Unit 1 Research Methods (can be examined in Unit 1&2) Correlations Unit 1 Research Methods (can be examined in Unit 1&2)
There is a positive correlation between shoe size and reading age
Just because there’s a correlation, this does NOT mean that one variable CAUSED another Explore the “Spurious Correlations” website
Even if there is a cause and effect relationship, have you got cause and effect the right way around?
What is a correlation? Where two variables are plotted against each other to see if there may be a relationship Pairs of data are plotted on a scattergraph Predictions can be made based on the data you have
Types of correlation
Just because there’s a correlation, this does NOT mean that one variable CAUSED another A cause and effect relationship cannot be established as the data was not collected in an experiment and extraneous variables have not been accounted for. The relationship could: a coincidence have a 3rd related variable be cause and effect
Extension A statistical test, e.g. Spearman’s rank correlation test or Pearson’s test can be used to mathematically decide if there is a correlation. These tests result in a number between 1 and -1 (1 = perfect +ve correlation, 0=no correlation, -1= perfect –ve correlation) A ‘critical value’ (cut off point) can be looked up depending on sample size
Evaluating the use of correlations Can be used to look for a relationship which a researcher can then do an experiment to try and establish causal relationship Can be used when it would be unethical or impossible to carry out an experiment, e.g. with diseases Cannot be used to determine cause and effect relationship Needs a very large sample size
Text book pages 88-91