Authors
- Stella Lampret Puž, Faculty of Health Studies, University of Rijeka, Department of Basic Medical Sciences, Rijeka, Croatia
Article type:
Review
Abstract:
In scientific research, the population represents the largest statistical group, encompassing all subjects studied within a specific time and place. Due to inherent variability among subjects, samples drawn from a population can never be identical but only similar. It is often impractical to study the entire population because of constraints such as limited research time, financial resources, and the actual availability of subjects. Therefore, researchers use sampling methodologies to select a subset of the population, called a sample.
Sampling not only produces a sample but also introduces sampling error, which is quantified statistically by the standard error. The key characteristic of a sample is representativeness, which depends on correctly calculating the sample size and selecting subjects randomly from the population. This distinction differentiates two main sample types: probabilistic (random selection) and nonprobabilistic (non-representative) samples. Probabilistic sampling methods include simple random sampling, stratified sampling, systematic sampling, and cluster sampling, while nonprobabilistic methods include convenience and quota sampling.
The sample size is the primary aspect a researcher can control during study planning. To ensure representativeness, inclusion and exclusion criteria must be clearly defined so that participants have relevant characteristics affecting the association under study. Sample size directly influences the statistical power of a study and must be calculated using appropriate statistical methods as discussed in this paper.
Keywords:
calculation, population; sample; sampling; standard error