Every time a statistic or a Gallup poll is published, journalistic power is at work.

Statistics and Gallup polls can at first glance seem like a reliable and objective way for the media to convey information. Still, the worn saying “lies, damn lies and statistics” encapsulates something essential about statistics and averages. When a newspaper conducts a questionnaire that receives answers from the readers, and the results are published sensationally as “the will of the people”, journalistic power is at work.

Already the phrasing of the question carries great importance. For example, when asked whether the respondents support the idea of higher taxes, the answer is often ‘no’. However, when they are asked whether the level of health care and other services should be maintained even with taxpayers’ money, the answer is more often ‘yes’.

When asked whether the respondents support the idea of higher taxes, the answer is often ‘no’. However, when they are asked whether the level of health care and other services should be maintained even with taxpayers’ money, the answer is more often ‘yes’.

Additionally, the result of polls is influenced by who is being asked. For example, newspapers have different readerships that are often politically profiled. The result of questionnaires that measure the political outlook of readers must therefore never be used to draw conclusions about the public as a whole.

The varying types of conclusions that can be drawn from statistics can be considered the third form of power use. For example, averages do not show how widely the results were dispersed or where the extremities lay. If half the population earns $10,000 per month and half $100 per month, the average earnings are $5,050 per month.

However, as the average doesn’t always paint a truthful picture, journalists are encouraged also to look at the median value of statistical information. The median is a value that represents the number that is in the middle when the results are put in order.

In this extreme example case, however, not even the median would paint a good picture because, with an even number of data, the median is the average of two middle numbers resulting in $5,050, the same as the average. The only truly truthful way in this case is to describe the situation with words or infographs.

Gallup polls can also be conducted in a discriminatory manner or so that they create a false feeling of power over laws and regulations. For example, if a magazine asks its readers wether a teacher who belongs to a certain ethnical, religious or sexual minority should be allowed to work in the profession or should be dismissed, this might at first glance seem a harmless use of freedom of speech.

But to ask such a question in a poll is irresponsible, as the question is basically asking the reader whether the constitutional fundamental rights of minority citizens should be limited. Posing such a question places the supposed reader to above people from the particular minority, and gives the impression that constitutions can be changed quickly and without lengthy democratic processes, just based on majority feelings or opinions.

Posing such a question places the supposed reader to above people from the particular minority, and gives the impression that constitutions can be changed quickly and without lengthy democratic processes, just based on majority feelings or opinions.

Gallup polls and statistics also legitimise popular and majority opinions. When it makes the news that the majority of the population is against immigration, for example, expressing such opinions out loud becomes more acceptable. Likewise, as an example, if a poll on party popularity is published before an election, it might indeed have a self-fulfilling effect on the matter.

Theoretically speaking, many societal phenomena such as politics and market economics are so-called “level two chaotic systems”. In more simple terms, this means that they are chaotic (very complex) but they also react to predictions made about them. This fact is by no means a reason not to publish a poll – it is just good to remember that publishing predications may later have an effect on the matter under scrutiny.