Sample

The choice a researcher always has to make is how large a sample to choose. It stands to reason that the larger the sample, the more accurate will be the results of the study. The smaller the sample, the less accurate the results. Statisticians have developed mathematical formulas that allow them to estimate how accurate their results are for any given sample size. The sample size used depends on how much money they have to spend, how accurate the final results need to be, how much variability among data are they willing to accept, and so on.

Interestingly enough, the sample size needed to produce accurate results in a study is often surprisingly small. For example, the Gallup Poll regularly chooses samples of people of whom they ask a wide variety of questions. The organization is perhaps best known for its predictions of presidential and other elections. For its presidential election polls, the Gallup organization interviews no more than a few thousand people out of the tens of millions who actually vote. Yet, their results are often accurate within a percentage point or so of the actual votes cast in an election. The secret of success for Gallup—and for other successful polling organizations—is to be sure that the sample they select is truly random, that is, that the people interviewed are completely typical of everyone who belongs to the general population. When invalid populations are used, erroneous predictions, such as those that took place relative to the 2000 U.S. presidential election, often occur.