Episode number 13 of Series “Against all odds: inside statistics”.
When setting up an experiment one must consider if the subjects differ in any systematic way that may effect the results. Blocking is used when subjects share the same characteristics but when there may be some variability, as in agricultural experiments, such as the example shown, of testing new varieties of strawberries. A census counts everything but obviously has inherent difficulties which are illustrated. Sampling is a way of circumventing these difficulties and some of the problems of true representation are shown.