CORRELATIONS
What is a correlation?
- A relationship between two variables
Correlational design
- In a correlational design, there are no independent or dependant variables, but co-variables
- Participant provides data for both variables
When do you use a correlation?
- To test a hypothesis about a relationship between two variables
- When looking for a relationship that would be unethical to manipulate for an experiment
How do you use correlations?
- Decide what co-variables you will be measuring, and operationalise your variables.
- Measure each participant on both co-variables
- Plot the values on the scatter graph to see if there is a relationship
Correlation coefficient
- The correlation coefficient is measured from -1 to 1
- A correlation can be positive, negative or no correlation. The larger the number, the stronger the correlation
Positive correlation
- Between 0 to + 1
- For example, the taller a person is, the heavier they are likely to be
- As one variable increases, the other also increases
Negative correlation
- Between 0 and - 1
- For example, the more alcohol you drink, the less you are able to remember
- As one variable increases, the other decreases
No correlation
- There is no relationship between the variables
- Close to 0
- For example, shoe size + IQ
Difference between correlations and experiments
- Experiments manipulate IVs, correlations don't
- In correlations the variables are simply measured
- Checks to see if two sets of numbers are related; in other words, are the two sets of numbers corresponding in some way.
Strengths
Weaknesses
Correlations only identify a link; they do not identify which variable causes which. There might be a third variable present which is influencing one of the co-variables, which is not considered.
Eg. stress might lead to smoking/ alcohol intake which leads to illness, so there is an indirect relationship between stress and illness.
Correlations are very useful as a preliminary research technique, allowing researchers to identify a link that can be further investigated through more controlled research.
Can be used to research topics that are sensitive/ otherwise would be unethical, as no deliberate manipulation of variables is required.