How do pollsters go about choosing people to answer their surveys?

Answer 1

The first stage, in any survey, is to establish some kind of register - the "sampling frame" - for the whole population that you wish to study. The accuracy of this frame can greatly influence the reliability of the survey results.

The simplest method then is the simple random survey in which each member in the population is allocated a unique identifier. A suitable sample size is selected. Then a random number generating procedure is used to select identifiers until the required number are selected. This is the list of people who will be surveyed.

Although statistically this list is representative of the whole population, you may be unlucky and get too many people of one kind or another. Also, the variability in the responses may differ between different subsets of the population. To overcome this, pollsters use stratified random sampling.

In this scheme the total population is cross-classified into subsets, or strata, according to a number of criteria. For each criterion, people in the same subgroup should comprise of broadly similar people while people in different subgroups should be different. The stratifying criteria may be gender, age group, race, religion, educational attainment, income band or social group, political affiliations - in general any factor that is thought likely to affect the members' responses to the variable(s) of interest. Information on some of these factors may be available from the sampling frame. The rest would have to be established using preliminary profiling questions.

A simple stratified scheme would be to treat each cross-classified subgroup as a population and take a simple random sample representing the same proportion in each subgroup. A more sophisticated process would be to estimate the variability (standard error) of the response variable(s) in each substratum. Then the sample size within each substratum is selected so as to bring the standard errors as close to one another as possible. Again, a simple random sample is chosen from each sub-population.

Each stage in the process:

• establish the sampling frame
• identify stratifying criteria
• estimate variability of response in substrata

is a highly skilled process which could requiring substantial resources, including access to large sets of data and the results of related surveys.

Answer 2

Pollsters typically use random sampling methods to select a subset of individuals from a larger population to participate in a survey. This helps ensure that the sample is representative of the population being studied. They may use techniques such as telephone surveys, online surveys, or in-person interviews to collect data.