What we think, and how sure we are.
Sleep research is genuinely uncertain in places. Effects that hold in a laboratory can shrink in a normal week, findings measured in one group do not always travel to another, and some questions are still being argued over. Rather than smooth that over, we rate it — and publish the definitions we rate against.
Six levels, defined in advance.
Every guide and every brief on Restadian carries one of these six ratings. The rating describes the state of the research, not how much we like the idea. Three of the six describe different kinds of not knowing, because a reader is owed the difference between a live argument, a question the research cannot settle, and a question nobody has studied.
Strong
A well-described mechanism supported by repeated findings across independent groups, measured in people, with results that point the same direction. New work is unlikely to reverse it.
Reasonable to act on. Individual response still varies, so the timing and the amount are yours to adjust.
Moderate
A real effect, consistent enough to be useful. How large it is, how long it lasts, or which people it was measured in still leaves real room to move.
Worth trying as an experiment you can reverse. Watch what happens over a fortnight, not a night.
Limited
Thin. Small studies, short follow-up, a single group, or a stand-in measurement rather than the outcome you actually care about. Plausible and under-tested.
Interesting, not directive. We report it so you know it exists, not so you rearrange your evening around it.
Mixed
Credible work disagrees, and the disagreement has not resolved. Different methods, populations, or definitions produce different answers.
Treat any confident claim in either direction with suspicion, including ours. We explain the disagreement rather than pick a side.
Insufficient
There is research, and it does not answer the question either way. Often because the thing being asked about cannot yet be separated from something it travels with.
Nothing here to act on. Where a mechanism points a direction we say so, and label it as mechanism rather than as outcome.
Not studied adequately
Nobody has properly looked, at least not in people. Animal work, laboratory models, and confident popular repetition all sit here until a human study exists.
We describe what is not known rather than filling the gap. A firm recommendation on this from anyone is not coming from evidence.
One question at a time.
A brief takes a single question, states what the evidence looks like, who it was studied in, and what is still unresolved. It is shorter than a guide and more specific: less "how to handle light", more "does this particular thing do what people say it does".
No briefs published yet
Briefs are published one question at a time, and only once the sources behind them are assembled. The scale above is already in force for everything that follows.