Prior to elections, it is difficult to avoid the polls that are being made. But a new study shows that the media often report errors by not taking into account the error margin.

“Then you do journalism that does not cover the material,” says Per Oleskog Tryggvasson, PhD student in journalism, media and communication at Gothenburg University.

Per Oleskog Tryggvasson, together with Jesper Stromback, has conducted a study that investigates how media reports on opinion polls. A mistake he believes the media do the report about very small changes without mentioning to the reader that the change is within the error margin.

“You can see a change that may not be there,” says Per Oleskog Tryggvasson.

The margin of error

When something is within the error margin, there are so small differences that it is not possible to determine if they occurred by chance or not. Despite this, media often report changes in a party’s opinion figures, even though these changes are within the margin of error. The worst mistake that Per Oleskog Tryggvasson thinks the media are doing is when they draw conclusions from this.

“That’s what I think is most blameful, it’s just as easy to explain random noise,” says Per Oleskog Tryggvasson.

The study covers 121 Swedish news articles written during the last two elections. In 44 percent of cases where news writers compared opinion polls, they have wrongly calculated the error margin. Most often, they have reported a change within the error margin, but without telling it.

Per Oleskog Tryggvasson has only looked at the last two years and has no examples from this year.

However, an example of SVT Science this year, when the Environmental Party fell by 0.5 percentage points and thus ended up under the 4% barrier. Then many newspapers wrote about how the environmental party ended up under the barrier. However, since the change was within the margin of error, it is not possible to say that something has actually happened since the previous opinion poll.

Lack of source criticism

Per Oleskog Tryggvasson believes that this reporting makes for a certain journalistic source criticism. When a batch increases or decreases within the error margin, there is no point in trying to explain why, because the increase simply can not be guaranteed.

– Then you get a false sense that there is something going on in the public opinion. Then there will be a positive or negative reporting about the parties that may not have any bearing in reality, “says Per Oleskog Tryggvasson.