Randomized controlled trials: randomize, control, blind
Explains the three design features that let a randomized controlled trial (RCT) show a drug works — randomization, a placebo control and double-blinding — and the ways real trials fall short.
Where this fitsConnects to 2 · Unlocks 2
- Connects to
- Primary effects and side effects — Trials are how side effects are separated from coincidence.
- The course lens: biopsychosocial, and no 'chemical imbalance' — The psychological part of the biopsychosocial lens is what placebo controls account for.
- Unlocks
- Expectancy: placebo and nocebo effects — Why expectancy is powerful enough to need a placebo group.
- Phases of drug trials: from animals to Phase IV — RCTs are the core of Phase III.
Walk through
A new drug appears
Once a psychoactive chemical is thought to have medical use, it must be tested: is it safe, and is it effective? Clinical trials — specifically randomized controlled trials — answer this, because their experimental design allows claims about causality.
Analogy and worked example
Analogy: Testing whether new running shoes make you faster: if you choose who wears them (not random), you might give them to the fittest runners. If there's no comparison group, everyone may get faster just from training. If runners know they have the 'fast shoes', they try harder (not blind). An RCT gives identical-looking shoes by coin flip and doesn't tell anyone — including the timer.
A trial finds that 60% of people improve on a new antidepressant. Can we say the drug works?
- Ask: compared with what? Without a control group we can't separate the drug from expectancy and natural recovery.
Tip: the ← and → keys move through the lesson.