To get input about a certain population, such as the students at a particular university, it is convenient to use a representative sample of students. The researcher gets input from this sample and extends the results of the research to the entire population. This method simplifies the process of research. There are different ways to get a statistically sound sample from the population. One such method is proportional allocation, which is a sort of stratified sampling method.
Stratified sampling divides the population into different strata based on a particular characteristic. For instance, a researcher could divide the population based on income into a low-income stratum, a middle-income stratum and a high-income stratum. The researcher should choose the characteristic in such a way that the samples chosen from within each strata are as representative of the strata as possible.
After the researcher splits a population into different strata, the question of how many people to sample from each stratum arises. If one stratum consists of 1,000 people, for example, and another of 2,000 people, it is necessary to draw samples that represent these larger groups in an adequate manner. One method of drawing samples from different strata is proportional allocation. In this method, the researcher draws the same proportion of people from each stratum, such as 5 percent of the stratum, to serve as a sample.
One major advantage of proportional allocation is that this is a simple method to execute. Choosing 5 percent of the population from each stratum is a relatively easy technique. There are other methods of sampling that entail drawing different numbers of people from each stratum, so as to adequately represent the diversity in views of the people in each stratum.
Another advantage of proportional allocation is that it produces a sample size that is representative of the size of the stratum within the population. If, for instance, one stratum consists of 1,000 people and another of 2,000 people, a proportional allocation could draw a sample of 1 percent from each stratum. This means that the researcher would choose 10 people from the first stratum and 20 people from the second stratum. Since there are more people in the second stratum than the first stratum, this sample is more representative of the population than choosing equal numbers of samples from each stratum.