1.4 Scales of measurement

Statistics are based on the use of different kinds of measurement scales. The following introduces the main types of scale:

Although the peoples of the Nordic countries can be represented in one statistical table, it is not possible to calculate their average or mean, for example, because nationality is a qualitative observation. The scale used here is a nominal scale, which can only be used to identify the biggest category, i.e. the most common value. In this case the biggest category is "Swedish".

Nordic peoples (mean population in 2006, 1,000 persons)

	Finnish	5 255
	Swedish	9 048
	Norwegian	4 640
	Danish	5 427
	Icelandic	300

Source: Nordic Statistical Yearbook 2007, Nordic Council of Ministers

What is the average educational level in Finland? We can calculate the arithmetic mean if education is translated into years of study. On the other hand, the number of years of study does not tell us very much about courses completed or about the actual level of learning. Indeed, education is often described using a scale which makes a crude distinction between a high and low level of education. A distinction is made between primary, secondary and tertiary education. This is known as an ordinal scale.

Finnish population over 15 by level of education in 2006 (1,000 persons)

Source: Statistics Finland, Education Statistics 2007

It is not possible to calculate the arithmetic mean from an ordinal scale. Indeed, the level of education in Finland could be described by saying that 59% of the Finnish population aged 15 or over have completed a secondary degree or higher.

Compared to the ordinal scale, the interval scale has the added feature that the different alternatives are thought to be equidistant from one another. A typical interval scale is a thermometer. It is possible to calculate the arithmetic mean from this kind of a scale, but there is only limited scope in comparing the values. The average temperature is often used to describe the climate of a region, but it is not possible to say that it is twice as warm in some region as somewhere else.

Different kinds of attitude scales are often interpreted as interval scales as well. In this case attitudes are measured with questions where the respondents are to say what they think about a certain statement on a scale from, say, "fully agree (1)" through "agree to some extent (2)" and "neither agree nor disagree (3)" to "disagree to some extent (4)" and "fully disagree (5)".

The Finnish consumer confidence indicator is calculated in a similar way. It summarises people's answers to questions about the country's and their own economic situation. As such, the measured value tells very little, but when compared to its long-time average (in July 2008 it was 6.5) allows conclusions to be made of whether the economic mood is optimistic or pessimistic.

Finnish people's assessment of the country's economic situation in one year's time and the average temperature in Helsinki, July 2008

Sources: Statistics Finland, Income and Consumption, Finnish Meteorological Institute, 2008

Where money values are concerned,it is customary to use a ratio scale. A ratio scale differs from an interval scale in that it has a real zero point. This means that it is also possible to make relative comparisons. For example, it can be said that incomes in a certain group are twice as high as in another group.

Finnish income earners by income range in 2006

Source: Statistics Finland, Income and Consumption 2008