The evidence library

What trials measured, outcome by outcome.

For each substance and each thing people take it for: one grade, the direction trials found, when they measured it, and every study behind it. Open to anyone, no account needed.

Coming soon Nothing published here yet. Every entry is read by a person before it appears.

How strong is the evidence?

One scale, in one blue that gets stronger with the evidence. Blue shows strength, not approval: there is no green for “good” and no red for “bad”.

The last two are not grades. “No effect found” is a finding. “Not studied” is a gap, and often a popular claim.

  1. High Consistent results across well-run trials; more research is unlikely to change it.
  2. Moderate Probably right; more research could change the size of the effect.
  3. Limited Some trials, inconsistent or small.
  4. Very limited A few small or low-quality trials.
  5. No effect found Trials measured it and consistently found nothing. A finding, not a gap.
  6. Not studied No trial measured it. Often a popular claim.

One grade, one outcome.

A grade belongs to one outcome, never to a substance. There is no overall score, no stars, no rating.

Never on dooz Substance: moderate

Moderate for what? A substance studied for six things has six answers.

How every grade reads Substance · outcome: grade

One substance, one outcome, one grade, with the studies right underneath.

How an entry is laid out.

A layout, not a published entry: the bracketed parts are where the reviewed facts go.

  1. The pair. The substance and the one outcome the grade is about.
  2. The grade. A level out of four, in words as well as blue.
  3. The direction and the clock. What trials found, and how long they ran before measuring it.
  4. The studies. Every one, linked to its source. Free, always.
1 [Substance] · [outcome]
2
Evidence: [grade], [n] of 4
3 · Measured[what trials found]
When trials measured it[after n weeks]
4 · [n] studies
[Study title] · [journal], [year] PubMed
[Study title] · [journal], [year] PubMed
[Study title] · [journal], [year] PubMed

Two different clocks.

How long a substance stays in the body and how long trials took to measure an effect are different clocks. Each claim says when its effect was measured.

Sources are never behind a paywall.

Every grade, every direction and every citation is free. dooz Plus buys depth against your own log, never access to the sources. What Plus adds.