How to Test If Your Data Are Normally Distributed

Many of the statistics used by analytical chemists are based on the assumption that the data are normally distributed. Sometimes, in the case of the standard deviation of a mean, the data tend to the normal distribution by theory, but most of the time we cross our fingers and use the normal statistics anyway. Although there is no method that can take three or four values and make a sensible statement about their distribution, there are tests for the assumption of normality for sets of data of at least 10 tests. 

A useful graphical procedure to test the normality of a set of data that can be implemented in a spreadsheet is the Rankit method. The procedure, shown in example 3.1 below, is as follows for n data  

Sort data into increasing order of magnitude. 

Write the cumulative frequency of each value, that is, how many data have an equal or lesser value. 

Calculate the normalized cumulative frequency = cumulative frequency/( + 1). 

---