaction: (1) I taught multivariate statistics for many years and the damage is probably permanent by now, and (2) my experience with using earlier editions of this book has been very positive.

Many psychologists dislike statistics and loathe conducting analyses but every now and then it becomes essential to do so, particularly if you are considering a return to study. Many funding bodies now require that basic statistics be supplied to them and sometimes even evidence of efficacy. In either case, Pallant’s book would be an excellent investment for you.

Julie Pallant is an Australian academic of many years’ experience and, like a good education, she begins at the beginning. The design of your study (or data collection exercise) is absolutely crucial to determining the kinds of conclusions that can be drawn from it. Pallant begins her book with a thorough consideration of the options that you have when designing a study. Tempting as it may be to think “I’ll just turn to the page about getting a Chi-square”, it’s well worth putting in the time to read the first three chapters carefully. This is the information which will give the numbers generated by research their vital context and meaning.  Most writers make methodology discussions as dull as ditch water – not Pallant – try this quote on for size:

People are notoriously unreliable – they don’t turn up when they are supposed to, the get sick, drop out and don’t fill out questionnaires properly! So plan accordingly. Err on the side of pessimism rather than optimism (p.4).

She is no remote, dull, maths professor!

Moving into the use of the software itself in chapter four, she immediately introduces the idea that you have choices and options in the ways that data and output are displayed. This alone should break into that miserable sense of loss of control that sweeps over many people facing a statistical software package. She advises you how to deal with the enormous tables that SPSS can sometimes produce and how to keep them under control for transfer to Word documents. Very clear instructions on setting up the file and creating the variables are supplied.

So what have you got at the end of this stage? A printout of your main variables ready for careful consideration, checking for errors and correcting any that are found.  Pallant tries to make it clear to the reader that, while it’s great to have a print out, that does not necessarily mean that all those figures are correct.  She gives a number of strategies to help people check for errors and then to correct them when they are found.

Once you have reached the point of having a lovely, clean data file, the choice of statistical analysis is largely determined by the kinds of hypotheses that you have, but all the same there are certain commonalities for all researchers. All students are strongly encouraged by statistics lecturers to get a full printout of all their variables, but judging from the people who came to see me about their analyses, very few had either understood what that meant or bothered to do it. It is very important to have a clear vision of exactly what your data consists of and Pallant explains how you can get that and why you should do so.

In the interests of testing the product on myself, I turned to a chapter on an area of SPSS which I rarely use – Chapter 7 – Using graphs to describe and explore data. The step by step coverage of the necessary instructions is very clear – and I learned something! Pallant is also careful to warn readers against the very human tendency to “believe their eyes”. It may look like a big difference, but it may not be significant. Once you have got the graph, you need to take the next step and test it to see whether there is a real difference. Mind you, the graph itself will almost certainly cheer up your report or thesis.

This edition has extended Pallant’s clear approach to several multivariate procedures and this provides a welcome alternative to her main competitor: Andy Field (Discovering statistics using SPSS: and sex, drugs, rock and roll). While many people love Field, other people find his very chatty style is distracting.

Any caveats? There is a section of SPSS called “Reports”, which offers what seems to be fascinating ways of manipulating and presenting data, but despite years of effort I have not yet managed to crack it in any useful sense. I certainly hope that Julie Pallant turns her talents to that section in her 6th edition.

 


Return to Journal Articles