Earlier in this section, we mentioned the term normal distribution, but what do we mean by this?
For any population, the individual values will be distributed in some way across the range of possible values. The normal distribution is a mathematical model which is used to describe one way in which the values may be distributed. Data which is normally distributed will tend to have more occurrences of values towards the middle of the range, and fewer at the upper and lower ends of the range. To illustrate this, let's look at an example of some data that is normally distributed.
The manageress of a dress shop has been keeping track of what dress sizes her customers have been buying. Over a four week period, she collected the results shown in the table below:
| SIZE | 8 | 10 | 12 | 14 | 16 | 18 | 20 |
|---|---|---|---|---|---|---|---|
| Week 1 | 1 | 2 | 9 | 10 | 9 | 2 | 0 |
| Week 2 | 0 | 3 | 8 | 6 | 5 | 3 | 1 |
| Week 3 | 2 | 1 | 7 | 11 | 6 | 4 | 1 |
| Week 4 | 1 | 2 | 6 | 5 | 10 | 1 | 0 |
| Total | 4 | 8 | 30 | 32 | 30 | 10 | 2 |
Looking at the totals in the bottom row, you can see that the number of dresses sold in each size follows the pattern described above, i.e. more in the middle range sizes and relatively few at either end of the scale. However, it's not always easy to see patterns in lists of numbers, so on the next page, the same data is presented graphically.