ABSTRACT
Subjects of a study are the entities that are measured, while the unit of analysis is the entity being analyzed. Research with human subjects in the United States is guided by the Common Rule as a result of historical racist abuse of power by some scientists. The Common Rule is guided by the principles of respect for persons (which leads to informed consent), beneficence (which leads to confidentiality), and justice (which leads to the establishment of the Institutional Review Board and special protections for vulnerable groups). Study samples are derived from populations, then narrowed to sampling frames, and potential subjects are selected through a sampling strategy that is either probability or non-probability. Sampling biases include selection bias and non-response bias. Study samples in HDFS should be representative of the population if they are to be generalizable: this is more important for some study designs (prevalence) than others (qualitative), with correlational and experimental designs in the mid-range. Sample size is not nearly as important as the representativeness of the sample.
