ABSTRACT
Now that we have established why description is important, let us look more closely at the sorts of enquiry we can undertake with descriptive analysis.
If we are confronted with a mass of data, it overloads us; we cannot assimilate it. UK universities, for example, are assessed on more than ten performance measures each year. Some have an equity dimension (per cent of students from schools and colleges in the state sector, per cent from lower socio-economic groups, per cent from low participation neighbourhoods). Others profile mode of attendance, type of degree, continuation rates (that is, failure), employment outcome and other university activities such as research. The quantities of data are enormous. Our immediate response is to simplify the data. One way of doing this is to compress the indicators into a single indicator for each institution. This would enable any institution to be compared with any other. This is what many newspapers in the UK do each year. The Higher Education Statistics Agency says that ‘[no] meaningful league table could fairly demonstrate the performance of all higher education institutions relative to each other’ (see www.hesa.ac.uk/index.php/content/ view/1168/141/, accessed August 2008). Notwithstanding this counsel of perfection, the rest of the world cannot handle a data matrix of 169 rows (the number of universities) and a number of columns that probably stretches into the hundreds. Even broken up
shall meet some of them in Chapter 13. What Anghileri and her collaborators did was show a marked difference in performance just by describing the data mathematically. And this is
what this chapter is all about, using statistical summaries to draw out the characteristics of data sets so that, by inspection, similarities and differences can be identified and assessed.
