The prediction a study is tested against
Definition
Hypothesis
A clear, testable prediction of what a study expects to find, written before the data is collected.
- A hypothesis is a precise prediction of what a study will find, written down before any data is collected.
- A study sets out two rival predictions, and the null hypothesis is the one saying that nothing is going on.
Key Idea
- The null hypothesis is the claim your results have to argue against.
- If the results show a clear difference or relationship you reject the null hypothesis, and if they do not, it stands.
What a null hypothesis actually says
Definition
Null hypothesis
A prediction that there will be no difference between the conditions, or no relationship between the variables, and that any pattern in the results is down to chance.
- A null hypothesis predicts that there will be no difference between the conditions, or no relationship between the two things measured.
- It adds that any small difference which does turn up in the numbers is down to chance rather than the independent variable.
- It is not a prediction that the opposite will happen, and it is not a prediction that the study will go wrong.
Example
- Comparing two conditions: there will be no difference in the number of words recalled in silence and with music playing.
- Looking for a link: there will be no relationship between the number of hours slept and the mark scored on a spelling test.
Putting the variables in properly
Definition
Operationalisation
Defining a variable clearly in terms of exactly how it will be manipulated or measured, so it can be tested in a precise, repeatable way.
- To operationalise a variable is to say exactly how it is set up or exactly how it is measured.
- Inside a null hypothesis that means naming the two conditions of the independent variable and naming the score, count or time that makes up the dependent variable.
- A null hypothesis that names only the topic is the most common way of all to lose this mark.
Example
- Not operationalised: there will be no difference in memory between the two groups.
- Operationalised: there will be no difference in the number of words recalled out of twenty between students who revise in silence and students who revise with music playing.
A sentence frame you can fill in
- For a study comparing conditions, begin with there will be no difference in, then the dependent variable with its measure, then between, then the two conditions.
- For a study looking for a link between two measured things, begin with there will be no relationship between, then the first variable with its measure, then and, then the second variable with its measure.
- Fill the frame straight from the scenario, using the numbers and the wording the scenario has already given you.
Checking your own null hypothesis
- Read it back and ask whether it clearly says no difference or no relationship.
- Ask whether both variables appear, each with the way it is set up or measured attached to it.
- Ask whether it is a statement rather than a question, and one sentence rather than a paragraph.
Common Mistake
- Do not write that there will be a difference and call it a null hypothesis, because that is the rival prediction.
- Do not soften it with might or probably, because a hypothesis is stated flatly.
- Do not name a direction in a null hypothesis, because saying which group will do worse is already predicting a difference.
Turning this into marks
- State no difference or no relationship in one flat sentence, with both variables operationalised.
- Lift the scenario's own measure into your sentence, such as the number of words recalled out of twenty.
- The mark most often lost comes from describing the aim of the study instead of stating the prediction.
Self review
- What does a null hypothesis predict?
- Why is there will be no difference in concentration not good enough?
- Write a null hypothesis for a study comparing test scores after revising with a friend and after revising alone.
- What does a null hypothesis say caused any small difference that does appear in the results?
