Evidence ratings
How we rate evidence
Last reviewed 16 May 2026
Every health claim on SleepUncovered gets one of four ratings. Here's what each means and what kind of research sits behind it.
The four ratings
Multiple high-quality randomised controlled trials or large cohort studies converge on the same conclusion. Effect sizes are consistent across populations. The finding has been replicated independently and is accepted in major clinical guidelines.
Examples: caffeine's effect on sleep, CBT-I's efficacy for chronic insomnia, sleep restriction increasing cognitive deficits.
Several reasonable-quality trials point the same way, but effect sizes vary, or replications are imperfect, or the evidence is still relatively new. The headline finding is likely correct; the details may shift.
Examples: melatonin's timing-shift effect, exercise on sleep quality, room temperature on sleep efficiency.
Some evidence exists but trials are small, short, or inconsistent. The mechanism may be plausible, but real-world effect sizes are unclear. We'll tell you the current best guess and flag the uncertainty.
Examples: magnesium for sleep in non-deficient adults, blue light filtering, most sleep supplements.
Early findings — promising but not yet replicated or confirmed at scale. We include emerging research because you might be researching it; we label it clearly so you know not to treat it as settled.
Examples: CBD for sleep, certain new pharmacological approaches, novel chronobiology findings.
How we decide which rating applies
We don't apply a rigid formula like GRADE; the site is editorial, not a systematic review. But we follow consistent logic:
- Strong requires either large replicated RCTs, or a finding accepted in major clinical guidelines (AASM, NICE, ESRS) with consistent underlying evidence.
- Moderate requires several decent trials pointing the same way, with some methodological caveats.
- Limited covers findings supported by some plausible evidence but not enough to call settled.
- Emergingis for early-stage findings that show promise but haven't been replicated at scale.
When evidence shifts
Ratings can change. If a major systematic review supersedes existing evidence, we update the rating and the article. The last-reviewed date on each article shows when we last looked.
Why we don't use a more formal system
Formal evidence frameworks (GRADE, Cochrane levels) are designed for clinical guideline-makers. They're thorough but inaccessible to most readers and over-engineered for much of what gets discussed in popular sleep content. Our four-level system is informally calibrated to GRADE but designed to be quickly understandable.