Bubbles in front of screens.

New study rejects the idea of filter bubbles

| 11 March 2021

There is little evidence that filter bubbles really exist or have any relevance to the increased polarisation in society. This is stated by Peter Dahlgren in a new Nordicom Review article. The study dismisses the idea of filter bubbles from both a psychological and technical point of view.

The impact of filter bubbles on democratic prosesses is something that has been discussed a lot in recent years. The idea of filter bubbles comes from journalist Eli Pariser and is about how algorithms on the Internet filter information based on previous searches. This would ultimately mean that we are surrounded solely by information in line with our existing opinions – sort of like a bubble.

In a new article published in Nordicom's scientific journal Nordicom Review, Peter Dahlgren, a researcher at the University of Gothenburg, has examined the existence of filter bubbles and their significance in society at large.

- Surely selective media use can have an impact, but to go so far as to say it has an effect on society at large – meaning that people do not see challenging information, and this, in the end, leads to increased polarisation or a collapse of democracy – there is no indication of, Dahlgren maintains.


People take part in information that both confirms and challenges

First, this study rejects the idea of filter bubbles from a psychological point of view. The idea that people only take part in information confirming their worldview is simply not true, accoring to Dahlgren

- We do indeed see that people take part in information that corresponds to their opinions to a slighty greater extent, but it is a difference of small degree.

Precisely these differences in degrees tend to dissappear when the discussion is about fitler bubbles.

- Namely, that people only choose information corrosponding to their worldview and discard other information, has been prominent in the public debate. But it is in fact very difficult to find people who truly do it. Most people take part in information that confirms and challenges their own perceptions, says Dahlgren.


Algorithms are not mindreaders

The idea of filter bubbles is even rejected from a techical point of view. It is based on a general notion that algorithms can read our thoughts, which is also not true.

- There is a widespread misconception about the force of algorithms. They are no smarter than the data we provide them. These so-called recommendation algorithms are based on the behaviour of a large group of people – not on an individual person. These algorithms assume that you will make the same choice as the group, but they cannot read your mind, affirms Dahlgren.


What a filter bubble is is unclear

The most important arguement against the idea of filter bubbles is the fact that it is not clear what a filter bubble is. The idea is often defended by the existence of algorithms on the Internet and by the fact that people are polarised, but according to the study, there is no causal link.

- If you say something exists, you have to be specific about what it is. In this case, one has assumed that filter bubbles are real and we have to do somthing about it – even though there is no evidence of them.

- One can say that the filter bubble is a product of the imagination, says Dahlgren.


Read more


Peter Dahlgren recently received his doctorate for his dissertation: ”Media Echo Chambers: Selective Exposure and Confirmation Bias in Media Use, and its Consequences for Political Polarization”