Appeal to popularity

“Every cook I know does it”

Jordan: Why are you searing the roast first?

Casey: It seals in the juices. Every cook I know does it, and every recipe says so.

Casey may or may not be right about the juices, but the headcount doesn’t settle it. Most of those cooks learned the idea from other cooks and recipes, not by weighing roasts cooked both ways, so a hundred of them repeating it carries little more evidence than one. The claim is about what happens to meat, and it’s answered by testing that, not by counting who believes it.

The best-selling blender

A shopper compares blenders online. “This one is the best seller on the site by a mile. It’s obviously going to last longer than the others.”

Here popularity is evidence, but of something else. Sales show that many people chose this blender, usually on price, looks, features and reviews written in the first few weeks. Almost none of those buyers knew how long it would last when they bought it, so their choice can’t tell you much about durability. The argument takes a fact about the product’s appeal and treats it as a fact about its quality.

“Everyone on the team does it”

A new hire notices that colleagues paste account passwords into the team’s group chat. When they ask about it, a coworker says: “Everyone here does it and nothing bad has happened, so it’s fine.”

This is the appeal to common practice: a practice is defended as acceptable because it’s widespread. Whether the practice is safe depends on who can read the chat and what happens if an account in it is compromised. The number of people doing it doesn’t change either. (The “nothing bad has happened” part is a different weak argument: a risk can be real and still not have caused a problem yet.)

• Bandwagon: everyone believes or does it, so you should too. “Everyone’s switching to this app. Don’t get left behind.”
• Appeal to common practice: it’s commonly done, so it’s acceptable. “Everybody rounds their hours up a little.”
• Snob appeal (appeal to an elite group): the right people believe or do it, so you should. “Serious runners don’t use that brand.” The crowd is smaller and more exclusive, but it’s still standing in for reasons.
• Appeal to vanity: linking a product or idea with an admired figure, so that you’ll want to be associated with it.
• Mob appeal: rousing a crowd’s feelings so that its shared enthusiasm does the work of evidence.

Patrick Hurley’s textbook sorts these into a direct approach (working up a crowd) and an indirect one (appealing to an individual’s relationship with the crowd), and counts the appeal to tradition (“we’ve always done it this way”) among the indirect forms. Douglas Walton’s book-length treatment distinguishes eleven subtypes and argues that many appeals to popular opinion are reasonable in context, especially when a group is deciding what to do.

Chronological snobbery makes the same kind of substitution along a different axis. An appeal to popularity asks how many people hold an idea; chronological snobbery asks when it was held, and rejects it for being old. “Nobody believes that anymore” commits both at once.

The bandwagon effect is a different thing: a term from the social sciences for the tendency of people to adopt a belief or behavior as more others adopt it. That describes what people do. The fallacy is an argument: a claim that popularity justifies the belief.

How many people believe or do something is a good reason when popularity is what makes the claim true, or when it reflects evidence the people had.

• The claim is about popularity or convention. Which side of the road people drive on, what a word means, what counts as polite at a dinner table and which file format everyone can open are all settled by what most people do. Here the headcount is the evidence.
• The question is what a group should do. When a club picks a meeting night or friends choose a restaurant, what most members prefer is a legitimate reason, not a distraction from one. Walton’s book argues that appeals to popular opinion are often reasonable when a group is deciding what to do.
• Many people judged independently, each with some access to the evidence. In a 1907 paper, Francis Galton analyzed 787 entries in a contest at a livestock show in Plymouth, England, to guess the weight of an ox after slaughter; the entrants included butchers and farmers. The middle estimate was within 1 percent of the true weight. (A 2014 re-examination of Galton’s worksheets corrected small errors in his figures: the median was 11 pounds over a true weight of 1,197, and the mean was exact.) Pooled estimates like this work only under conditions: people form their own judgments, and those judgments vary. In one experiment, letting people see others’ estimates narrowed the spread of answers without making the group more accurate, and made individuals more confident anyway.
• The agreement is among experts. A consensus of people who have studied a question is a different argument from popularity. It’s closer to an appeal to authority, and it’s judged by whether they really are expert on this question and reached their views separately.
• As a reason to look, not a verdict. A crowded restaurant or a widely used tool is a fair reason to give it a try, as long as you treat that as a weak starting point that your own experience can overturn.

The test: is the claim about what people do, or does their agreement reflect evidence they each had? Or is the count itself the only reason on offer?

Standing on the right

A visitor to London stands on the left side of an escalator and gets tapped on the shoulder. Their friend says: “Stand on the right. Look, everyone else is.”

This appeals to what everyone does, and it’s correct, because the question is itself about convention. There’s no fact about escalators that makes the right side better; standing there is right because it’s what people expect. This is the convention condition above.

The jelly bean jar

A teacher has thirty students each write down, without talking, a guess at how many jelly beans are in a jar. She tells the class that the middle guess is likely to be closer than a typical individual guess.

She’s leaning on what the group thinks, but each guess came from a student looking at the jar themselves, and nobody could copy anyone else. To the extent that their errors point in different directions, they partly cancel out. If the students had called out their guesses one after another, later guesses would tend to drift toward earlier ones, and the argument would weaken. This is the independent judgments condition above.

Relying on other people is usually sensible. Nobody can check most of what they believe, and a great deal of what anyone knows comes from others. The habit misfires when agreement isn’t independent: a belief passed from person to person gains supporters without gaining evidence, yet each new supporter makes it sound better established.

Agreement also feels like confirmation. In the experiment on pooled estimates described above, people who saw others’ answers grew more confident after their answers converged, although the group was no more accurate. The Stanford Encyclopedia of Philosophy’s entry on fallacies describes ad populum as a relative of the appeal to authority, where the source appealed to is popular opinion or common knowledge rather than a named authority. That kinship helps explain why it persuades: it borrows the credibility we reasonably give to sources.

• Ask how the people came to believe it. Did they observe it, test it, or hear it from each other?
• Ask what the popularity is evidence of. Sales measure appeal at the moment of purchase; repetition measures how easily an idea spreads. Neither is automatically evidence about the claim.
• Separate the question from the count. Imagine the same claim held by one person. What would you need to know to evaluate it? That’s still what you need to know.