Any scientific study faces the dilemma of population or sample study. Obviously, a much higher accuracy would be obtained if the whole group, the population, were analyzed than a small representative portion, called the sample. It is observed that it is impracticable in most cases to study the population due to distances, cost, time, logistics, among other reasons.
The alternative practiced in these cases is to work with a reliable sample. If the sample is reliable and provides inferences about the population, we call it statistical inference. For the inference to be valid, good sampling is necessary, free from errors such as lack of correct population determination, lack of randomness and error in sample sizing.
When randomly removing two elements from the same population and exposing only one element to a given factor (advertising, for example). The impact on both elements is assessed.
Only one element is selected and exposed to the factor. A comparison is made by considering the before and after.Next: Sampling