Random data from some process

Use this diagram to show that a data set can seem to have an outlier, clusters or skewness even when the data are genuine measurements from a process with no such characteristics. Randomness has not been formally discussed yet but you could mention that these features may appear 'by chance'. The Another sample button generates random data from a normal distribution (mean = 4.5, st devn = 1.75).

Take several samples, looking for the most 'extreme' evidence of an outlier. (Click Remember as you do this to retain the most extreme example so far.)

Similarly, look for the most 'extreme' evidence of clusters.

Finally, look for the most 'extreme' evidence of skewness.