Purity test vs innocence test

"Purity test" and "innocence test" are two names for the same mechanism: a checklist of experiences where every yes costs you a point and the total counts down from a perfect score. There is no methodological difference between the labels — the difference that matters is which questions are on the list, and that difference makes scores from different versions genuinely non-comparable.

If you and a friend took different versions, you are not comparing yourselves. You are comparing two questionnaires.

Why the two names exist

The Rice Purity Test carried the older name, from a campus tradition of "purity" questionnaires that predates the internet. "Innocence test" is the softer rewording that spread later, partly because "purity" reads as moralistic to a modern audience and partly because search demand split: people looking for a gentler, less explicit version tended to type the other word. The scoring is identical. Nobody rebuilt the mathematics; they renamed the file.

Where a real difference shows up is in question selection. Most lists marketed as innocence tests just trim the explicit sexual and hard-drug items out of the purity list, which changes what the score can reach and not only what it says. A few, including ours, go further and ask about a different subject entirely.

The four tests on this site

Rice Purity Test — the classic 100-item list, covering every cluster: romance, sexual experience, substances, parties, law and general mischief. Written for adults and older teens who want the canonical version. This is the reference scale: a score quoted online with no version named means this list.

Teen Purity Test — an age-appropriate subset for 13 to 17: dating, school rule-breaking, parties, first drinks. The explicit sexual and hard-drug items are removed rather than softened, so its score cannot be mapped back onto the classic scale.

Couples Purity Test — shared experience between two people instead of one person's history, for partners taking it together. A low number means the couple has done a lot together; it says nothing about either individual.

Innocence Test — not a censored copy of the classic list; a separate 100-item list that shares no questions with it. It asks about travel, driving, work, money, injury, performing in front of people and sitting through a funeral. Exactly one item mentions alcohol and none ask about sex, so a low score here means a full life rather than a wild one.

Each test page states its own question count and scale above the first question. Side-by-side, the same information is on the compare all tests page.

Why the scores are not interchangeable

The arithmetic makes this concrete. Suppose you have done twenty things that appear on both lists.

On a 100-item list, twenty checks gives you 80. On a 50-item list containing those same twenty items, twenty checks gives you 60. Identical behaviour, identical honesty, a twenty-point gap — created entirely by the denominator.

The second effect is subtler and cuts the other way. When a version removes its most extreme questions, it also removes the items that only a small minority ever check. What remains is the part of the list almost everyone eventually checks, so the achievable floor rises: on a trimmed list you may be unable to score below 30 no matter what you have done, because the questions that would take you lower are not there. A trimmed test compresses the bottom of the range and expands the middle.

So there are three separate reasons two scores may not be comparable:

• Different length. Each check is worth a different fraction of the total.
• Different question mix. A version without the legal or hard-drug clusters cannot reach the same low values.
• Different wording of the same question. "Have you ever been drunk?" and "have you ever had an alcoholic drink?" are answered yes by very different numbers of people, and clone sites rewrite items freely.

The only clean comparison is same test, same version, same wording. That is why the score history stored on your history page records which test produced each number, and why the distribution on our live statistics page is kept separate per test rather than pooled.

The clone problem

Search for "purity test" and you will find dozens of sites offering "the" test with a hundred questions. Many are not the same hundred questions. Items get reordered, reworded to be milder or more explicit, dropped when they read as dated, and replaced with new ones about phones and social media. The page still says "out of 100", so the number looks authoritative and comparable when it is neither.

Two quick checks tell you how much weight a given copy can carry. First, count the items and check the arithmetic: a list that says "out of 100" while offering 87 checkboxes, or that deducts more than one point for some answers, is not scoring the way the number implies. Second, read the first ten items and the last ten. The canonical shape runs from very mild opening items — hand-holding, a first kiss — to genuinely extreme closing ones; a list that stays mild throughout has been trimmed, whatever it is called, and its floor is much higher than 0.

What you cannot use as a check is the presence of phone questions. Older mirrors have none because their wording was frozen before smartphones; our own classic list has four, because we added them. Modern items mean the copy has been revised, which is a fact about that copy and not a verdict on it — see where the Rice Purity Test came from.

Can you convert a score between versions?

Not reliably, and the reason is worth understanding. A conversion would require knowing which specific items each person checked, not just their total — because the only way to restate a score on another scale is to re-count their checks against that other list. The total throws that information away. Two people at 70 on a trimmed test might sit at 70 and 45 on the classic list, depending entirely on whether the items the trimmed version omits happen to apply to them.

The closest honest approximation is to retake the version you want the number in. It costs a few minutes, it uses your real answers instead of a fudge factor, and it is the only method that produces a number you can defend when someone asks how you got it.

Which one to take

Take the classic 100-question test if you want the number people mean when they quote a score, and you are an adult or a teenager comfortable with explicit questions.

Take the teen version if you are 13 to 17. It is not a censored consolation prize; it is a list built around experiences that are actually available at that age, which makes the result more informative rather than less.

Take the couples version with a partner. It asks about the two of you, so it produces a conversation rather than a comparison, and the failure mode of the classic test — turning a relationship into a scoreboard — does not apply.

Take the innocence test if you want a life-experience score rather than a purity one — it asks about places you have been, jobs you have had and things you have survived, and it is the only one of the four you could run at a work event without anyone regretting it.

Whichever you choose, read the result as a description of which clusters you checked rather than a rank. What every score means breaks that down value by value, what counts as a good score explains why the ranking framing fails, and how to read your percentile covers what comparing yourself to other submissions can and cannot tell you.