Your Account

What is Sampling?

Sampling is the process of selection of a subset of individuals from a population to estimate characteristics of the entire population. In statistical terms, a population can be thought of as a group or certain number of individuals defined by a particular characteristic (such as all individuals of species "x") or by separation from other similar groups in space (such as all individuals of species "x" in area "c").

Thus the word "sample" can be used to denote the subset of individuals collected from a population or the data of interest collected from the individuals of that subset.


Why do we sample?

When we want to learn about a particular group, it may be impossible to study each and every member of the group due to limitations of time and effort. Thus a small subset of the group is selected using an appropriate sampling method (such as random or systematic sampling) and studied. As long as the subset is selected through sampling in an appropriate manner, its characters will be representative of the entire group.

For example if one wishes to know what the average density of a particular species of plant is in an area, one can randomly sample the area using an appropriate unit of sampling such as a quadrat (which is basically a unit of area, the outer boundaries of which are shaped like a square) count the number of plants in the unit of area represented by the quadrat. Once a sufficient sample size of such quadrats is reached the average density of plants in the data set should be very similar to the average density of plants in the actual area beings studied.

Before going ahead it'd be good to go over an important aspect of statistical theory that is crucial in certain types of sampling.


The Importance of  Randomness


How do we Sample?

Different Statistical Sampling Methods/Frameworks -

The types of sampling methods can be broadly divided into two types. These are Probability Sampling and Non-probability sampling.


Probability Sampling (Also known as Random Sampling) -

Probability sampling is a type of sampling which always involves random sampling. 


Simple Random Sampling -

In this type of sampling the method for sampling is designed in such a way that each individual in a population has an equal chance of being selected. Sampling is then carried out based on this design.


Stratified random Sampling -

In this type of sampling, the various categories or subgroups of a population are demarcated before sampling and random sampling is carried out in each of these various groups. For example if you want to know the density of a particular type of plant in an area that is heterogenous (i.e. It is not internally similar to itself but has a variety of different components), like an area having forest, grassland, and farmland random sampling can be undertaken in each of these separate subcategories of the area.


Non-random Sampling -


Systematic sampling -

Systematic sampling involves an underlying design in the sampling process. This usually involves standardization of the sampling protocol prior to sampling based on a more or less subjective criteria such as sampling quadrats for plants by spacing the quadrats out a standard distance away from each other (usually in what would look like a grid like pattern). This has certain benefits as well as drawbacks. The benefit of this method is that very often the sampling may be spread out evenly across an area and more of the study area is covered in a shorter duration than would be possible by random sampling. It is also supposed to be more efficient as points in a study area closer to each other would be more likely to be similar to each other than points further away. The main drawback of this method is that the underlying non random design (such as deciding to sample using quadrats spaced evenly a certain distance away from each other), may cause the sample to be non-representative of the true population (see randomness).


Regarding sampling units -

A sampling unit can be thought of as a standardised unit of measurement in which samples are obtained when sampling. There is a large variety of standardised sampling units that are commonly used depending on what is being sampled. For example, quadrats (rectangular plots of varying sizes) are often used as sampling units when measuring plant biodiversity, quadrats are placed randomly and the number and type/species of plants found in each randomly placed quadrat is recorded. For fish a unit that is often used is "catch per unit effort" where, "effort" can be described as the number of times a net is cast to capture fish in it and the "catch" constitutes the number of fish and different types/species of fish being caught each time the net is thrown into the water.


Methods for Sampling Biodiversity


Randomization and Systematization of Sampling Methods


About the Author of this Article


McDonald, J.H. 2009. Handbook of Biological Statistics, 2nd ed. Sparky House Publishing, Baltimore, Maryland.