Science Puzzle
The Missing Half of the Trial
A company tests a new cough syrup on 100 people with coughs. Within a week, 80 of them have recovered. The company announces that the syrup is 80 percent effective.
What is the fatal flaw in this claim?
The Answer
There is no control group. The number 80 percent is meaningless on its own, because coughs get better by themselves. Most people with an ordinary cough recover within a week whether they take anything or not.
If you had given 100 similar people nothing at all, perhaps 78 of them would have recovered anyway. In that case the syrup did almost nothing, and the impressive-sounding 80 percent is simply the recovery rate of a cough.
A control group exists to answer one question: compared to what? It supplies the number you must subtract before you can claim any effect. Without it you cannot separate what your treatment did from what time, the body, and chance were going to do regardless.
The other options are not nonsense. A company testing its own product is worth noting, and sample size matters. But neither is fatal here. You could fix the sample size, use independent testers, and follow up for a year, and the claim would still be worthless, because there is still nothing to compare it against.
The principle: The control group. An outcome figure means nothing without a comparison group that received no treatment, because it cannot otherwise be separated from what would have happened anyway.