Glossary
Publication bias.
Direct answer
Publication bias occurs when studies with a clear or welcome result appear more often and more quickly than studies finding no difference. The accessible literature then presents an unduly flattering picture, and any synthesis of that literature, however careful, inherits the same imbalance.
The fault does not lie with journals alone: null results are frequently never submitted, being judged uninteresting. Certain patterns raise suspicion, such as small studies showing almost nothing but positive findings. The most effective countermeasure remains registering a study publicly before it begins.
In equine research publication bias weighs heavily, because many studies are small and carried by manufacturers, federations or clinics. A small study finding no difference slips easily into a drawer. Human medicine requires clinical trials to be registered, whereas veterinary registration remains distinctly patchy.
The evidence letter
Once a month: new syntheses, checked studies, and what has changed in the state of the evidence.