Testing theories: Hypotheses
From Research Methods in Psychology
Contents |
[edit] SECTION OVERVIEW
IN THE LAST SECTION WE LOOKED AT PSYCHOLOGICAL THEORIES, AND TALKED ABOUT HOW THEORIES GENERATE HYPOTHESES. IN THIS SECTION WE WILL HAVE A CLOSER LOOK AT HYPOTHESES, AND CONSIDER HOW A THEORY MODIFIES A HYPOTHESIS.
[edit] What are Hypotheses?
|
Crucial Concept: A hypothesis (plural hypotheses) is a testable statement, derived from a theory, which is used to test that theory. |
To understand the relationship between theories and hypotheses we need first to consider what we mean by science. Science as a whole is not necessarily about test tubes, chemicals, smells or electronics.
|
Crucial Concept: Science is about an approach to research. Psychology is a science and uses a scientific approach. |
(It is interesting to note that sometimes students of subjects such as biology, chemistry or physics may consider that psychology is not a science, but often do not know what it is that makes a science into a science. )
Instead of thinking about any difficult, abstract psychological theories, let’s think about a theory we are all pretty familiar with - let us consider the Theory of Gravity. An apple fell from a tree, and landed on Isaac Newton’s head, or so the story goes. He wondered why the apple fell, and decided to make up a theory that explained why the apple fell. He called his theory the Theory of Gravity. As we said before, a good theory should explain what happened, make further predictions and have implications. The theory of gravity says (roughly) that there is an attractive force between objects that is dependant on the mass of the objects. The earth, of course, has a large mass and is therefore extremely attractive to objects - therefore things tend to fall towards the ground. The Theory of Gravity explains why the apple fell, makes a large number of further predictions (a theory of gravity that applied only to apples would not be a very good theory) and has a very large number of implications - for instance about the movement of heavenly bodies such as the planets and the sun.
[edit] Testing Theories by Testing Hypotheses
| Crucial Concept: A hypothesis is a testable statement derived from a theory. A hypothesis is a statement that predicts what will happen. |
Hypotheses (usually) come in two flavours. These are the null hypothesis, and the alternative hypothesis (sometimes called the experimental hypothesis). They are sometimes written as H0 (the null hypothesis) and H1 (the alternative hypothesis). Crucial concept: experimental hypothesis or (sometimes alternative hypothesis), H1, is a statement that is made by your theory.
Suppose your theory says that all swans are white. The alternative hypothesis H1 will state, “There are no black swans.”
|
Crucial concept: The null hypothesis, H0, is the negation of your theory. |
The null hypothesis, H0, will state the opposite. “There is at least one black swan.”
In this case, the theory and the hypothesis are very similar. There is really only one hypothesis that can be derived from the theory. Usually there are many hypotheses that can be derived from the theory.
If we have another look at the theory of extraversion that we covered earlier, (if you have jumped in at this section, you might want to go and have a quick read of that section), we can derive more hypotheses from these two theories. The biological theory of extraversion, which states “Introverts are more easily aroused than extraverts.” We can generate hypotheses from this theory, and for each hypothesis, we can generate a corresponding null hypothesis.
For example, we might hypothesise that extraverts have a higher pain threshold than introverts.
H1: Extraverts have a higher pain threshold than introverts. H0: Extraverts do not have a higher pain threshold than introverts.
We could hypothesise that extraverts will wake up more slowly than introverts.
H1: Extraverts are less aroused 10 minutes after waking, than introverts. H0: Extraverts are not less aroused 10 minutes after waking than introverts.
[edit] The Importance of the Null Hypothesis
At this point, you might be asking why we seem to be so interested in the null hypothesis. Surely, the alternative hypothesis, the one that is proposed by our theory, is the more important one?
Well, yes it is. However, we can never prove the alternative hypothesis. What we do instead is see if we can disprove, or reject, the null hypothesis. If we can’t reject the null hypothesis, this doesn’t really mean that our alternative hypothesis is correct, our alternative hypothesis was a bit vague anyway, but it does provide support for our alternative hypothesis. Because of this, this approach is sometimes called ‘the method of indirect support.’
In psychology, we can’t usually make definite statements about the null hypothesis. We collected data to test the null hypothesis, and it is the results of the analysis of these data that we can talk about, rather than the null hypothesis itself.
We say that it is either likely or unlikely that we would have found these results were the null hypothesis to be true. According to the results of our study, we make one of two statements, we either ‘reject the null hypothesis,’ or we ‘fail to reject the null hypothesis.’
Note that these two statements can be written in many different ways, but we should always talk in terms of either rejecting, or not rejecting the null hypothesis. Some versions that you might come across (or be tempted to write) are shown in the table below. Although they all look equivalent, you should only use the first version.
[edit] One-Tailed and Two-Tailed Hypotheses
Hypotheses come in two kinds: one-tailed and two-tailed hypotheses.
A two-tailed hypothesis states that there will be a difference between two (or more) measures, or a relationship between two (or more) measures. An example of a two –tailed hypothesis is that ‘Introverts and extraverts will differ in the amount of saliva they produce when lemon juice is placed on their tongue.’
|
Crucial concept: A one-tailed hypothesis states that there will be an effect, and that it will go in a particular direction. |
An example of a one-tailed hypothesis: ‘Introverts will salivate more than extraverts when lemon juice is placed on their tongue.’ A one tailed hypothesis is sometimes called a directional hypothesis (because it states a direction for the effect).
|
Crucial concept: A two-tailed hypothesis states that there will be an effect, but does not state the direction of the effect. |
An example of a two-tailed hypothesis: ‘There will be a difference in the amount of saliva produced by introverts and extroverts when lemon juice in placed on their tongue.’ A two-tailed hypothesis is sometimes called a ‘non-directional’ hypothesis because it specifies that there is an effect, but doesn’t specify a direction for that effect.
All hypotheses talk about variables, when a hypothesis mentions a variable, it mentions it in an abstract, psychological way. We need to move from the abstract, psychological representation, to a more concrete representation – this is known as operationalising the variables. We will have a look at what we mean by this, by looking at a study that is examining the effects of a depressant drug on complex skill performance.
[edit] Variables
|
Crucial Concept: A variable is something that can change (or vary) between people. |
Hypotheses propose that there will be a relationship between two (or more) variables, and every study in psychology involves at least one, and usually more than one variable. We will use the example of a psychologist who has a theory about the effects of depressant drug consumption on complex motor skill performance. They devise an experimental hypothesis: ‘depressant drugs will effect complex skill performance.’ There are two variables, depressant drug consumption, and complex motor skill performance, and we are interested in their relationship.
This theory is interested in complex skill performance, but we can think of many different complex skills that we could assess. In the following figure we have a circle labelled ‘Complex skill performance.’ The theory talks about ‘complex skill performance,’ and so we will label this as the theoretical level of the variable.
‘Complex motor skill’ is a rather vague, fluffy name for a variable. We need to make it a little more concrete – we need to conceptualise complex motor skill performance. The next stage is to decide what sort of complex motor skill we will measure. There are many ways that we could conceptualise complex motor skill: walking, driving, throwing darts at a dartboard, throwing balls of paper at a waste paper bin, typing. We need to decide which of these methods we will use. When we have decided, we have defined the variable at the conceptual level. The psychologist, as you can see in Figure 3 decides to conceptualise complex motor skill as ‘walking.’ Therefore, the psychologist has decided to measure ‘walking’, but this could be done in many ways. How will we actually measure the variable ‘walking’? What operations will we perform to measure this? Because we are now talking in terms of the operations that we will use, we are operationalising the variable. We can operationalise the measurement of walking in many ways, for example:
- Time taken to walk 5 metres.
- Number of times participant falls off stepping stones while crossing river.
- Number of times participant steps off a straight line.
Each of these would be a different (and equally legitimate) operationalisation of the conceptual variable ‘walking.’
This whole process of moving from a theoretical variable, to a conceptual variable to an operational variable is shown in the figure below.
|
Crucial Tip When you are describing a variable, you need to describe it at each of the levels. It is no use to say “Participants filled in the following questionnaire” without saying what the questionnaire was measuring. Similarly, it is no good to say “We measured participants’ attitudes towards hospital food,” without saying how you measured the attitude. |
[edit] Summary
In the past, theories used to be tested by collecting data that supported the theory – this was called the inductive method. Popper argued that this wasn’t a very good approach. Instead, he said we should try to prove ourselves wrong. If we fail to prove that we are wrong, this means we can have greater belief that we are right. This is called the hypothetico-deductive method.
There are two types of hypothesis: the experimental (or alternative) hypothesis – H1, and the null hypothesis – H0. It is the null hypothesis that is tested. If the null hypothesis is rejected, this provides indirect support for the experimental hypothesis.
There are two types of hypothesis: a one-tailed hypothesis, predicts that there will be an effect, and specifies the direction of the effect. A two-tailed hypothesis specifies that there will be an effect, but does not specify the direction of that effect.
Theories and hypotheses make mention of variables, but there are many different ways of conceptualising and operationalising a variable – you need to make it clear in your practical reports how you did this.
