Qualitative data – what people say, descriptive material of one kind or another – are not intrinsically numerical in character. Ultimately, of course, no data are numerical: measurements of physical phenomena like temperature and air pressure are reflections of more fundamental processes but on a precise interval scale. And the familiar, everyday, measurement of time can seem almost natural: but of course all of these measurement scales are man-made systems related to the physical world.
The applications of number to qualitative data are much more limited and we have to bear these limitations in mind but they are still useful in the sense that they can add something to meaning, not least a level of precision.
There is a distinction we need to make clear. Descriptive statistics – things like averages (known as means in statistical science), ranges and frequencies enable you to describe in a tidy, summary format, usually as tables or figures, rather a lot of data. For example, if you have used recording schedules (Chapter 11) with 150 people, you might want to show the frequencies of the different ageranges in a histogram (Figure 20.1). This distribution would be tedious (and confusing) to describe in words but can be seen almost at a glance.
Inferential statistics are those which enable you to make an inference about differences or relationships between two sets of data. Here we shall deal with two ways of ordering data involving these kinds of statistical analyses: putting data into categories; and putting data into ranks.
When analysing data into categories for a number of different respondents you can do it in two ways as described in the previous chapter: on one