Only accepted work is recordedAboutGuidesContactSign in
CitePepCitePep
Negative results

Null findings and failed replications, on the record, so nobody repeats them blind.

Most experiments that don’t work are never published, so other labs run them again. Here a null result or a failed replication can be recorded in a few minutes, checked by an expert, and linked to the claim it tested.

Record a null resultSearch null results
Null resultRecorded 12 Sep · checked by ReviewerProbiotic dosing did not reduce thermal stress in Pocillopora under +2°CA second independent lab · n = 96 colonies · pre-registered
Tested claimSantos et al. (2024)Probiotics cut thermal stress by 30%
ResultNo difference (Δ 2%, 95% CI −6 to 10)
Why it mattersSaves ~8 months of tank time
Linked to1 living review · 2 replication calls
Counts on the lab’s record like any other contribution
How it works

Small to record, valuable to find

1Describe itThe claim you tested, your design, n, and the result with uncertainty.
2Attach dataOptional but encouraged. It gets its own data record.
3An expert checks itFor design and analysis, not for whether it “worked”.
4It’s findableLinked to the claim, the living review and any replication call.
Recent null results

What didn’t work, so you don’t have to find out again

Null resultChecked · 12 SepProbiotic dosing did not reduce thermal stress in Pocillopora under +2°CA second independent lab · n = 96 · pre-registered
Tested Santos et al. (2024)
Failed replicationChecked · 3 SepA 10-minute writing task did not reduce exam anxiety5 labs · n = 1,240
Tested Moreno & Chen (2024)
Null resultChecked · 22 AugVitamin D did not change HbA1c over 6 monthsKarolinska · n = 310
Tested a common supplement claim
Null resultAwaiting checkShading nets did not lower thermal stress at 4 m depthA marine research centre · n = 40
Tested field guidance