Obtain a representative group to measure variables by
Random sample = every person in the population being studies has an equal chance of inclusion
but allow for some biases:
Unrepresentative samples
Organization of survey (order of questions, response options, wording)