Random Selection

Random Selection

Introduction

Selection refers to a situation in which data are not representative of the underlying population of interest. In particular, selection occurs when unobserved factors that determine whether an observation is in the data set also help determine the value of the quantity of interest. For example, in a study of willingness to volunteer, it would make little sense to only select those participants who volunteer to participate. An analysis of data that suffer from selection generally produces biased estimates and renders statistical inference problematic. Solutions to the problem commonly involve modeling the selection process and the outcome of interest at the same time to account for how selection influences the observed values. (Cochran, 1977)

Random selection is a precise, scientific procedure whereby each unit in a population has an equal chance of selection for inclusion in a sample. This concept underlies the premise of probability sampling and is central to probability methodologies and generalizability. Random selection eliminates sampling selection bias but introduces random error in its place. Random selection is the only valid strategy for obtaining a representative sample in research.

Each methodology emphasizes different sample selection procedures. Probability sampling methods all use a random selection procedure to ensure that no systematic bias occurs in the selection of elements. The term random selection is commonly used in the context of experimental research, while the term random sampling is more common to survey research; however, the underlying principle is the same.

The most simple and common example of random selection is the tossing of a perfect coin. Each toss of the coin can result in two outcomes, heads or tails, with both outcomes having an equal probability of occurrence, that is, 50%. Moreover, each event is totally independent of previous selections and is also mutually exclusive, that is, a head and a tail cannot occur at the same time.

In practice, researchers sample an accessible population because a sample of a theoretical population would be quite an arduous undertaking and the cost of such a sample would be quite prohibitive. In addition, trying to obtain a sample from a theoretical population would be quite difficult because no accurate listing of the population may exist.

Samples therefore offer a practical solution to the study of populations. Sampling may thus be defined as the process of selecting units from a population of interest. It is a key component of sound research design because the manner in which ...