The Salk polio vaccine trial.

The exact design and results are detailed in the text, pages 4-6.

Several points to note:

• The size of the experiment can't make up for a bad design.
• There two effects to be careful of:
  • The Hawthorne effect, just being part of a study changes people's reactions.
  • The placebo effect, believing that a drug is going to help, talking to a doctor can start a chain of improvements (see below for a good example).
• The reason for randomizing is to avoid confounding some unknown factor that volunteers or chosen subjects may have (rats in higher cages, all one grade level, parental consent) with the treatment effect.

Randomized controlled double blind (placebo) is the best design for a simple experiment.