Why Hiding Like Counts Might Actually Help Your Mental Health

In 2021, the Wall Street Journal published what became known as the Facebook Files — a series of internal documents leaked by whistleblower Frances Haugen. Among them was a 2019 internal study by Instagram's own research team, which found that the platform "is harmful for a sizable percentage" of its users, particularly teenage girls, and that a significant driver was social comparison enabled by visible metrics.

Source: Wall Street Journal, "The Facebook Files," September 2021; internal Facebook/Instagram research documents published as part of the Congressional record

The company had run the study internally. It understood what was happening. And it continued making like counts visible by default.

This is not a conspiracy. It's an incentive structure. Visible metrics drive engagement. Engagement drives ad revenue. The harm to users is real and documented — and it's outweighed, from the platform's perspective, by the commercial benefit.

What social comparison theory actually says

Leon Festinger proposed social comparison theory in 1954: humans evaluate their own opinions and abilities by comparing themselves to others, particularly when objective information isn't available. We need a reference point. Other people serve as that reference point.

Source: Festinger, "A Theory of Social Comparison Processes," Human Relations, 1954

Social media didn't create social comparison. It made it quantitative. In physical life, social comparisons are approximate and context-dependent. Online, they're numerical and stripped of context. Your post got 12 likes. Someone else's got 3,400. That number is immediate, visible, and easy to interpret as a verdict.

What makes this particularly corrosive is what's called upward social comparison — comparing yourself to people who appear to be doing better. Research consistently finds that upward social comparison reduces wellbeing. And social media, by surfacing the most popular content by default, systematically presents you with the top of the distribution. You're not seeing an average — you're seeing the outliers, and your brain doesn't naturally correct for this.

The hidden-likes experiment

Instagram ran a test in 2019 — initially in Canada, then expanded to seven countries — that hid public like counts from posts. The liker could still see how many likes their own post received, but other users couldn't.

The outcomes were mixed in ways that are instructive. Younger and more casual users reported reduced pressure and anxiety. Some influencers and professional users reported disorientation — their entire metric of success had become invisible. Instagram eventually made it an opt-in setting rather than a default, which effectively reversed the test for most users.

The interesting data point from the test: removing visible metrics reduced the compulsive checking behaviour that tends to follow posting. People who couldn't see their like count checked back less obsessively. The metric had been driving a behaviour loop as much as the content had.

Vote scores and the problem with Reddit

Reddit's upvote/downvote system is particularly interesting because it operates as explicit crowd endorsement, presented as a quality signal. A post at the top of a subreddit with tens of thousands of upvotes carries a social proof that makes it feel authoritative.

The problem is that the score reflects what people engaged with emotionally in the first few hours — which is not the same as what is accurate, balanced, or worth your attention. Removing vote scores from the display removes a cue that influences your reading before you've read anything.

How to hide metrics

Instagram (native): Settings → Posts → Hide like and view counts. This hides counts on other people's posts. You can still see your own.

Twitter/X (native): No native option. Requires a browser extension.

Reddit (native): No native option for hiding vote scores.

SocialFocus: Chrome extension that hides metrics across multiple platforms including Twitter/X and Instagram. Free tier covers most platforms.

Devoke: Hides like counts, repost tallies, vote scores, and view counts across Twitter/X, Reddit, and news sites by default — alongside rewriting sensationalist content. The two problems tend to travel together.

The argument against

Some people find metrics useful as a genuine signal. A post with 40,000 upvotes on a technical subreddit has probably been vetted by a large community. A news story with massive shares might indicate genuine public importance.

This is fair. The counter-argument is that the same signals also surface outrage and misinformation at scale — and that the compulsive checking behaviour metrics drive is probably a greater cost than the signal loss.

If you're not sure, hide the metrics for a week and see whether anything you actually needed disappears. Usually, it doesn't. The useful information is in the content. The number was a proxy for it that turned out to be a less reliable one than it felt.