Doing Statistics With SPSS

Alistair W Kerr, Howard K Hall, Stephen A Kozub

• Publisher: SAGE Publications Ltd
• Year: 2002
• Print ISBN: 9780761973850, 9780761973843
• eBook ISBN: 9781446230718
• Edition: 1
• Topic: Social Sciences
• Subtopic: Social Science Research & Methodology
• Index: Social Sciences

`This book is to be commended, particularly, for putting the tool of statistics into the familiar context of a research study. In so doing it emphasizes the neglected pre-analysis stages of a research study. Indeed the performing of a data analysis, this book reminds us, should be the mere icing on an already well cooked cake? - Psychology Learning & Teaching

Doing Statistics With SPSS is derived from the authors? many years of experience teaching undergraduates data handling using SPSS. It assumes no prior understanding beyond that of basic mathematical operations and is therefore suitable for anyone undertaking an introductory statistics course as part of a science based undergraduate programme. The text will: enable the reader to make informed choices about what statistical tests to employ; what assumptions are made in using a particular test; demonstrate how to execute the analysis using SPSS; and guide the reader in his/her interpretation of its output. Each chapter ends with an exercise and provides detailed instructions on how to run the analysis using SPSS release 10. Learning is further guided by pointing the reader to particular aspects of the SPSS output and by having the reader engage with specified items of information from the SPSS results.

This text is more complete than the alternatives that usually fall into one of two camps. They either provide an explanation of the concepts but no instructions on how to execute the analysis with SPSS, or they are a manual which instructs the reader on how to drive the software but with minimal explanation of what it all means. This book offers the best elements of both in a style that is economical and accessible.

Doing Statistics with SPSS will be essential reading for undergraduates in psychology and health-related disciplines, and likely to be of invaluable use to many other students in the social sciences taking a course in statistics.

1	THE NEED FOR STATISTICS

When you enrolled for a degree course, you were probably under the impression that much of your time would be spent playing sport, participating in outdoor pursuits, meeting new friends, enjoying your leisure activities in the many local hostelries, and even attending the odd lecture, seminar or workshop to learn about important phenomena in your field of interest. To discover that you had to register for a course in data handling is likely to have generated little enthusiasm, and more likely caused you some concern. So, why is it that many degree programmes require students to undertake a statistics course?

The answer to this question lies in the fact that statistics play a vitally important role in research. They help answer important research questions and it is the answers to such questions that further our understanding of the field and provide a knowledge base for academic study. This body of knowledge resides, for the most part, in the textbooks and academic research journals that grace the shelves of university libraries. In the future this knowledge base will continue to expand as many more answers to important research questions are provided.

At this point you may be thinking that you have no interest in research or in becoming a researcher. Although this may be true it is also a fact that during the next few years you will be asked to read a considerable amount of information as part of your academic programme of study. This activity will require you to interpret the results of research employing many different research methods. Consequently, it is imperative that you become familiar with the research process so that you become an informed consumer of the literature in the field, and can think critically about everything you read. An awareness of how to answer research questions and read research articles will enable you to discern valuable information that forms part of the knowledge base in academic study, from unsubstantiated speculation that adds little to our understanding of important issues in the field. Therefore, not only will this book help you understand the research process, but it will help you reflect critically on the research literature.

It is not uncommon for students who are just beginning their university careers to be directed towards the library by academic staff and asked to read selected articles in preparation for a lecture or seminar. However, most undergraduate students find the task of reading a research article to be quite intimidating. If the task can be avoided, the article will rarely be read. If it cannot be avoided, the following strategy appears common. There is a tendency to read the abstract in minute detail, because it is short and concise. Once this task has been completed attention is turned towards the introduction and discussion sections because they appear to be most understandable and contain what most neophyte researchers believe to be the relevant information. Next, the methods section is typically given a cursory glance, or at best it is skimmed because it tends to be dull and boring and simply describes what the researchers did. Then the results section is ignored completely because the statistics appear incomprehensible. Finally the discussion section is read because, like the introduction, it is partially comprehensible. Although this is an understandable strategy when one first encounters research articles, it is not one that is going to produce either informed consumers of research, or university graduates who are independent critical thinkers. Worse still, it can lead to the acceptance of conclusions that a critical examination of the methods and results sections would have shown to be at best speculative and at worst incorrect. Changing this strategy so that you are able to feel confident about reading the results section of a research article is a central aim of this text. However, simplifying the task of understanding the statistics section in a research article is not easy. Even the most dedicated knowledge seeker can be turned off by terms such as canonical correlation, factorial analysis of variance, or discriminant function analysis. However, to avoid getting bogged down with statistical terminology it would be especially wise at this point to keep in mind that statistics are simply tools that researchers employ to help answer research questions.

Using the Appropriate Tools

The selection of the most appropriate tool to perform a job is a crucial issue whether you are a carpenter, a surgeon, a motor mechanic or a research scientist. The tool selected by a carpenter obviously depends upon the job that is being done at the time. Similarly, which statistical tool a research scientist selects depends upon the question being asked. Unfortunately, it is all too noticeable that when students are asked to conduct research they often base the selection of their statistical analysis upon tests which they might have heard mentioned once or twice, rather than upon their suitability for the job. This is analogous to a homeowner who attempts to perform all the major home improvement projects with the first tool that is pulled from a toolbox.

Let us suppose that our homeowner had chosen to work on a project that involved installing a new door on a shed. To complete the job correctly requires the door frame to be rebuilt, a little material to be removed from the bottom of the new door, the door to be rehung on its new hinges, and a new lock attached. All these jobs could be performed efficiently if the homeowner possessed the correct tools for each job. However, the only tool that our homeowner possessed was a hammer. Unfortunately, rather than making an expensive trip to the local DIY’ store to purchase the appropriate tools, the homeowner attempted to get by with the aid of this one trusty implement. Some early success was met because the door frame only required nailing into position. For this task the hammer proved to be the correct tool for the job. Taking a little off the bottom of the new door with a hammer was more difficult, but it was achieved with a few hefty whacks. The wood splintered and cracked around the bottom of the door, but it was only a shed so it didn’t really matter that the job wasn’t well done. Next, our homeowner attached the door to the new frame by bashing all the screws into place, again splintering the wood, and bending the screws. This meant the door hinges didn’t hold the door in place well but it looked fine from a distance, and it was better than the old door that suffered from dry rot. Finally, the homeowner forced a new lock into place with the aid of the trusty hammer. Unfortunately, the lock no longer worked after the homeowner’s efforts with the hammer. The homeowner thought that if the new door was tied to the frame with a piece of string it would look just fine –that is, until it needed opening.

Fortunately, most property owners do not go about making home improvements in this fashion. They attempt to utilise the correct tools for each job. However, the point to be considered here is that it was possible for our homeowner to complete the job on the shed door with only the use of the hammer. Upon closer inspection, however, the results indicated that the job was a complete disaster. Similarly, it would be possible for a researcher to select the wrong statistical techniques to analyse data and still be able to draw conclusions from the analyses. Only upon close scrutiny of the results by someone versed in research methods would they be found to be meaningless. Consequently, whether you are a homeowner or a novice researcher, it is recommended that you have an understanding of what tools are suitable for the particular jobs you are likely to undertake. This involves knowing how to use each tool, and recognising when each should be used. Knowledge such as this will allow you to avoid the drawbacks of selecting the wrong tools for the job. That is, picking up your hammer and treating everything as if it were a nail.

Numerophobia

Even when students become aware of the importance of using statistics to help answer research questions, many feel anxious about their mathematical expertise. This is not an uncommon phenomenon. Many intelligent people who are capable of competent performance in other areas experience ‘numerophobia’ (i.e. a fear of numbers) at the thought of having to solve more than the most basic mathematical problems. Fortunately, the subject of statistics is not mathematics, although it does involve the use of some simple mathematical operations. In using this manual you will need to know how to add, subtract, multiply, use brackets, and understand the principle of squaring and taking the square root of numbers. Beyond this you will not find yourself encountering any mind-blowing mathematics. What might cause a few problems are the symbols and notation that are used to describe particular operations or specify particular statistics. However, just as when learning a foreign language one has to become familiar with new words and symbols in order that the language makes sense, the learning of unfamiliar statistical symbols and notation will take time, patience and some considerable thought before they become second nature to you. The meaning of these symbols and notation will be explained to you as you progress through the manual. You can also find a brief description of symbols and notation in Appendix 1. However, if you still find that you do not understand something, consider the alternative strategies that might be available to you to help overcome this barrier. It is important to remember that if you continue to invest effort into understanding the principles covered in this manual you will begin to develop some confidence in the subject and overcome many potential barriers to understanding. In contrast, if you adhere to a view that your lack of understanding cannot be changed, no matter how hard you try, then it is unlikely that you will develop any confidence in the subject matter and this may simply reinforce the numerophobic condition.

The importance of this fact is illustrated by the following story. It describes an incident that took place at the end of the third or fourth lecture of an introductory statistics course being taught by one of the authors of this text. A disgruntled student approached the lecturer and announced angrily: ‘I hate this course! I don’t understand any of it.’ When the lecturer enquired about which aspects in particular were causing the problem, the student looked vacantly at him and replied, ‘All of it. I just don’t have a clue about anything!’ At this point, the lecturer thought for a while and asked why the student hadn’t communicated this lack of understanding during the lecture when he had attempted to solicit feedback by asking if everyone understood what was said. Back came the reply: ‘Do you think I’m stupid? I’m not going to put up my hand in front of the class and say that I don’t understand. You’ve got to realise that just because there’s silence when you ask if we understand, doesn’t mean that we do. Anyway, I don’t want everyone thinking I’m a half-wit, so I’m not going to give them the opportunity, and publicly embarrass myself in the process.’ When quizzed as to why one might perceive others in the class to be so judgemental, the student responded by saying ‘Everyone else seems to understand the subject, and I don’t.’ ‘But why does it matter if others really do know more than you?’, said the lecturer. ‘Surely, what is important is that you feel you’ve learned something in the course.’ This response was met by a torrent of verbal abuse and a diatribe on why it would probably be better to give up on the course rather than to suffer the indignity of realising that others might be more competent and to experience the humility of publicly displaying this lack of understanding.

Realising he was inflaming an emotionally charged situation, the lecturer tried another strategy. He enquired if there were parts of the lecture that made any sense that day. At this point, the student began to describe the different measures of central tendency and their properties. ‘You see’, said the lecturer, ‘You do understand something.’ The lecturer then stepped up to the whiteboard and drew a long horizontal line, representing the continuum of student understanding. The continuum ranged from no knowledge of statistics to a sound understanding of the subject. He then asked the student to mark a cross at some point on this line which signified the level of understanding of statistics at the start of the course. The student marked the board at the extreme left above the words ‘no understanding’. Then the lecturer asked the student to repeat the process by marking a cross on the line to signify current understanding. The student responded by inserting the cross a couple of centimetres to the right of the first mark. At this point, the lecturer looked at the student, and noting that there didn’t seem to have been much progress made in understanding, said: ‘As I see it, my job is to help you move along this continuum. How far you get in this class depends on how willing you are to invest in reading the assigned material, completing the practical exercises, and asking questions to help you comprehend what it is you’ve read. Everyone is likely to end up at a different point on the continuum by the semester’s end, but that’s OK. What is important is that you begin to develop a clearer understanding of the subject.’ Realising that it was more important to focus upon personal improvement and understanding than on how one’s performance compared with that of others in the class, the student departed with a little more reassurance that continued investment could bring rewards.

We would like to be able to recount that this student is now a statistical consultant for a nationally known research organisation, but you would probably not believe this. However, the point of this true story is very simple. Your understanding of statistics will be a direct function of your personal investment in the subject matter. Therefore spend time reading around the subject and asking questions about issues of concern. You will reap the benefits in the future as you read and learn more about the research process and the role it plays in your field of interest.

Stages of the Research Process

Identifying the Research Question

Whether you are reading published research, or attempting to conduct a research project, the ...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Contents
• 1 THE NEED FOR STATISTICS
• 2 DESCRIPTIVE STATISTICS
• 3 THE NORMAL DISTRIBUTION
• 4 INTRODUCTION TO EXPERIMENTAL RESEARCH
• 5 SAMPLING ERROR
• 6 BETWEEN-GROUPS ONE-WAY ANALYSIS OF VARIANCE
• 7 ANALYTICAL COMPARISONS
• 8 WITHIN-GROUPS ONE-WAY ANOVA
• 9 FACTORIAL ANOVA
• 10 MORE ADVANCED APPLICATIONS IN ANOVA
• 11 LINEAR REGRESSION
• 12 MULTIPLE REGRESSION
• 13 TWO FINAL EXERCISES
• APPENDIX 1 GLOSSARY
• APPENDIX 2 TEST SELECTION TABLE
• APPENDIX 3 ANSWERS TO THE EXERCISES
• APPENDIX 4 ILLUSTRATION OF HOW CHAPTER 13 MIGHT BE COMPLETED USING THE DATA ON PAGE 74
• Index