As they decide which content to include, survey designers also specify the population of households, persons, providers or organizations that will be covered by the procedure. Selecting the population also defines the limits of statistical estimates that can be validly reported. Statisticians implement the definition of a population by specifying its frame.
A survey may collect information from each member of the population or a sample that allows analysts to infer results for the entire specified population. Probability procedures allow analysts to infer statistical results for an entire population based on information collected from a smaller sample. The sampling procedure selected determines the weights analysts apply to generate valid estimates for the entire population—or specific parts of it.
When designing a survey, researchers and statisticians use a standard questionnaire to collect information. Each respondent is asked the same questions and information is recorded in the same way.
This differs from procedures typically used in diagnostic interviews where practitioners may probe widely and vary their questions based on observation and the health conditions they observe.
Surveys therefore involve less professional judgment and rely more on tested and validated procedures than on an extensive knowledge of the field. By relying on standard procedures they maximize comparability but loose detailed information and the expert insights clinicians provide.