Most UK adults have used online safety tools, but few are satisfied with reporting

Only 18 percent of UK adults who used a social media reporting tool said they were always satisfied with the outcome, even though most people know the tool exists. That is one result of a nationally representative survey by Jonathan Bright, Florence E. Enock, Pica Johansson, Helen Z. Margetts, Francesca Stevens and colleagues, who studied how people use the tools platforms offer to protect users: reporting, blocking, unfollowing and controls over feeds and replies. In the UK, new legislation makes it a requirement for large platforms to offer such tools, but, the authors say, little was known about how users engage with them.

How the survey worked

The survey ran in January 2023, using the Qualtrics platform, with 1,067 UK adults who passed quality checks. It was built to be nationally representative on age, gender and ethnicity, took about twenty minutes, and was approved by the ethics committee of The Alan Turing Institute. It asked about seven tools in three groups:

For each tool people said whether they knew of it, how much they had used it and whether the outcome was what they hoped. They were also asked about content that platforms had removed from their accounts and about appeals.

What people knew and used

Awareness of platform safety technology. Figure from the paper, CC BY 4.0.

• Awareness. More than 90 percent knew of blocking and unfollowing, and 86 percent of reporting. Awareness of preventative tools was lower: 53 percent for hiding engagement and 40 percent for deactivating a feed algorithm.
• Use. Among people who knew a tool existed, use was high: over 85 percent had unfollowed, over 80 percent had blocked, and more than 70 percent had used each preventative tool. Only about 60 percent of those aware of reporting had ever used it. In all, over 80 percent of the sample had used at least one tool.
• Satisfaction. More than 60 percent said they were always satisfied with blocking or unfollowing, but only 18 percent with reporting. For preventative tools it was about 50 percent for limiting responses and 30-35 percent for the others.
• Exposure. About 67 percent said they had seen content they considered harmful and 31 percent had been sent it directly. About 26 percent had had content removed by a platform, and about 16 percent more than once. Of those, about 48 percent appealed, and about 30 percent of those appealing got a response they regarded as satisfactory. Only 278 people in the sample had gone through an appeal, so that number is small.

Use of platform safety technology among those aware of it. Figure from the paper, CC BY 4.0.

Who uses the tools

The authors ran seven regressions for who had used each tool among those who knew about it. They stress that the models explain little of the variation.

What explains use of platform safety technology. Figure from the paper, CC BY 4.0.

• Concern about harm raised use of both preventative tools (deactivating the feed, hiding engagement) and blocking and reporting, by 18 percent (blocking) to 43 percent (reporting) per standard deviation of concern.
• Past exposure mattered more, which the authors call "learning the hard way." Those who had witnessed or received harmful content were 78 to 144 percent more likely to have unfollowed, blocked or reported. Receiving it also made people more likely to alter the feed and to hide engagement.
• Digital literacy made no difference for preventative tools, but raised blocking and unfollowing, which the authors find a little concerning for access.
• Gender. Women were more likely to report (38 percent), unfollow (40 percent, which did not hold up when outliers were removed), block (58 percent) and limit responses (43 percent), even after accounting for concern and exposure.
• Politics. Right-leaning users were more likely to report and hide engagement, against the authors' expectation, but the evidence is weak.
• Social media use. Infrequent users were less likely to use most tools. Older users were less likely to unfollow, and education had no effect.

What the authors conclude

Preventative tools should be made more visible to users, since most people reach them only after being harmed. Reporting, the authors say, is where platforms most need to be transparent and effective, and earlier studies found the same disappointment. Their limits: only people who knew a tool were modelled (the samples ranged from 330 to 981), the survey did not ask which harms people had seen, the digital literacy items were simple, and it covered only adults, not children and teenagers.

Sources

• Authors of the original: Jonathan Bright, Florence E. Enock, Pica Johansson, Helen Z. Margetts and Francesca Stevens. "Understanding engagement with platform safety technology for reducing exposure to online harms," arXiv:2401.01796 (2024), licensed under CC BY 4.0. Changed here: rewritten in hubnx's own words, with the paper's references, regression tables and appendices left out. Figures are the paper's own, used under CC BY 4.0.