Yes, a biased sampling method that selects from a subgroup of a population would likely reduce the spread of the data, have fewer outliers and lead to positive kurtosis. Conversely, small sample sizes, subjective outcome measures and poor measurement approaches might all lead to either a more widely spread population or would suffer more from the impact of outliers.
OK, we’ve looked at shape, let’s move on to the statistical descriptions of dispersion. First, an important concept.
Standard deviation (SD) is the most commonly used measure of dispersion but which of the following is true if you went from having 1,000 samples to 10,000?
LaNts and Laminins 