ABSTRACT
In experiments, a special type of sampling bias results from what researchers call mortality. Because mortality does not occur at random, the result is presumed to bias the sample. Even if a researcher does not make a statistical adjustment based on the difference between the demographics of a population and a sample, it is still worthwhile to know how closely a sample reflects a population on important demographics. By collecting demographic information on the participants, the researcher can provide the reading audience with a more accurate description of the individuals who constituted the sample for the study and how their responses differed across the groups. If a researcher draws a sample of the population and obtains a sample in which only 10% of the respondents are Asian American, he or she can make adjustments when completing statistical analysis to weight the results of the Asian American respondents in the sample.
