Correlations

• what is a correlation analysis?

  • A statistical technique for analysing the relationship between two sets of numerical scores (co-variables).
  • Correlations are represented using scattergrams
• what is an experiment?

  • In an experiment, an IV is manipulated to measure the effect on a DV.
  • Since, in an experiment, there is a deliberate change in the IV, it is possible therefore to infer that the IV causes any observed changes in the DV.
• what does a correlation involve?

  • A correlational analysis does not involve any manipulation of variables: ‘deliberate change’.
  • It is therefore not possible to establish a ‘cause and effect’ relationship between (co-)variables with a correlational analysis: we ‘can’t show causation’.
• what is a correlation co-efficient?

  • A correlation coefficient has a maximum value of +1 (perfect positive correlation) and a minimum value of -1 (perfect negative correlation).
  • The coefficient tells us how closely the co-variables are related.
• what are strengths of correlation analysis?

  • Easy to see a connection.
  • Easily replicable: repeat findings.
  • Cheap (economical) (use of secondary data!?) and quick to do.
  • A useful ‘primary tool’ to help research (what does this mean?)  (What might be triggered if sufficient relationship between co-variables is found!?)
  • Used when it would be impossible/ impractical or unethical to carry out an experiment i.e. can’t force people to smoke to see if there is a link causing cancer.
• what are weaknesses of correlation analysis?

  • Methods of measurement may lack validity.
  • Large amount of data needed to be informative.
  • Don’t know which of the two variables cause the correlation: a common problem is misinterpretation.
  • Third variable problem (intervening variables).