- A correlation is the measurement of extent in which pairs of values on two variables change together.
- They describe the relationship between 2 variables in statistical terms. Not a research method.
- No IV or DV but described in co-variables because they both vary & measured. Neither is set or controlled.
- Strength of a correlation described in co-efficiency ranging from +1 to -1 (perfect + perfect -)
- With a co-efficiency less than 0 describing a – correlation and above 0 + (0.3 as weak, 0.3-0.7 moderate and above 0.7 strong)
- Data is collected from various research methods and experiments. Can be analysed to see if relationship between 2 Co-variables.
- Both measures usually come from the same researcher and tend to use Interval or Ordinal data
- Investigate a relationship plot variable A on one axis & variable B on the other (scatter graph) & a straight line for line of best fit. The gradient is the same as co-efficient.
Strengths | Weaknesses |
Ethical – unethical to study number of cigarettes smoked & lung capacity – allow researcher to investigate things without setting up a study to encourage & control ppt smoking
Validity – both sets of data are provided by the same person; each ppt acts as their own control –individual differences will not affect results Practical –easy to carry out with input from a researcher (e.g nothing needs to be set up artificial as uses pre-existing data) – correlations can decide if new research should be generated Reliable – involve an easily replicated method (self-report questionnaires) proved as standardized procedure – can repeat to test for consistencies |
Internal validity – impossible to claim one co-variable causes the other as there could be a third, unknown variable – misunderstanding & erreous conclusions. Unable to determine cause & effect
Validity – methods used to gather data usually self-report such as questionnaires – demand characteristics or socially desirable answers may be given lacking validity Validity –correlations may use secondary data e.g hospital records which may be out-dated or have been subject to bias interpretation – invalid & data flawed. |