What is a randomized controlled trial (RCT)?

A randomized controlled trial is an experiment that splits volunteers into two groups by chance. One group gets the real treatment, the other gets a dummy version, and nobody picks their own side. Because chance makes the two groups alike, any extra improvement in the treated group points back to the treatment itself.

Say a company believes its new pill lowers blood pressure. The obvious check, comparing people who take it against people who don't, may be broken before it starts. People who seek out a new pill are often sicker or wealthier or more health obsessed, and any of that moves blood pressure on its own. An RCT removes the choosing: a coin flip decides who gets the pill and who gets a dummy pill with no active ingredient, so both groups carry the same mix of sick and healthy, and the pill is the one systematic difference left. The dummy matters as much as the randomizing, because hope and attention improve people too, and both groups deserve equal doses of hope. Usually neither the patients nor their doctors know who received which, the arrangement called double blind. The costs are real: volunteers, money, years, and some questions cannot be randomized at all, since nobody can assign people to smoke for a decade. But when a claim is called the gold standard of evidence, this design is what earned it the title.

A gardener testing fertilizer on one continuous half of a field learns nothing, because the sunny half was always going to grow better. So he cuts the field into small plots and lets a coin decide which ones get fertilizer. Sun, soil and drainage now land evenly on both sides, leaving fertilizer as the one thing the treated plots have that the others don't. When those plots grow taller, it has to be the fertilizer talking.