What an alternative hypothesis claims
Definition
Alternative hypothesis
A precise, testable prediction that there will be a difference between conditions or a relationship between variables. It is sometimes called the research or experimental hypothesis.
- An alternative hypothesis predicts that there will be a difference between the conditions, or a relationship between the two things measured.
- It says the independent variable, and not chance, lies behind the pattern in the results.
- You may also see it called the research hypothesis or the experimental hypothesis, and the three names mean the same thing.
Building one from a scenario step by step
- Work from this scenario: your class wants to find out whether phone notifications slow people down on a ruler drop task.
- Find the independent variable and name its two conditions, which here are doing the task with notifications switched on and doing it with the phone switched off.
- Find the dependent variable and name the exact measure, which here is the distance in centimetres the ruler falls before it is caught.
- Decide whether you can justify saying which way the result will go, or whether you should only predict that the two conditions differ.
Example
- The finished sentence: students catch the ruler after a shorter fall in centimetres when their phone is switched off than when notifications are switched on.
- Everything that matters is in it: the two conditions, the measure and the direction.
Directional and non-directional versions
Definition
Directional hypothesis
An alternative hypothesis that states which way the results will go, for example which condition will score higher or whether a correlation is positive or negative.
- A directional alternative hypothesis names which way the result will go, such as which condition scores higher or which is faster.
- A non-directional alternative hypothesis predicts a difference without saying which condition comes out on top.
- Choose the directional version when there is a clear reason to expect one particular outcome, and the non-directional version when there is not.
Example
- Directional: students catch the ruler after a shorter fall in centimetres when their phone is switched off than when notifications are switched on.
- Non-directional: there is a difference in the distance in centimetres the ruler falls before it is caught between students whose phone is switched off and students whose notifications are switched on.
Writing one for a relationship rather than a difference
- Some studies do not compare two conditions at all, but measure two things in the same people and look for a link between them.
- The alternative hypothesis then predicts a relationship, so use the word relationship rather than difference.
- The directional version says which way the link runs, using ordinary words such as the more and the higher.
Example
- Non-directional: there is a relationship between the number of minutes spent revising and the mark scored out of twenty on a class test.
- Directional: the more minutes students spend revising, the higher the mark they score out of twenty on a class test.
Repairing a weak hypothesis
- Music affects memory is too vague to test, because neither variable has been pinned down.
- Repair it as: students recall fewer words out of twenty when music is playing than when the room is silent.
- I think revising with flashcards might work better is hedged and written as an opinion.
- Repair it by stating it flatly: students who revise with flashcards score higher on a twenty mark test than students who re-read their notes.
Common Mistake
- Do not name the topic without the measure, because the sentence cannot be tested as it stands.
- Do not add a direction you cannot justify from the scenario or from what the question has told you.
- Do not describe the aim of the study instead of predicting its result.
Turning this into marks
- Operationalise both variables inside the one sentence you write.
- Lift the scenario's own measure, such as out of twenty or in centimetres, straight into that sentence.
- The mark most often lost comes from giving a non-directional hypothesis when a directional one was asked for, or the other way round.
Self review
- What does an alternative hypothesis predict?
- Rewrite sleep affects concentration as a testable alternative hypothesis.
- When would you choose a non-directional version rather than a directional one?
- What is the difference between an alternative hypothesis for a difference and one for a relationship?
