Non-probabilistic sampling – What is it, definition and concept | 2022

Non-probabilistic sampling is one where not all the subjects of the statistical population have the same probability of being chosen to be part of the study that is being developed.

That is, this type of sampling implies that the interviewer or researcher does not select randomly or randomly, among the entire population, the individuals that are part of the sample on which he works.

The above, in principle, may not seem to correspond to what would be a reliable investigation. However, as we will see throughout this article, there are circumstances in which non-probability sampling is justified.

Before continuing, we will make some concepts clear. First, the statistical sample, which is a subset of data that belongs to a larger one that we call the population. The latter comprises the total number of individuals who share a characteristic that is to be investigated.

For example, the population could be Spaniards between 30 and 45 years old. Meanwhile, a sample would be 1,000 adults who meet this characteristic and who were selected in various cities in Spain.

Types of non-probability sampling

Among the types of non-probabilistic sampling we can mention

  • For convenience: In this case, the investigator selects those individuals who are closest to him or who are more accessible. Example: A university professor conducting research among students at the university where he teaches.
  • by installments: In this case, a non-probabilistic sampling is done because it seeks to meet a certain condition. For example, that 50% of the sample are men, and the other 50%, women.
  • discretionary, opinionated or intentional: The researcher, according to his judgment, chooses a group of people who he considers to be suitable for the study. For example, if you want to conduct a survey of doctors working in a hospital, the researcher could select those whom he considers to be more suitable because he trusts their judgment.
  • Snowball: In this case, it is usually about investigating a group of people with very particular and unusual characteristics, or when said information is not easily accessible. Then, the researcher, once he finds some people with the desired characteristic, asks them to call other individuals with the same characteristic. For example, this could be in research on a rare disease.
  • casual or accidental: The researcher selects, without prior judgment, the individuals who will be part of the sample. For example, this often happens when street surveys are conducted. The interviewer goes at a certain time and to a specific place in the city and asks questions of the people who pass by and they agree to answer. Not all the target population has the same probability of passing through that street and, moreover, of meeting the interviewer.
See also  No more lies: 89% of funds did not beat the market

Advantages and disadvantages of non-probability sampling

Among the advantages of non-probabilistic sampling we can mention:

  • It implies lower costs and time because the researcher can choose those subjects that are easily accessible.
  • It is possible to control the characteristics of the sample, as we explained in quota sampling.
  • It is ideal for studies on populations with unusual characteristics or handling sensitive data. For example, people who are in a situation of illegality in a foreign country. It is also used for pilot studies or that allow us to know roughly the characteristics of a population.

However, there are also some disadvantages of this type of sampling:

  • Subject selection, especially in the case of discretionary sampling, can be affected by researcher bias. This would lead to the data not being entirely reliable.
  • It does not always guarantee adequate representation of the target population, again, especially if the selection is discretionary.
  • The results, given the points mentioned above, are often not entirely accurate.

Example

It seeks to conduct a study on young university students in Italy. The researchers choose four universities in the country, from different cities, and from them they select students interested in participating in the research.

In this case, it is a non-probabilistic sampling because the entire Italian university population is not being considered, even though this is the target audience for the analysis.

Leave a Comment