Geographic Information Systems: An Introduction, 3rd Edition

3.4 Levels of Measurement

Attributes may be grouped in different ways. A common method of division, as previously mentioned, is to differentiate between qualitative and quantitative attributes. Qualitative attributes represent the type of data, such as type of vegetation or type of language. We can therefore say that qualitative data is the name of the attribute. On the other hand, quantitative attributes are related to quantities and are normally represented by figures, such as number of inhabitants in a city and the acreage of a property. In addition, data that can be ranked in a series, such a good, better, best, can be placed in a separate group known as ordinal data. In cartography is it usual to divide data into qualitative, quantitative, and ordinal.

Attributes may also be ranked in four levels of accuracy: ratio, interval, ordinal, and nominal (Figure 3.10). The most accurate are ratio or proportional attributes, such as length and area, which are measured in respect to a fixed origin or starting point and on a continuous scale at fixed intervals. Interval data comprise numerical data that are less accurate. An interval scale places the data on a number scale and, in principle, with a randomly selected starting point and arbitrary interval. Temperature, for example, can be measured in Celsius and Fahrenheit and years according to different calendars. A segment of these number lines without the origin and without definition of the unit of measure gives only the intervals ...