Key Concepts About Data Normality

The distribution of environmental chemical data often lacks normality. The normal or nearly normal distribution should have a skewness value close to 0.  If skewed, the distribution could be positively skewed (right skew: mean > median), or negatively skewed (left skew: mean < median). The asymmetries of environmental chemical data usually show right skewness.

It is important to check the normality of data distribution because the data normality will determine whether a parametric (which assume a normal distribution) or non-parametric hypothesis test is appropriate, and whether it is necessary to transform the variable of interest.