Counting plan for the first cohort

How we will count the results of the first cohort — announced in advance, before any data. It is a test of the method, not of people: those who leave stay in the count, and poor results are published on a par with good ones.

Title. “Fire of Ethics”: feasibility of a 30-day online practice of everyday conduct — first cohort (pilot).

1. Have any data been collected for this study already? No. The plan was published on 7 Nov 2026, the day the cohort was created: nobody has joined it yet and no baseline questionnaire has been filled in. Baseline questionnaires open to those who join, before the start; course entries and all other questionnaires start on 23 Nov 2026.

2. Main question. Are the course and an honest measurement feasible: what share of those who start records at least 21 of the 30 days and answers the questionnaires within their windows? No causal hypothesis is tested: there is no comparison group.

3. Key dependent variable. The share of participants who recorded at least 21 of the 30 course days, among everyone who opened Day 1 (logged in after the start or recorded at least one day), including those who left. A recorded day means the day's minimum was ticked or at least one of the four evening lines was written; Days 1–30 only. Cut-off: entries made before 22 Dec 2026, 09:00 (Moscow time); recount: everything entered by the end of Day 44 (5 Jan 2027). Benchmarks: 60% or more — feasible; 40–59% — acceptable; below 40% — not feasible in its current form; for each outcome, the change for the second cohort is written down in advance.

4. Conditions. One group, no assignment. One cohort of up to 12 people, led by the programme's founder.

5. Analyses. The key variable is reported as “k of n” with a percentage. Secondary, descriptive only: change in the total score of the Satisfaction With Life Scale (SWLS, Russian version, 5 items) from D0 to D30 and to D90, within-person pairs only, at least five pairs; and change in three self-made self-rating scales (1–5) from D0 to D30 and to D90, within-person pairs only — mean change, 95% interval, number of pairs, at least five pairs, no p-values; an interval crossing zero is reported as “undetermined”. The share with at least one course practice still alive at Day 90 — as three numbers (alive / not alive / no answer) out of everyone who started, with the two bounds. Weekly pulse — as a course total. Everything else is exploratory.

6. Outliers and exclusions. Nobody is excluded from the denominator: those who left stay in the count. Only administrator accounts are excluded. Questionnaire answers outside their windows (D0 after Day 1; D30 outside Days 28–44; D90 before Day 90) are not analysed and are reported separately. Days first recorded later than the next day are counted and reported separately.

7. Sample size. Everyone who opens Day 1 in the first cohort (start 23 Nov 2026; up to 12 places; joining until 29 Nov 2026). No other participants are added to this analysis.

8. Other. All numbers are computed by the platform's code, which outputs only cohort totals; nobody reads individual questionnaires or entries. Numbers for groups smaller than five are not published. Deviations from this plan are listed as dated “Transparent changes” lines together with the report. The report comes in three releases: 22 Dec 2026, January and March 2027. About the programme and its source: https://fireofethics.com/o-proekte/#istochnik.

The same in brief — on the About page, in the block “What and how we will count”.