Do YouTube titles tell the truth? Honesty measured across twenty subjects

Across 3,243 measured videos, mean title honesty is 81 out of 100 — and it is not the same 81 everywhere. By subject it runs from 87.2 in science down to 72.9 in travel, fourteen points apart.

We score how accurately a video’s title describes what is actually inside it, on every video we measure. This report is that score on its own — where it is highest, where it is lowest, and how tightly it tracks the other things we measure about the same videos.

A note on the numbers. Unlike the rest of this site, the figures on this page are frozen at the date above rather than recalculated on each visit. A report people cite should not move under them. When we run this census again it will be published as a new edition with a new date.

Fourteen points separate the most and least honest subjects

Twenty subjects, ranked by mean title honesty, highest first.

Science, engineering and AI and machine learning lead, all above 85. Travel trails the whole list at 72.9, a full three points behind the next-lowest subject. The middle of the ranking is tight — fourteen subjects sit within six points of each other, between 78.0 and 84.2 — so read the two ends as the real separation and the middle as an ordering.

The same subjects that oversell also pad

Line the padding chart up against the honesty chart, subject for subject, in the same order.

Science sits at the top of the honesty chart and the bottom of the padding chart. Travel sits at the bottom of one and the top of the other. The two charts are not a perfect mirror image — engineering and gaming both carry more padding than their honesty rank alone would predict — but the overall drift is unmistakable: as you move down the honesty ranking, the bars in the padding chart get taller more often than not.

This is not us measuring the same thing twice under two names. Title honesty is scored against the video’s own description of itself; padding is scored against how much of the runtime restates or sets up rather than adds. They are independent measurements that happen to move together, which is the more interesting fact than either one alone.

The relationship holds video by video, not just subject by subject

Averaging by subject could in principle be hiding a weak relationship inside a strong-looking chart. It is not. Across all 3,002 individual videos carrying both a title honesty score and a padding measurement, the two move in opposite directions together, consistently enough to show up clearly at every level we cut the data.

Grouped into four honesty bands, the pattern is stark: videos scored 90 and above carry 34.9% padding on average; videos scored under 60 carry 56.7% — nearly twenty-two points apart, on a signal a viewer can read from the results page before clicking anything. We published that breakdown in full in the padding census, and again from the verdict side in the one-in-five report, where titles scored under 60 predicted a skip 72.1% of the time against 5.6% for titles scored 90 and above.

We are not claiming an inaccurate title causes padding, or the other way round. The more likely reading is that both come from the same decision: a video built to earn a click on a promise it does not quite keep is usually also built to stretch past the point where the promise runs out.

What a subject’s honesty score tells you

Treat a low subject score as a reason to read titles more skeptically in that subject, not as a reason to avoid it. Seventy-three is still an average, not a ceiling — plenty of travel videos have accurate titles, and the score tells you the odds you are working with before you know anything else.

Treat the title itself as the fastest signal you have, before playtime is spent. Across all twenty subjects, the video-level relationship between title honesty and padding is stronger than the subject a video sits in. A well-titled travel video is a better bet than a poorly-titled science one.

Expect the subjects with the least honest titles to also be the ones with the least accurate runtimes. If a subject is a poor fit for the explanatory format — its core idea compresses to a sentence — that shows up as both padding and overselling, together, because the video is compensating for the same shortfall in two ways at once.

How this was measured, and what it does not say

Every measured video carries a title honesty score, comparing what the title promises against what the video actually delivers. The reading is ours. Method and its limits are in what value density means.

Four subjects rest on fewer than eighty measured videos — travel, gaming, comedy and DIY & crafts. Their position at either end of a ranking is consistent with the padding and verdict data we have published separately, but read their exact figures as an illustration rather than a tight estimate.

Padding, used for the second chart, is measured on explanatory material only — 3,002 of the 3,243 scored videos. Honesty is scored on all 3,243; the chart’s own n notes the difference per subject.

Correlation, not causation, both across subjects and across individual videos. The relationship is consistent and it is not small, but nothing here establishes that fixing one measurement would move the other — both are more likely downstream of the same editorial decision.

The sample is not YouTube. It is 3,243 English-language explanatory videos across twenty subjects, assembled to build a recommendation library rather than to be representative of the platform.

The figures are frozen at 26 August 2026. Scores are recalculated as the underlying measurement improves, so the live numbers on the shelves will already have moved slightly from these. This edition will not be edited; the next census will be published separately.