Control Group
A control group is the group in an experiment that does not receive the treatment under investigation, serving as the baseline against which the treatment group is compared.
Why not just measure the treatment group?
Without a control, you cannot tell whether observed changes come from the treatment, the Placebo Effect, natural recovery, regression to the mean, or just time passing.
Types:
- No-treatment control: receives nothing
- Placebo control: receives a sham treatment indistinguishable from the real one
- Active control: receives the current standard treatment (when withholding would be unethical)
- Wait-list control: receives the treatment later, so untreated during the comparison window
"Patients improved after the treatment"
Meaningless without a control. Many illnesses improve on their own, and patients seek treatment when symptoms are at their worst, so the next measurement looks like improvement (regression to the mean).
A control group is necessary but not sufficient. Pair with random assignment and blinding for a strong causal claim.