If YouTube algorithms are radicalizing users, the data doesn’t show it

We have all seen It happens: Watch a video on YouTube and your recommendations change, as if Google’s algorithms think the topic of the video is your life’s passion. Suddenly all of the recommended videos (and probably plenty of ads) shown to you are on the topic.

Most of the time, the results are funny. But there was a constant flow of stories about how the process radicalized people, sending them into deeper and deeper burrows until all of their viewing was dominated by fringe ideas and conspiracy theories.

A new study released Monday examines whether those stories represent a larger trend or are just a collection of anecdotes. While the data cannot rule out the existence of radicalization online, it certainly suggests that it is not the most common experience. Instead, it seems marginal ideas are just part of a larger community that is self-reinforcing.

Big data

Normally, the challenge of doing a study like this is to get data on the video viewing habits of people without their knowledge and potentially modify their behavior accordingly. The researchers got around this problem by obtaining data from Nielsen, which simply tracks what people are watching. People allow Nielsen to track their habits and the company anonymizes the resulting data. For this study, the researchers obtained data from more than 300,000 viewers who collectively watched more than 21 million videos on YouTube during a period from 2016 to the end of 2019.

Most of these videos had nothing to do with politics, so the authors used the literature to identify a large collection of channels that previous research had labeled based on their political orientation, ranging from far left to centrist. through the far right. To this list, the researchers added a category they called “anti-alarm clock.” While not always overtly political, a growing collection of channels focuses on “opposing progressive social justice movements”. While these channels tend to align with right-wing interests, ideas are often not presented that way by the hosts of the videos.

In total, the channels ranked by researchers (just under 1,000 of them) accounted for only 3.3% of total video views during this period. And those who watched them tended to stick to one type of content; if you started watching left-wing content in 2016, you were likely to watch it again when the study period ended in 2020. In fact, based on the amount of time spent on video, you were very likely to watch more of that content in 2020, possibly due to the controversy of the Trump years.

(The exception to this rule is far-left content, which was viewed so infrequently that it was impossible to see statistically significant trends in most cases.)

Almost all types of non-margin content also experienced growth over this period, both in terms of total viewership and time spent watching videos on those channels (with the exception of far-left and far-left content. far right). This finding suggests that at least some of the trends reflect an increasing use of YouTube as a substitute for more traditional broadcast media.


Since viewers have mostly watched one type of content, it is easier to see them as forming separate groups. The researchers tracked the number of people in each group, as well as the time they spent watching videos over the four-year period.

All the while, the dominant left was about as important as the other groups put together; it was followed by the centrists. The ruling right and the anti-revival started the period at roughly the same level as the far right. But they all showed different trends. The total number of far-right viewers has remained stable, but the time they have spent watching videos has increased. In contrast, the total number of mainstream right-wing viewers increased, but the time they spent watching was not much different from that of the far-right.

Anti-clock viewers showed the highest growth rate of any group. At the end of the period, they spent more time watching videos than the centrists, although their population remained smaller.


Source Link