1.1 What does a statistic tell us?

The world can be observed in many ways, producing different kinds of information. We can read about news and look at photos in a newspaper. As a child we learned a lot about the world through fairy tales and stories, even though most of us knew they were not true. Intuition and feelings can also be knowledge, developed by the human brain as a loose summary of many experiences. The opposite of intuition are statistics which condense observations into exact numbers. The procedures for obtaining information are also exactly specified in statistics. Statistics are precisely defined information.

Not all information can be described with a statistical table. Statistics tell us about magnitudes obtained with measurements. They often describe very simple things because it is very difficult to device unambiguous measures for complicated matters. A table comprised of numbers leaves out many aspects that are necessary in forming a truthful picture of reality. Length, weight and income can be measured, whereas happiness, traits of character or socialising with other people are very difficult to measure. Measures describing these matters are often argued about.

Why should information be condensed to numbers? The main reason is that numbers are easy to compare. Another reason relating to statistics is the desire to draw a representative picture of reality. When we read a piece of news about a traffic accident caused by a drunken driver we can understand the tragedy of the event but we cannot assess how widespread such events are. In order to evaluate the risks from drink-driving to traffic and society we have to sum up all traffic accidents caused by drunken drivers.

Statistical tables are useful for capturing and describing observations that concern a larger group of people or phenomena. Tables are needed for describing, for example

A useful starting-point in interpreting statistics is that figures can only illustrate broad lines of development. Although statistical figures are often very precise, it is practical for the reader to roundthem. A useful rounding rule could be, for instance, to look only at the first three digits. If there are more than three digits in the statistics, the smallest digits are often more or less random. This is due both to measurement inaccuracies and to random variation in real phenomena.

Finnish population by age group on 31 December 2006

Source: Statistics Finland, Population Statistics

Even though these statistics indicate the respective numbers down to the accuracy of a single person, it is perfectly adequate to study how these groups compare with each other at the level of thousands of persons.