Sampling and Unbiased Sampling Methods

Chapter 5. Sampling: Sampling and Sampling Methods

Sampling and Unbiased Sampling Methods

Sampling and Sampling Method

Definition

Sampling is the procedure of selecting individuals from the population you want to investigate.

How the individuals that will make up the sample are selected is called the sampling method.

Unbiased Sampling Methods

Simple random sampling
Stratified random sampling
Cluster sampling

#\phantom{0}#
There are many different ways to select individuals from a population, but the main priority should always be the same: to select a sample that is as representative of the population as possible. Ideally, you want the sample to be a miniature version of the population. Inferential techniques are used to study a relatively small number of individuals in the hopes of generalizing the results of those studies to the larger population. It is, therefore, of great importance that the sample accurately reflects the characteristics present in the population.

One important way in which you can enhance the representativeness of a sample is to make sure to use an unbiased sampling method.
#\phantom{0}#

Unbiased Sampling Method

A sampling method is unbiased if all members of the population are equally likely to be selected for the sample.

#\phantom{0}#
Three examples of unbiased sampling methods are:

Simple random sampling
Stratified random sampling
Cluster sampling
#\phantom{0}#

Simple Random Sampling

Definition

Simple random sampling is a basic sampling method where the individuals of the sample are selected from the population as a whole, usually with the help of a random number generator.

Each individual is chosen entirely by chance and each member of the population has an equal probability of being selected.

#\phantom{00000}# Simple random sampling

#\phantom{0}#

Stratified random sample

Definition

When taking a stratified random sample, the population of interest is first divided into groups of individuals that share at least one common characteristic. These subgroups are called strata.

A simple random sample is then drawn from each of the strata and these samples are combined into a final sample that will be used for data analysis.

Stratified sampling guarantees that the final sample will contain at least some individuals from each group.

A drawback of stratified sampling is that extensive knowledge of the population is required in order to subdivide a population into relevant strata.

#\phantom{000}# Stratified sampling

Example

Imagine a researcher who is interested in the sleeping habits of Olympic athletes. To study these habits, the researcher travels to the 2018 Winter Olympics to interview the participating athletes. Close to #3000# athletes across #15# disciplines are scheduled to compete in the Olympics and the researcher decides to draw a sample of #200# athletes to conduct his research.

If the researcher would use simple random sampling to select the sample, it is quite possible that some disciplines would be over- or underrepresented in the sample. A simple random sample drawn from the population could, for example, contain a disproportionate amount of speed skaters or perhaps no snowboarders at all. In order to prevent this from happening, the researcher decides to use stratified random sampling to select the sample.

The population of Olympic athletes is first divided into strata based on the discipline the athletes are participating in. A simple random sample is then taken from each stratum in proportion to the size of the stratum compared to the population. So for example, if 150 out the 3000 athletes (5%) participated in the speed skating events, then the researcher wants his sample to contain 10 speed skaters (5% of 200). This procedure is repeated for all the other disciplines. Finally, all the sample subsets are combined into a final sample that can be used for statistical analyses.

#\phantom{0}#

Cluster Sampling

Definition

Cluster sampling divides the population into a number of subgroups, called clusters. Next, simple random sampling is used to select one or more of these clusters. The elements in each selected cluster are then sampled.

If all the elements in a sampled cluster are selected for the final sample, this is called one-stage cluster sampling. If simple random sampling is used to select a subset of elements in the sampled cluster(s), this is called two-stage cluster sampling. If this process is repeated more than two times, it is called multistage cluster sampling.

Quite often, the division of a population into clusters is done on a geographical basis, such as dividing a country into cities or a city into different neighborhoods or streets.

#\phantom{00}# One-stage cluster sampling