Bandwagon effect

How it works

In the 1950s, Solomon Asch sat people in a room with a group of actors and asked a question a child could answer: which of these three lines is the same length as that one? When the actors confidently gave the wrong answer, about a third of the real participants’ answers went along with them, and roughly three in four participants went along at least once. Some said afterwards that they knew the group was wrong and did not want to stand out. Others had started to doubt their own eyes. (Later reviews found that conformity rates vary a lot by culture and era, and that a single dissenting voice in the room cuts the effect dramatically. Remember that detail.)

Online, the room has been replaced by numbers. Likes, shares, views, “12,000 people have signed”, a comment section that all agrees. Your brain reads those numbers as a crowd and does what it did in Asch’s lab. The difference is that this crowd can be bought by the thousand. Bots, sockpuppets and paid engagement farms exist precisely to manufacture the look of consensus, a practice called astroturfing when it pretends to be a grassroots movement.

The effect snowballs. In a 2013 experiment, researchers randomly gave some new comments on a news site a single extra upvote. That one vote made other readers more likely to upvote too, and lifted the comment’s final score by about a quarter on average. An earlier study using an artificial music market found that when people could see what others had downloaded, hits became bigger, flops became smaller, and which songs became hits turned largely unpredictable. Early numbers, real or fake, decide what you later think is popular.

Example

@future_now_official, 6h 48,000 signatures in one day for the party’s new initiative. The people have spoken. Media silence as usual.

@ben_f, 5h Signed. If this many people back it, it can’t be wrong.

@comment_archaeologist, 2h Scrolled the 300 supportive comments: 210 accounts created this month, most with one post. The counter on the petition page also doesn’t change when you reload. Might be real support, but these numbers aren’t proof of it.

What to notice: the argument in the second post is purely the number. Nobody has discussed what the initiative actually says. The crowd is the claim.

Why it works on you

• Other people are usually a good shortcut. If everyone runs out of a building, following them is smart. The shortcut only breaks when the crowd is fake, and online you cannot see the crowd. You can only see a counter.
• Disagreeing with a group is physically uncomfortable. Asch’s participants sweated and fidgeted. A thread of fifty people agreeing produces a milder version of the same pressure, and silence feels safer than a lone comment. This is how closed groups keep their members quiet.
• Popularity gets mistaken for correctness. “Everyone knows” and “the silent majority” are claims about a crowd that nobody can see. See ‘common sense’ & the silent majority for the political version.

What to do

When the numbers impress you: ignore them for a minute and read the thing itself. Then look at the crowd: account ages, profile photos, whether the comments say anything specific. A thousand “so true” comments from accounts born last week are not a thousand people.

When you catch yourself staying quiet: remember that in Asch’s lab, one person giving a different answer was enough to free most of the others. You do not need to win. You need to be the one visible dissent.

In a comment section: do not say “these are all bots”. You will usually be wrong, and it is a lazy dismissal. Say what you can see: “many of these accounts are new and identical” is checkable; “bots!” is just a feeling. For shopping and ratings, the same checks apply (see fake reviews).

Don’t confuse it with

• Expert consensus. “Most climate scientists agree” is a statement about people who have checked the evidence, not about a crowd that has not. The bandwagon is about numbers replacing reasons.
• Genuinely popular things. Some things are popular because they are good. The question is never “is it popular?” but “would I think this without the counter?”