Survey research and design in psychology/Tutorials/Correlation/Effects of outliers

From Wikiversity
Jump to navigation Jump to search
Effects of outliers
Gnome-settings-background.svg View the accompanying screencast: [1]

The purpose of this exercise is to explore the impact of bivariate outliers on linear lines of best fit and correlations:

  1. Try out the following applets: (if the device you are using doesn't have the latest version of java installed, these applets may not work)
  2. Create some distributions and try out the options noting:
    • correlation
    • line of best fit
    • mean X and Y lines
    • moving data points
  3. Explore the impact of an outlier by adding data points or moving data points around until you can answer:

1. Where would you add an outlier in order to maximise the correlation?

Answer

As far to the ends of the line of best fit as possible

2. Where would you add an outlier in order to minimise the correlation (or make the correlation go as far as possible in the opposite direction)?

Answer

Place the outlier as far as possible to the ends of a line which would run perpendicular to the line of best fit, crossing at the mean for X and the mean for Y

3. Where would you add a point which does not change the correlation?

Answer

On the mean for X and Y