Posted: November 18th, 2015
Biological Research Skills 5: Tests of relationships and association.
Correlation and Regression.
Both of these tests look at relationships between 2 variables but there are differences between them. Correlations just look at whether there is a relationship between the way that 2 variables vary, but does not imply a cause and effect relationship. Regression analysis attempts to explain the relationship between 2 variables by fitting a straight line (the line of best fit). Providing the straight line explains the relationship, it can be used to predict the effect of increasing the independent variable on the dependent variable. In most cases the independent variable would be the one we control in an experiment, such as temperature, and the dependent variable would be the one we are measuring, such as Optical density (OD) or growth..
Correlation.
Like the other tests covered in this module, there are both parametric and non-parametric methods for determining the correlation between 2 variables. The parametric test is the Pearson’s product-moment correlation.
As this is a parametric test it has a couple of assumptions: that the data are measured on a continuous scale and that both variables are normally distributed. If these criteria cannot be met then you must use the non-parametric Spearman’s rank correlation instead.
Example:
The following data was collected by measuring the wing span and body length of 8 birds.
Wing span: 22.2, 22.4, 23.1, 23.2, 24.0, 24.1, 24.5, 24.9
Body length: 12.5, 13.2, 13.8, 14.8, 14.9, 15.2, 15.3, 15.5
Enter the data into 2 separate columns. The next step should be to plot a scatterplot. Select ‘Graphs’ and ‘Scatterplot’ from the menu. The following box appears:
Place an order in 3 easy steps. Takes less than 5 mins.