The Regulation Index (RI-1) and the Regulation Profile (RP-1)
Two self-report instruments for nervous system regulation in adults — published in full, with their scoring, their method, their limits, and the live distribution of how everyone who has taken them actually scores. Free to use in your own practice or research, with attribution.
What this is
Almost everything written about nervous system regulation is qualitative. You can read a great deal about what dysregulation feels like and very little about how to measure it, and nothing at all about what a given measurement means relative to anyone else. These two instruments exist to close that gap, and this page publishes them completely, so that a number produced by them can be checked, argued with, reused and cited.
The Regulation Index (RI-1) is the short one: five items, about thirty seconds, a seven-day recall window, scored 0–20. It is built to be answered weekly for months, which is a constraint on length before it is anything else — a tool nobody opens measures nothing.
The Regulation Profile (RP-1) is the long one: twenty-four items across six domains, about eight minutes, scored 0–96, intended roughly quarterly. Its purpose is not a more precise total. It is the per-domain read: RI-1 can say you are at 11 of 20, and only RP-1 can say recovery is what is carrying that, and it is not sleep.
Neither instrument reproduces a third-party scale. The PSS, WHO-5, MAIA and BPQ are variously copyrighted or licensed for non-commercial use only; what these borrow, and credit item by item below, are the published constructs, worded from scratch.
Population norms — live
This is the part that did not exist until now, and the part that makes a score interpretable. 11 of 20 means nothing on its own: a person reading it for the first time has no way to tell whether it is unremarkable or alarming. A distribution answers that.
RI-1: the five items, verbatim
Each item targets one construct that the autonomic and stress literature treats as a marker of nervous system state. Response anchors are scored 0–4 in the order shown, so 4 is the resourced end of every item — higher is better throughout, with no reverse-scored items to get wrong.
“How much of the last 7 days felt like the week was running you?”
- 0 Almost all of it
- 1 Most of it
- 2 About half
- 3 A bit of it
- 4 Hardly any of it
“When something wound you up this week, how long did your body take to come back down?”
- 0 Hours, or it never really did
- 1 About an hour
- 2 Twenty or thirty minutes
- 3 A few minutes
- 4 Almost straight away
“How many of the last 7 nights did you wake up actually restored?”
- 0 None, or one
- 1 Two
- 2 Three or four
- 3 Five or six
- 4 Every one
“Did you catch your body tightening up before it turned into a problem?”
- 0 Never — only afterwards
- 1 Once or twice
- 2 Sometimes
- 3 Most times
- 4 Nearly every time
“By the end of an ordinary day this week, what was left in the tank?”
- 0 Nothing at all
- 1 Almost nothing
- 2 Enough to get by
- 3 A decent amount
- 4 Plenty
RI-1: scoring and bands
Sum the five items for a total out of 20. No weighting, no subscale transformation. The bands below are interpretive language, not cutoffs — they exist so a total arrives with a sentence rather than alone.
| Total | Band | Read |
|---|---|---|
| 0–5 | Running on reserves | Very little is coming back in. At this level a one-minute reset is relief, not repair — the load itself is what needs to change. |
| 6–9 | Activated | The system is spending most of the week switched on. Recovery is happening, but it is losing to the demand. |
| 10–13 | Holding | Coping, on thin margins. The band where a regular practice makes the most visible difference, because there is something to build on. |
| 14–17 | Recovering well | The body is coming back down between demands. That is the thing that actually matters, and it is working. |
| 18–20 | Resourced | Good capacity and good recovery. Worth knowing what this feels like from the inside, so the slip is noticeable. |
Change threshold
A week-on-week difference of fewer than 3 points is reported as “about the same”. This is a stated noise floor, not a validated minimal important difference: perceived load and sleep swing hard, and a two-point move on a five-item scale is not distinguishable from a different Tuesday. Cadence is once every seven days, never more often — daily self-report on items like these produces straight-lining within weeks and corrupts the data it collects.
RP-1: the six domains
Four items per domain, each scored 0–4, giving a domain subtotal out of 16 and a total out of 96. The five RI-1 constructs, plus one that only a longer instrument has room for.
| Domain | Construct | What it reads |
|---|---|---|
| Load | Perceived load and controllability | How much of the week is set by things you did not choose, and how much you feel able to steer. |
| Recovery | Return-to-baseline speed | How fast the body comes back down once wound up. The single most useful thing to move. |
| Sleep | Sleep health and restoration | Not hours. Whether you get to sleep, stay asleep, and wake with something in the tank. |
| Noticing | Interoceptive awareness | Whether bodily change registers early enough to act on, rather than afterwards. |
| Capacity | Allostatic load and end-of-day capacity | What is left at the end of an ordinary day, and how much of the week runs on empty. |
| Bracing | Chronic defensive bracing, hypervigilance | Background holding — jaw, shoulders, breath, scanning — that stays on when nothing is wrong. |
A spread of 3 or more points between the highest and lowest domain is the finding worth acting on, and it is the reason the long instrument exists: two people with the same total and different spreads need different things. Change threshold on the total is 12 points, scaled from RI-1’s 3 for the wider range.
How the norms are computed
One observation per person
Every distribution on this page is computed over each participant’s most recent completion only — never over every row recorded. This matters more than it sounds. Somebody who has checked in weekly for a year would otherwise contribute fifty-two observations against a newcomer’s one, and since long-term users are precisely the people whose scores are improving, norming over raw rows would drag the reference upward and make every new arrival look worse than they are.
Mid-rank percentiles
A percentile is everyone below the score, plus half of everyone level with it, over the total. Half the tie is used because a tie carries no information about order within it; the simpler below-only calculation would tell the single most common score in the distribution that it sits at the bottom of its own tie.
Two floors
An individual is not placed against fewer than 30 people, and a metric is not published here at all below 50. The first floor is about harm rather than statistics: a person in a bad week reading “bottom 10%” off a sample of nineteen has been told something false about themselves that they will believe. Percentiles are also reported to the nearest five rather than the nearest one below 200 people, because “the 37th percentile” off ninety observations claims a resolution the data does not have.
Refresh and privacy
Recomputed once a day, so a published figure holds still between page loads and can be cited with a date. The stored aggregate contains counts of scores and nothing else — no user identifiers, no email addresses, nothing joinable back to a person.
What this cannot do
There is no clinical cutoff anywhere in either instrument, and no percentile in this dataset should be read as one. Neither instrument screens for, detects or diagnoses anything — not ADHD, not anxiety, not depression, not burnout, not any autonomic condition. They are self-tracking tools for noticing your own direction of travel over weeks.
The norming population is self-selected, as described above. It is also unstratified: no age, sex, diagnosis or medication status is recorded against a score, which means these norms cannot be broken down by any of those and cannot be compared to a matched sample. That is a deliberate consequence of not collecting the data, and it is a real limitation rather than a modesty note.
If you need a validated measure — for research, for clinical decisions, for anything that goes in a chart — use a validated measure. These are honest instruments for noticing change, published openly so that what they are and are not is checkable.
Licence and citation
Both instruments, the item wording, the scoring, the bands and the published norms are released under Creative Commons Attribution 4.0. Use them, reproduce them, build on them, including commercially. Clinicians, coaches, app developers and researchers are explicitly welcome. The one condition is attribution: credit PKJ Coaching and link to this page.
If you adapt the items, please version your adaptation under a different name. Two different sets of wording sharing one identifier is how a measurement becomes uninterpretable, which is the same reason pkj-ri-1 is stored alongside every score here.
Machine-readable norms: /api/norms — JSON, CORS-open, cached six hours. Take the measurement yourself at /measure; it is free, needs no account and takes about a minute.