Definition of Margin of Error

Many concepts have a double dimension, the colloquial and technique. This is what happens with the tag "margin of mistake".

In its everyday sense

If someone says they have "no margin for error" in relation to a draft, you are indicating that you cannot make any mistakes for any reason. On the contrary, if it says that "it has a small margin of error" it communicates that a possible mistake does not have serious consequences. We must bear in mind that the meaning of margin depends on the context of the language in which it is used.

In statistics

The statistics is a tool mathematics that allows to establish measurements on any type of field. With it, it is possible to know specific data on aspects of different nature, such as demographics, voting trends, diseases and a long etcetera. An important piece of information for statistical studies is to establish the limit of error or margin of error for a sample.

The margin of error is, in short, the largest possible error in relation to some numerical data

In this sense, there are two types of error margins, the absolute and the relative. The first refers to the exact measurement of something. In this way, if an object is actually 15 cm but when we measure it we make a mistake and determine that it measures 14.9 cm, the absolute margin of error will be 0.1 cm (this implies the subtraction between the actual measurement of the object and the measurement made of the same).

The relative error is stated as follows: the absolute value divided by the actual value. Continuing with the previous example, the absolute value is 0.1 cm and the actual value is 15 cm, so the relative error will be as follows: 0.1: 15, which is equal to 0.00666 cm.

The statistical margin of error in sociological surveys

These types of calculations are widely used in elaboration of surveys in which the opinions of citizens about some aspect of reality are measured, for example their assessment of a candidate or a proposal politics. Although statistics is a neutral and objective tool, in practice the information it provides does not always correspond to the reality of the facts.

In this way, the following question should be asked: why do sociological statistical measurements present so many errors? This question has two possible answers:

1) some statistics have been "cooked", so their final results do not adequately express what they intend to measure and

2) the people surveyed do not always tell the truth, so their answers do not allow us to know the reality of an issue.

Photos: Fotolia - get4net - euroneuro