The different sampling methods and how to choose

The different sampling methods and how to choose


The sampling method determines the reliability of the results of market research or data analysis. This involves selecting a representative sample of the population, in order to draw useful conclusions from an optimized survey in terms of the company’s resources. Two types of methods can be distinguished: random or probability sampling and non-random sampling.

> Download this free kit and do in-depth market research. ” align=”middle”/>The choice naturally falls on the non-random sampling method. Illustration: the company’s after-sales service deliberately isolates customers who have reported malfunctions on a product to identify frequent breakdowns. In theory, however, the probabilistic method is more rigorous, because the sample constituted on a statistical basis is more likely to be representative. In the example: the after-sales service draws lots from a determined number of customers who have purchased the product; in this case, the company not only identifies frequent failures, but also measures their frequency to assess whether or not they are significant. The analysis is enriched, and the results are more usable.

When the company does market research, data analysis, a poll or any other type of survey, the first step is to define the target population. When the volume is considerable, querying the whole is not realistic, or too consuming in time, money or human resources. To streamline the effort, the company selects a sample. It is at this time that a sampling method must be chosen. Prior to its implementation, the company determines the relevant size for the sample, ie the number of individuals who make it up.

The different sampling methods

Non-random sampling methods

Non-random, or non-probability, sampling methods consist of selecting individuals from the target population on the basis of subjective criteria. Five methods coexist.

  • Convenience or convenience sampling: the interviewer questions individuals at random. The individual is accessible and ready to answer, this practical aspect facilitates the task of the interviewer. Example: the restaurateur takes the time to gather the opinion of a few tables at the end of his evening service.
  • Voluntary sampling: the interviewer proposes the questionnaire to the entire target population or to a specific segment. It draws conclusions from the answers obtained on a voluntary basis, without consideration for the silent majority. Example: the marketing team sends a satisfaction questionnaire by email to the customer segment of the online store in its contact base, to identify recurring points of dissatisfaction.
  • Judgmental or purposive sampling: the interviewer makes the decision to interview a specific segment of the target population, based on his own reasoning which leads him to assume that the segment is representative. In the previous example: the marketing team only solicits the opinion of online customers whom it has observed systematically opening the weekly newsletter, excluding a large part of this segment from the contact base.
  • Sampling by quotas: the interviewer questions equal parts of categories of the population. Each category is defined on the basis of objective criteria such as gender, age or socio-professional category. In the example: the satisfaction questionnaire is sent to 20 men and 20 women from the contact database, it does not matter that the online clientele is mainly made up of women.
  • Snowball sampling: this sampling method is used when the target population is small. The aim is to expand the number of respondents by calling on the network of each individual questioned: each new recruit is invited to involve relatives who share the specified characteristics with them. Example: the company is interested in the potential of a range of organic food for cats; she contacts the few people she has identified have a cat, and asks them to broadcast the survey.
See also  which professional social network to choose?

These relatively instinctive non-random sampling methods have advantages and disadvantages:

  • The advantages of non-random sampling: the method is simple and quick to implement, the upstream thinking is minimal and the effort required to obtain answers is considerably reduced.
  • The limit of non-probability sampling: the composition of the sample is not based on a system of probabilities. Under these conditions, the panel of respondents does not necessarily include all the types of profiles relevant to the survey, and the profiles are not represented proportionally. The results therefore risk being biased and must be analyzed with caution.

Random sampling methods

With random, or probability, sampling methods, each individual in the target population has a probability of being interviewed. Statistically, each respondent profile is represented proportionally. Four random sampling methods can be implemented.

  • Simple random sampling, EAS: the investigator extracts the list of contacts from the target population, then has a sample generated automatically on the basis of chance. It’s sort of a draw, where each individual has the same chance of being selected for the survey.
  • Systematic sampling: a variant of EAS, this sampling method does not use chance to select individuals. Each contact on the list has a number, and the interviewer constitutes the sample at regular intervals. Illustration: the customer base comprises 10,000 contacts; the interviewer chooses the interval + 10: he selects contact no. 1, then no. 11, then no. 21, and so on until the database is exhausted.
  • Cluster sampling: the whole population is divided into groups, then the interviewer randomly selects a defined number of groups. He interviews all the individuals in each group. Example: the restaurant chain does a nationwide satisfaction survey; the company can hardly interview all customers; she chooses 3 clusters: the restaurant in Lyon is the first cluster, the restaurant in Bordeaux is the second cluster and the restaurant in Bayonne is the third cluster. All customers from Lyon, Bordeaux and Bayonnais are interviewed.
  • Stratified sampling: this sampling method is implemented in two stages. First, the investigator divides the target population into homogeneous sub-groups, on the basis of objective criteria. In a second step, the interviewer selects a sample within each stratum, using the random sampling method of his choice. Example: the company is thinking about the terms of payment for its new SaaS software, it is hesitating between a monthly subscription and pay-per-use and wishes to query its contact database for this purpose; the contact base is divided into three strata based on income level; for each stratum, the company uses cluster sampling: it surveys all low-income mobile app users, all middle-income computer users, and all high-income tablet users.
See also  How to Earn Money on TikTok? 6 Complementary Methods

Despite some practical limitations, probability sampling methods are recommended for their increased reliability.

  • Disadvantages of random sampling methods: an in-depth analysis of the survey objective and the available population data must be carried out beforehand. The interviewer must also obtain the participation of the entire sample, which requires convincing recalcitrant individuals to respond. Last obstacle: the high number of responses implies a significant downstream workload.
  • The advantage of random sampling methods: the results obtained from the sample can be extrapolated to the entire population. On this basis, the company is able to base strategic decisions.

How to choose a representative sample?

To take advantage of the results of large-scale market research, it is essential to obtain responses from a representative sample. The sample is representative when it constitutes a faithful image of the target population, not only in terms of quantity but also in terms of standard profiles.

To choose a representative sample:

  1. Determining the sample size: Of course, the sample size should be proportional to the population size. The choice of sample size is also dictated by the company’s own operational constraints. Illustration: questioning a considerable number of individuals by telephone, when the results of the investigation are urgent, seems unrealistic.
  2. Choose a random sampling method: the representativeness of the sample is essentially linked to the choice of the sampling method. The investigator who has the necessary time and means preferably opts for a probability sampling method, in order to exploit the results in a secure manner. Simple, systematic, cluster or stratified random sampling: the method is determined with regard to the number of individuals to be interviewed on the one hand, and the technical means available on the other.
  3. Eliminate errors: When consolidating survey results, significant discrepancies can be identified to eliminate responses that appear to be anomalies.
See also  How to make a flowchart? (+ examples)

To go further, download this free market research kit and use it to learn more about your customers, your industry and your competitors.Bottom-CTA: Market Research Kit

Leave a Comment

Your email address will not be published.