Communicating statistics using maps

Communicating statistics using maps

A map is a powerful communication tool that provides an intuitive and simplified picture of a set of data. However, maps can also be seductive and highly misleading, and it is important for readers to be aware of some fundamental factors.

When choosing a cartographic form of expression, there are several choices to make that entail advantages as well as disadvantages. The most commonly used cartographic presentation is known as the choropleth map. In such maps, each area (the interactive atlas uses hospital referral areas (health trust areas)) is assigned a shade of colour depending on which class/group the data value falls within. The advantage of this is that it is easy to identify the areas described, while the disadvantage is that large hospital referral areas will dominate visually, while smaller areas get far less attention.

Grouping or classification is another factor that one should be aware of when interpreting maps. All the variables presented are in principle continuous variables (rates are at interval/ratio level). In order to present this in a meaningful form on a map, the information must be simplified – we need to classify (group) the information.

An unfortunate effect of such generalisation is that it may conceal big differences between data values in the data set and/or emphasise minor differences between variable values on the map. In order to counteract this, a method known as Jenks natural breaks has been chosen for the interactive atlas, and four group are used (in some cases fewer). This classification method uses an algorithm to maximise homogeneity within each group as well as heterogeneity between groups, which means that the groups are defined so that we get the biggest possible differences between groups and the smallest possible differences between the hospital referral areas within each group. The number of hospital referral areas in each group will therefore vary between patient samples. In some cases, one group can consist of a single hospital referral area and another group of a large number of areas, while for other patient samples, the hospital referral areas can be evenly distributed between groups.

The interactive atlas includes a table and a bar chart that are intended to be used alongside the map to interpret the variations observed.

Maps are intended to provide a simplified picture of reality, but maps are also designed on the basis of a number of subjective choices. These choices influences how other people perceive reality. Just as with figures and statistics, one should take a critical approach when interpreting maps.

Therefore, be aware of the rhetorical possibilities of the map.