Organizing popularity games questions for a quiz night

Popularity games questions are opinion prompts: unlike pub-quiz knowledge questions, they have no fact-based answer key. Their answer is the distribution produced by the room. Building a bank for an opinion round looks like an easy writing job because there is nothing to verify, but each prompt still needs an answer sheet in the form of the room’s count.

That changes what a source note is. For a knowledge question the note records where the fact came from. Here it records a measurement: who was asked, in what words, on what day, and what they said. Without that, a prompt you reuse is a prompt you cannot compare with anything.

What to take away

What a source note looks like here

Write four things beside every prompt, and write them at the time.

The exact wording used. Prompts drift when they are read aloud, and a prompt read slightly differently produced a slightly different question. If you plan to compare across rooms, the wording has to be fixed.

The option set as offered, in the order it was offered. Order affects answers, and adding or removing an option changes the whole distribution.

The count, in full. Not the winner, the numbers. A prompt that split evenly and a prompt that went nine to one are useful for completely different jobs, and only the full count tells you which you have.

The room. Roughly how many people, and anything about them that plainly mattered. A prompt that produced a good split in one room can be unanimous in another, and knowing which rooms produced which result is what makes the bank reusable.

Kept that way, the bank is a record of measurements rather than a list of ideas, and the scoring machinery it feeds is in popularity games.

Writing a prompt worth testing

The failure mode is unanimity, but no wording can guarantee a split until a prompt is tested in the room where it will be used.

Pit like against like. Two options that are strong in different ways divide a room. One strong option against a weak one does not, and the second is what a hurried writer produces, because the strong option comes to mind first and the second is chosen to fill the sentence.

For a celebrity round, test this straight choice: “Which artist’s live show would you rather see: Beyoncé or Taylor Swift?” A bank entry can preserve that exact wording and option order, then mark the count and room “not tested” until the prompt is used. That keeps a celebrity prompt from being presented as a proven split before you have results.

Remove the respectable answer. Any option that reads as the correct thing to say will attract people who are answering in public, and the round measures politeness instead of preference.

Ask about one thing. A prompt that asks which option is better and more enjoyable is a double-barreled question, and it produces answers that cannot be interpreted, because a table that split the two criteria had to pick which one to answer.

Keep it short enough to hold, since a prompt is heard once and answered in seconds.

The craft of questionnaire construction says short, single-barreled, neutrally worded items apply here, with one added constraint: read aloud in a noisy room.

The same discipline shows in easy popularity games, where each round is built so no answer can be wrong.

Avoid a scale where a choice will do. Asking people to rate something out of ten produces a pile of sevens. Asking them to choose produces a distribution.

Fifty is a bank, and banks need structure

Fifty prompts you cannot navigate is worse than twenty you can.

Tag each prompt with its shape: a straight choice, an ordering, an estimate of what fraction of the room will say something. For example, tag the celebrity prompt above as a straight choice and mark its split untested until a room supplies a count. A round wants a mix, and the mix is what stops the format from feeling repetitive.

An ordering example is: “Rank comedy, drama, horror, and science fiction by which kind of movie you would most want to watch tonight.” An estimate example is: “What fraction of this room would rather see a live concert than watch a movie at home?” Both can be tested and tagged with the result they produce, rather than treated as proven splits in advance.

A sample entry before its first use could read: “Which artist’s live show would you rather see: Beyoncé or Taylor Swift?” Options, in order: Beyoncé; Taylor Swift. Full count: not tested. Room: not tested. After the prompt is used, replace those placeholders with the full count and a description of the room.

Tag each with the split it produced last time. When you build a round you want a few prompts that reliably divide a room and a couple that reliably do not, because a near-unanimous result is a good reveal once per round.

Tag each with anything that dates it. Most opinion prompts do not expire, which is one of the format's advantages, but any prompt whose options refer to something current will need the same time check that daily and seasonal quizzes items need.

Retire prompts that produced discomfort rather than disagreement. A prompt that made a table go quiet is not a hard prompt, it is the wrong prompt, and the note against it should say so.

Keep the subject out of the room

The strongest rule in this format is about who the prompts are about.

Prompts are about things. They can be about anything at all outside the room: works, formats, categories, choices, hypothetical situations. That is a very large space and it never runs out.

Prompts are not about the people present. Anything that asks the room to rank, compare or characterize the guests is a different kind of game, it needs everybody's agreement in advance, and it does not belong in a scored round where declining to answer costs points.

The line is easy to check. If the answer to a prompt is a person in the room, cut it. If the answer identifies a person in the room by implication, cut it too.

Reading them out

Delivery matters more here than in a knowledge round, because the prompt is the entire stimulus.

Read the prompt and then the options, in a fixed order, and then read the options again. Nobody remembers four options heard once.

Do not editorialize between options. A tiny inflection is a recommendation, and the round is then measuring your delivery.

Collect everything in writing at the same moment. This is the single most important mechanic in the format, and it is the one most often skipped for speed.

Reveal the count before scoring. The distribution is the interesting part, and reading it out first makes the scoring feel like a consequence rather than a verdict. For a prediction round, say in advance whether teams are predicting the most popular option or the exact count; the stored full count is the answer key. For a preference round, there is no correct answer, so do not score it as a knowledge question. Collect each team’s choice in writing before the reveal, then apply the rule announced before play. The general principle that a question should have one defensible reading, argued in guess the celebrity questions, applies to the prompt even though it does not apply to the answer.

Common questions

How is this different from writing knowledge questions?

The verification work disappears and the wording work multiplies. A knowledge question can be clumsy and still have one answer. A clumsy prompt has no answer at all, which is the same standard applied in career quizzes questions from the opposite direction.

Can a prompt have a right answer?

Only in the prediction version, where players are guessing what the room will say. Then the right answer is the count, and it is exact.

How many options should a prompt offer?

Two for a clean split, four for an ordering. More than four and the tail of the list is answered by whoever was still listening.

What if the same prompt gives opposite results in two rooms?

That is the format telling you something true about the two rooms. Record both and use the prompt where the split is useful, which is usually the room that divided.

Do I need fifty?

You need enough to build several rounds without repeating, and to retire the ones that do not work. Fifty is about right for a series; twenty is enough for an evening.