Advantege of sampling in any Survey?

Time and cost.





There is not enough man hours available to survey everyone...and not everyone will give you a response.

Advantege of sampling in any Survey?
Cost.





It is cost prohibitive to survey everyone in the population to get their opinion or find out some attribute about them. If we want to know if Americans agree with some point or another, we can't afford the time and cost to talk to over 200 million Americans. So, by sampling, say, 1000 people, you can get a "good enough" answer without talking to everyone.





This is usually accomplished by either a pure random sample or a stratified random sample. A stratified random sample for a survey might use "strata" to ensure that people of all age groups, gender, and socio-economic status are surveyed in proportion to their occurance in the population and not concentrated into those that might happen to be at home when the survey was performed. For instance, a pure random sample conducted during the day might bias toward the unemployed and stay at home moms and be less likely to include employed people.





These samples will have an "error" associated with them. This might be expressed as 51% of Americans agree with such and such, +/- 1%. This error is the "standard error" and is an estimate of how different the answer would be if you conducted the survey again with a different random sample of people.





Let's say we want to cut the error in half. So we now know the answer that 51% of Americans agree, +/- 0.5%. One aspect of sampling is the central limit theorem and it tells us that to cut the error in half requires a sample that is four (4) times larger, hence four times costly.