Sample

The sample is the group you measure in order to speak about the group you do not. Two properties decide whether that is allowed: size, which sets the error margin, and representativeness, which sets who the conclusion is about. Of the two, representativeness is the sterner master: a thousand self-selected respondents say less than two hundred well drawn. The classic pitfalls are self-selection, measuring whoever is easy to reach, and silently generalising from customers to the market. Reading research therefore starts at the draw: who is in, who is missing, and for whom does the conclusion then hold?

Covered in