Why purity scores fall with age

A Rice Purity score can only go down over a lifetime. That is not a claim about morals or about generations — it is a property of the instrument. The test is a checklist of things that have happened to you, and nothing that has happened can stop having happened. Once an item is true, it stays true at every future retake, so your own score is a number that never rises.

This is why "the average score drops with age" is close to a tautology, and why treating it as a finding about people is a mistake worth avoiding.

The ratchet

Start at 100. Subtract one point per statement that applies. Because the statements are past-tense — held hands, been on a date, been arrested — each one is a latch: it flips once and never flips back.

Say someone takes the test three times across twelve years and answers honestly every time:

Nothing about this person's character changed between rows. What changed is how many years they had been alive and in situations the list asks about. The same three rows would appear for almost anyone who keeps taking the test, with different numbers.

Some items are unreachable at 16 for reasons that have nothing to do with virtue

The canonical list contains eight statements about trouble with the law and seven about alcohol and drugs. A number of them are effectively locked until adulthood by law, money or logistics rather than by choice. A sixteen-year-old has not been in a bar because bars will not admit them.

So a high teenage score partly measures a shorter list of available items. That is worth stating plainly, because the teenage version of this comparison is where the moralising usually starts — "scores are lower than they used to be" gets said about a number that is mechanically lower for anyone who has lived longer.

Our teen test exists partly for this reason: it asks a different 50 statements aimed at a life that is still in progress, so the score is not a countdown of things a teenager has not had the chance to do yet.

Cross-sectional is not longitudinal

There are two completely different comparisons hiding behind "scores fall with age", and only one of them is about ageing.

Following one person over time (longitudinal). This is the ratchet above. It is real, it is arithmetic, and it tells you nothing except that time passed.

Comparing different people at one moment (cross-sectional). This is what any site's by-age table actually shows, including ours. It puts today's 18-year-olds beside today's 30-year-olds — two different groups of people, born twelve years apart, who grew up with different norms and who chose to visit a quiz site on the same day.

A gap between those two rows can come from ageing, from generational change, or from which group is more likely to submit a score. The table cannot separate them, and neither can we. What it can honestly do is describe the people who submitted.

Selection makes the age gradient steeper than it should be

Add the effect covered in the percentile explainer: who submits is not random.

A nineteen-year-old comparing numbers with a dorm floor is a common reason to take this test. A thirty-five-year-old typing "rice purity test" into a search bar usually has a different reason — curiosity about the thing itself, or a partner who mentioned it. Those two motivations select differently. The younger group includes more people taking it socially, which includes more deliberate zeroes for a screenshot, and also more people who genuinely have few items ticked.

Neither pull is measurable from the score alone, so a by-age average is a description of volunteers per band and never a population figure. That is why our by-age page publishes nothing for a band until it holds 30 results, and why it says so on the page rather than filling the cell with a number three people produced.

What a lower score at 30 actually tells you

That thirty happened.

If you want the number to say something more specific, the useful move is to stop reading the score and start reading the items. The per-question pages show what each statement asks and which block of the list it belongs to; the score pages show how many statements a given total implies. A drop from 69 to 56 over six years is thirteen more statements — and which thirteen is the only part of that sentence carrying information about a life.

If you are comparing with someone older or younger

Three things worth saying out loud before the comparison turns into a scoreboard:

1. Adjust for time, not for virtue. A twelve-year age gap is a twelve-year gap in opportunity, and the arithmetic does the rest.
2. Compare items, not totals. Two people twenty points apart may differ in one block entirely and match everywhere else.
3. The same instrument, or none. A classic-list score is not comparable with a couples, college or innocence score — different question sets, different scales, as the comparison page sets out.