This edition first published 2020 © 2020 by John Wiley & Sons Ltd
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.
The right of Michael Nelson to be identified as the author of in this work has been asserted in accordance with law.
Registered Office(s)
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office
9600 Garsington Road, Oxford, OX4 2DQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of Warranty
The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting scientific method, diagnosis, or treatment by physicians for any particular patient. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging‐in‐Publication Data
Names: Nelson, Michael (Nutritionist), author.
Title: Statistics in nutrition and dietetics / Michael Nelson.
Description: Hoboken, NJ : John Wiley & Sons, 2020. | Includes bibliographical references and index.
Identifiers: LCCN 2019030279 (print) | ISBN 9781118930649 (paperback) | ISBN 9781118930632 (adobe pdf) | ISBN 9781118930625 (epub)
Subjects: MESH: Nutritional Sciences–statistics & numerical data | Statistics as Topic | Research Design
Classification: LCC RM217 (print) | LCC RM217 (ebook) | NLM QU 16.1 | DDC 613.2072/7–dc23
LC record available at https://lccn.loc.gov/2019030279
LC ebook record available at https://lccn.loc.gov/2019030280
Cover Design: Laurence Parc | NimbleJack &Partners | www.nimblejack.co.uk
Cover Image: © filo/Getty Images; PHN Courtesy of Public Health Nutrition Research Ltd
To Stephanie
Dr. Michael Nelson is Emeritus Reader in Public Health Nutrition at King's College London, and former Director of Research and Nutrition at the Children's Food Trust. He is currently Director of Public Health Nutrition Research Ltd (http://www.phnresearch.org.uk/).
His early career with the Medical Research Council sparked a keen interest in nutritional epidemiology, statistics, and measurement validity. Research interests have included the diets of UK school children and links between diet and poverty, cognitive function, behaviour and attainment, and monitoring the impact of standards on school lunch take-up and consumption. He collaborates nationally and internationally to promote a strong evidence base for school food policy. He has published over 200 peer-reviewed articles and other publications in the public domain.
January 2020
Worldwide, there is no basic statistics textbook that provides examples relevant to nutrition and dietetics. While it could be argued that general medical science statistics texts address the needs of nutrition and dietetics students, it is clear that students find it easier to take on board the concepts relating to statistical analysis and research if the examples are drawn from their own area of study. Many books also make basic assumptions about students' backgrounds that may not always be appropriate, and use statistical jargon that can be very off‐putting for students who are coming to statistics for the first time.
The book is aimed at undergraduate and postgraduate students studying nutrition and dietetics, as well as their tutors and lecturers. In addition, there are many researchers in nutrition and dietetics who apply basic statistical techniques in the analysis of their data, for whom a basic textbook provides useful guidance, and which helps to refresh their university learning in this area with examples relevant to their own field.
The level of the material is basic. It is based on a course that I taught at King's College London over many years to nutrition and dietetics students, physiotherapists, nurses, and medical students. One of the aims was to take the fear and boredom out of statistics. I did away with exams, and assessed understanding through practical exercises and coursework.
This book takes you only to the foothills of statistical analysis. A reasonable competence with arithmetic and a little algebra are required. For the application of more demanding and complex statistical techniques, the help of a statistician will be needed. Once you have mastered the material in this book, you may want to attempt a more advanced course on statistics.
The aim of this book is to provide clear, uncomplicated explanations and examples of statistical concepts and techniques for data analysis relevant to learning and research in nutrition and dietetics. There are lots of short, practical exercises to work through. These support insight into why various tests work. There are also examples of SPSS1 output for each test. This makes it is possible to marry up the outcomes computed manually with those produced by the computer. Examples are taken from around the globe relating to all aspects of nutrition, from biochemical experiments to public health nutrition, and from clinical and community practice in dietetics. All of this is complemented by material online, including data sets ready for analysis, so that students can begin to understand how to generate and interpret SPSS output more clearly.
The book focuses on quantitative analysis. Qualitative analysis is highly valuable, but uses different approaches to data collection, analysis, and interpretation. There is an element of overlap, for example when quantitative statistical approaches are used to assess opinion data collected using questionnaires. But the two approaches have different underlying principles regarding data collection and analysis. They complement one another, but cannot replace one another.
Two things this book is not. First, it is not a ‘cookbook’ with formulas. Learning to plug numbers in to formulas by rote does not provide insight into why and how statistical tests work. Such books are good for reminding readers of the formulas which underlie the tests, but useless at conveying the necessary understanding to analyze data properly or read the scientific literature intelligently. Second, it is not a course in SPSS or Excel. While SPSS and Excel are used to provide examples of output (with some supporting syntax for clarity), it is no substitute for a proper course in computer‐based statistical analysis.
The book provides:
All of the exercises have worked solutions.
Some students say, ‘Why do we have to do the exercises by hand when the computer can do the same computations in a fraction of a second?’ The answer is: computers are stupid. The old adage ‘garbage in, garbage out’ means that if you don’t have insight into why certain tests work the way they do, a computer will generate output that might be meaningless, but it won’t tell you that you’ve made a mistake, or ask ‘Is this really what you wanted to do?’ So, the purpose of the textbook and supporting learning materials is to help ensure that when you do use a computer, what goes in isn’t garbage, and what comes out is correct and provides meaningful answers to your research questions that you can interpret intelligently.
Finally, it is worth saying that some students will find this textbook providing welcome explanations about why things work the way they do. Others will find it annoyingly slow and detailed, with too much explanation for concepts and applications that seem readily apparent. If you are in the first group, I hope you enjoy the care with which explanations and examples are presented and that it helps to demystify what may at first seem a difficult topic. If you are in the second group, read quickly to get to the heart of the matter, and look for other references and resources for material that you feel is better suited to what you want to achieve. However hard or easy the text seems, students in both groups should seek to make friends with a local statistician or tutor experienced in statistical analysis and not try and do it alone.
There are many unique features in this textbook and supporting material:
This textbook is based on over 20 years of teaching experience. There are four parts:
This introduces concepts related to the scientific method and approaches to research; populations and samples; principles of measurement; probability and types of distribution of observations; and the notion of statistical testing.
This covers the basic statistical tests for data analysis. For each test, the underlying theory is explained, and practical examples are worked through, complemented by interpretation of SPSS output.
Most undergraduate and postgraduate courses require students to collect data and/or interpret existing data sets. This section places the concepts in Part 1 and the learning in Part 2 into a framework to help you design studies, and determine sample size and the strength of a study to test your hypothesis (‘Power’). A Flow Chart helps you select the appropriate statistical test for a given study design.
The last chapter explores briefly how to present findings to different audiences – what you say to a group of parents in a school should differ in language and visual aids from a presentation to a conference of your peers.
It would be desperately unfair of me to set exercises at the end of each chapter and not provide the solutions. Sometimes the solutions are obvious. Other times, you will find a commentary about why the solution is what it is, and not something else.
No textbook is complete these days without online resources that students and tutors can access. For this textbook, the online elements include:
For lecturers delivering courses based on the textbook, I have prepared brief teaching notes. These outline the approach taken to teach the concepts set out in the textbook. I used traditional lecturing coupled with in‐class work, practical exercises, homework, and research protocol development. My current practice is to avoid exams for any of this material. Exams and formal tests tend to distract students from revision of study materials more central to their course. Some students get completely tied up in knots about learning stats, and they fret about not passing the exam, ultimately to their academic disadvantage.
The principal aid for tutors and lecturers is slide sets in PowerPoint. These save hours of preparation, provide consistent format of presentation, and build on approaches that have worked well with literally thousands of students that have taken these courses. When using the slides outside the context of teaching based on the text book, please ensure that you cite the source of the material.
A complete set of SPSS files for the examples and exercises in the text book is provided.
The page on Learning Resources includes website links and reviews of the strengths of a number of sites that I like and find especially helpful.
Unsurprisingly, there is a wealth of websites that support learning about statistics. Some focus on the basics. These are mainly notes from University courses that have been made available to students online. Some are good, some are not so good. Many go beyond the basics presented in this text book. Diligent searching by the student (or tutor) will no doubt unearth useful material. This will be equivalent to perusing the reading that I outline in the Introduction to Chapter 1.
Flow Charts are useful to find the statistical test that best fits the data. Appendix A10 in this book shows one. There are more online. Two that I like are described in more detail on the Learning Resources page. I have also included links to sites for determining Power and sample size.
Finally, guidance on the use of Excel and SPSS in statistics is very helpful. There are many sites that offer support, but my favourites are listed on the Learning Resources page.
I would like to thank the hundreds of students who attended my classes on research methods and statistics. They gave me valuable feedback on what worked and what didn’t in the teaching sessions, the notes, and exercises. Irja Haapala and Peter Emery at King's College London took over the reins when I was working on other projects and made helpful contributions to the notes and slides. Charles Zaiontz at Real Statistics kindly helped with the Wilcoxon U table, and Ellen Marshall at Sheffield Hallam University very helpfully made available the data on diet for the two‐way analysis of variance. Mary Hickson at the University of Plymouth made helpful comments on the text. Mary Hickson, Sarah Berry, and Wendy Hall at King's College London, and Charlotte Evans at the University of Leeds kindly made data sets available. Thanks to the many colleagues who said, ‘You should turn the notes into a book!’ Stephanie, Rob, Tom, Cherie, and Cora all gave me great encouragement to keep going and get the book finished. Tom and Cora deserve a special thanks for the illustrations of statisticians. The Javamen supplied the coffee. Finally, I would like to thank Sandeep Kumar, Yogalakshmi Mohanakrishnan, Thaatcher Missier Glen, Mary Aswinee Anton, James Schultz, Madeleine Hurd, and Hayley Wood at Wiley’s for bearing with me over the years, and for their support, patience and encouragement.
This book is accompanied by a companion Website:
www.wiley.com/go/nelson/statistics
The Website includes:
The ideas upon which these skills are founded – an understanding of the scientific method, an introduction to different models of scientific investigation, and the statistical tools to understand the significance of research findings – form the core of this book. Practical, worked examples are used throughout to facilitate an understanding of how research methods and statistics operate at their most fundamental level. Exercises are given at the end of each chapter (with detailed, worked solutions at the end of the book, with more examples and solutions online) to enable you to learn for yourself how to apply and interpret the statistical tools.
I have a grown‐up son and a grand‐daughter, age 6 and ¾. They are both very artistic. When I asked them to put their heads together and draw a picture of a statistician by way of illustration for this book, this is what they came up with (Figure 1):
‘What’s that!’ I cried. ‘He’s hideous!’
‘Well’, they explained, ‘the eyes are for peering into the dark recesses of the student’s incompetence, the teeth for tearing apart their feeble attempts at research design and statistical analysis and reporting, and the tongue for lashing them for being so stupid’.
‘No, no, no’, I said. ‘Statisticians are not like that’. So here was their second attempt (Figure 2):
‘That’s better’, I said.
They interpreted the new drawing. ‘Statisticians may appear a bit monstrous, but really they’re quite cuddly. You just have to become familiar with their language, and then they will be very friendly and helpful. Don’t be put off if some of them look a bit flabby or scaly. This one can also recommend a great dentist and a very creative hair‐stylist’.
Because computers can do in a few seconds what takes minutes or hours by hand, the use of computer statistical software is recommended and encouraged. However, computers are inherently stupid, and if they are not given the correct instructions, they will display on screen a result which is meaningless in relation to the problem being solved. It is vitally important, therefore, to learn how to enter relevant data and instructions correctly and interpret computer output to ensure that the computer has done what you wanted it to do. Throughout the book, examples of output from SPSS are used to show how computers can display the results of analyses, and how these results can be interpreted.
This text is unashamedly oriented toward experimental science and the idea that things can be measured objectively or in controlled circumstances. This is a different emphasis from books which are oriented toward qualitative science, where descriptions of how people feel or perceive themselves or others are of greater importance than quantitative measures such as nutrient intake or blood pressure. Both approaches have their strengths and weaknesses, and it is not my intention to argue their relative merits here.
The examples are taken mainly from studies in nutrition and dietetics. The aim is to provide material relevant to the reader’s working life, be they students, researchers, tutors, or practicing nutrition scientists or dietitians.