DozeClock › Corrections policy and log

Corrections policy and log

Updated October 3, 2026

If a number on DozeClock is wrong, we want to know, we fix it in the open, and the fix stays on record.

How to report an error

Use the contact form and tell us the page and the figure. A link to the source that disagrees helps, but is not required. You do not need to give a name or an email.

What we do with a report

We check the figure against the original publication listed on the sources page and re-run the formula from the methodology page. If the page is wrong we correct the data file, which corrects every page built from it, and add a dated entry to the log below. We aim to do this within seven days. If we think the page is right we keep the figure and, where the confusion is reasonable, add a sentence that explains it.

How changes are marked

Each page shows two dates. "Updated" is the date of the data snapshot the page was built from. "Last reviewed" is the last time the formulas and sources were checked, even if nothing changed. A correction that changes a number is always logged here with the old and the new value; we do not edit silently. Wording changes that do not alter a figure or a conclusion are not logged.

What we will not do

We do not remove a correction from the log, and we do not attribute pages to people who did not write them. DozeClock is made by a small team, and we are not clinicians, as the about page says.

Log

DateWhat changedFigures changed
2026-11-02Brought the Underwater page back with its own measurements, and corrected the White Noise Lite row on the app comparison: its listing now needs iOS 15 and has no in-app purchases.None
2026-10-19Folded the eleven nap-start-time pages (12 pm to 5 pm) into the naps page as rows of one table, and rewrote the seven nap-length pages around the study that measured each length. Old addresses redirect to the matching row; no figures changed.None
2026-10-18Folded the Underwater scene page into the sounds page as a section, added scene-specific research to the Fire, Beach and Forest pages, and rewrote the nine sleep-needs-by-age pages with survey data for each age. The old address redirects to the section; no figures changed.None
2026-10-17Rewrote the seven "is N hours of sleep enough" pages with their own study evidence for each number of hours. No figures changed.None
2026-10-16Folded 28 quarter-hour wake-up pages into their hour (6 pm to 3 am bedtimes); each is now a row of the hour page, and all eleven hour pages were rewritten with their own text. Old addresses redirect to the matching row; no figures changed.None
2026-10-15Folded the nine caffeine-cutoff pages (one per bedtime, 9 pm to 1 am) into the cutoff page as one grid with a row for each bedtime, and folded 48 bedtime pages into their hour: quarter-hour alarms are now rows of the hour page, and alarms from 12 to 2:45 am and 11 am to 1:45 pm are rows of the full table on /bedtime/. The eight morning hour pages (3 am to 10 am) were rewritten with their own text. Old addresses redirect to the matching row; no figures changed.None
2026-10-14Merged 22 short hours-by-age pages into the seven "Is N hours of sleep enough?" pages, where each age group is now its own section. Moved the power-nap guide into the nap page and the teenager guide into the teenager sleep-needs page, so each question has one page. Every old address redirects to the matching section. No figures changed.None
2026-10-05Dropped 73 draft pages that were never published (quarter-hour bedtime and wake-up variants, and hours-by-age variants); their addresses redirect to the hour or hours page that already covers them. No published page or figure changed.None
2026-10-05Rebuilt the 12 caffeine-by-drink pages: each now shows the measured spread for that drink by size, brand or brew from USDA FoodData Central, FDA and maker pages, with the source of every row. Headline mg values are unchanged.None
2026-11-03The clock-change page still said "Fall back Sunday, Nov 1, 2026" in its title and answer two days after the clocks went back, and its planner was still set to that change. It now says the clocks went back on November 1, 2026 (October 25 in the EU and UK), gives the next change, March 14, 2027 in the US and March 28, 2027 in Europe, adds a section for the week after the change, and the planner opens on March 2027. The state section said 49 of the 50 states change their clocks; it is 48 plus Washington, DC (the count had included DC). Every date on that page, and the clock-change lines on the home page and the bedtime calculator, is now worked out from the day the site is built, and the site rebuilds itself the day after each change.States that change: 49 → 48 (plus DC)
2026-10-27Measured the 30 recordings of the Fall Asleep app and 6 reference noise files (spectral slope, energy centre, loudness swing; method and table on the open data page). The Beach scene had been described as closest to brown noise; measured at the app's starting levels it falls 3.4 dB per octave, closest to pink, and the page now says so. "Steady rain is close to pink noise" on the rain and thunderstorm pages became the measured range: the rain loops fall 0.6 to 1.9 dB per octave, between white and pink, and the whole Rainy Day scene measures pink.Beach: brown → pink
2026-10-07Removed the "34% better performance, 54% better alertness" figures attributed to the NASA cockpit-nap study (Rosekind et al. 1995) from the nap pages, the nap calculator, the home page and the methodology page. We could not find them in the report text we opened. They are replaced by counts the report does give: 25.8 minutes of sleep on average in a 40-minute rest window, 81 against 124 vigilance lapses and 34 against 120 microevents of reduced alertness (rest against no-rest group).34% / 54% → removed
2026-10-04Added the clock-change planner for the November 1, 2026 change, with its sources. Added an hour-specific section and an hour table to twelve pages that had not yet been published (wake-up alarms from 2 pm to 11 pm, and the 6:00 pm and 4:00 am bedtimes).None
2026-10-04Review. Every bedtime, wake-up, caffeine and sleep-cycle page re-computed from the data files by an independent script; the three calculators tested in a browser. Source links re-opened: 18 of 19 loaded. The remaining one (the AASM/SRS 2015 adult consensus) resolves by DOI to the publisher, Oxford Academic, whose bot check our tools could not pass; the DOI itself is valid. Added: source links on the methodology page, this policy, the open data tables.None
2026-10-03First publication; snapshot of all data files.—

Photo by Jeremy Bishop on Unsplash

Last reviewed by the DozeClock team: formulas re-run against the data files, sources re-checked. How we calculate · Corrections policy and log