the Statistician
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’.
FIGURE 1 A SCARY STATISTICIAN.
Using Computers
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.
FIGURE 2 A FRIENDLY STATISTICIAN.
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.
CHAPTER 1 The Scientific Method
Learning Objectives
After studying this chapter you should be able to:
Describe the process called the scientific method: the way scientists plan, design, and carry out research
Define different types of logic, hypotheses, and research designs
Know the principles of presenting data and reporting the results of scientific research
1.1 KNOWING THINGS
What can I know?
—Immanuel Kant, philosopher
The need to know things is essential to our being in the world. Without learning we die. At the very least, we must learn how to find food and keep ourselves warm. Most people, of course, are interested in more than these basics, in developing lives which could be described as fulfilling. We endeavour to learn how to develop relationships, earn a livelihood, cope with illness, write poetry (most of it pretty terrible), and make sense of our existence. At the core of these endeavours is the belief that somewhere there is the ‘truth’ about how things ‘really’ are.
Much of the seeking after truth is based on feelings and intuition. We may ‘believe’ that all politicians are corrupt (based on one lot of evidence), and at the same time believe that people are inherently good (based on a different lot of evidence). Underlying these beliefs is a tacit conviction that there is truth in what we believe, even though all of our observations are not consistent. There are useful expressions like: ‘It is the exception that proves the rule’ to help us cope with observations that do not fit neatly into our belief systems. But fundamentally, we want to be able to ‘prove’ that what we believe is correct (i.e. true), and we busy ourselves collecting examples that support our point of view.
Some beliefs are easier to prove than others. Arguments rage about politics and religion, mainly because the evidence which is presented in favour of one position is often seen as biased and invalid by those who hold differing or opposing points of view. Some types of observations, however, are seen as more ‘objective’. In science, it is these so‐called objective measures, ostensibly free from bias, that are supposed to enable us to discover truths which will help us, in a systematic way, to make progress in improving our understanding of the world and how it works. This notion may be thought to apply not only to the physical and biological sciences but also the social sciences, and even disciplines such as economics. There are ‘laws’ which are meant to govern the ways in which things or people or economies behave or interact. These ‘laws’ are developed from careful observation of systems. They may even be derived from controlled experiments in which researchers try to hold constant the many factors which can vary from one setting to another, and allow only one or two factors to vary in ways which can be measured systematically.
It is clear, however, that most of the laws which are derived are soon superseded by other laws (or truths) which are meant to provide better understanding of the ways in which the world behaves. This process of old truths being supplanted by new truths is often a source of frustration to those who seek an absolute truth which is secure and immutable. It is also a source of frustration to those who believe that science provides us with objective facts, and who cannot therefore understand why one set of ‘facts’ is regularly replaced by another set of ‘facts’ which are somehow ‘more true’ than the last lot. It is possible, however, to view this process of continual replacement as a truth in itself: this law states that we are unlikely1 ever to find absolute truths or wholly objective observations, but we can work to refine our understanding and observations so that they more nearly approximate the truth (the world ‘as it is’). This assumes that there is in fact an underlying truth which (for reasons which we will discuss shortly) we are unable to observe directly.2
Karl Popper puts it this way:
We can learn from our mistakes.
