19.08.2015

Research sample selection key to valuable conclusions - ConQuest

Market research is one of the most popular methods of obtaining information used by companies and institutions around the world. Professional analysis can be a very valuable source of knowledge, and the subject matter of research itself is almost limitless. Conducting such a survey seems easy when the target group is a few, a dozen or a few dozen individuals, but how to conduct a survey on a population of several hundred thousand? It would be almost impossible to interview all the residents of Wroclaw or to check the quality of every egg in a 300,000-strong shipment. The only fair solution in such a case is to select "representatives," i.e. a specific part of a given collectivity that will take part in the survey.

What, then, makes it possible for us to analyze a study conducted on even a hundredth part of the population under study by reference to the whole, maintaining the reliability of the conclusions drawn? The key is the selection of the research sample. A properly selected research sample has a fundamental impact on the quality and credibility of the entire study. The selection of the sample is a key step in the implementation of the analysis and thus poses quite a challenge. In order to make a proper selection of a research sample, it is important to keep in mind the four basic steps of the process:

  • Defining the target population
  • Determine the sample size (this step has a direct bearing on the accuracy of our analysis - the larger the group of people who take part in the survey, the more accurate the results will be)
  • Choice of sampling method
  • Sampling the plan

The sampling method should be decided based on the type of information sought. Unfortunately, it is often determined by the availability of respondents. This process can be carried out according to two schemes:

Random (probabilistic) testing

As the name suggests, this selection scheme involves the use of random sampling tools. The researcher has no influence on the selection process, and all individuals in the target group are equally likely to participate in the study. Within the framework of a random survey, several techniques are additionally distinguished.

The first of these is simple random selection. This is the easiest of the random methods and involves drawing an appropriate number of units from the population under analysis.

Another method is stratified random selection. The first step in this method is to characterize the structure of the population in terms of a particular characteristic or combination of characteristics. Then draw an appropriate number of units from each stratum so that the sample structure agrees with the population structure.

Example:

If in a given population the number of smokers is 45% and the number of non-smokers is 55%, the researcher, taking into account addiction as a characteristic that shapes the population structure, should draw such a number of smokers and non-smokers that in the drawn sample they constitute 45% and 55%, respectively.

The third method is systematic random selection. It involves setting a certain frequency of drawing individuals from the population, e.g. every tenth person in the study group will be selected.

Team selection as another technique included in the random survey is used in the absence of having a complete list of units of the analyzed population. It is necessary to draw teams (can be, for example, localities, institutions, schools) in order to obtain lists of their members, and then, based on them, a random draw is made.

Non-random (non-probabilistic) survey

Unlike the first scheme, a non-random survey rejects the random sampling procedure. It is also not fully independent, since the interviewer is given the opportunity to influence the selection of units, or more precisely, the mechanisms used.

One of the techniques of a non-random survey is purposive selection. It consists in selecting such units that are the source of the most valuable information on a given phenomenon.

Another technique is random selection. This method is based on the availability of the interviewees. An example would be, for example, an interview on the street, when the interviewer quizzes people who happen to be within his reach.

The third way is quota selection. It is very similar to one of the random survey methods, since it too is based on describing the structure of the population under study in terms of a given characteristic or combination of characteristics, and then selecting such a number of people that the sample structure agrees with the structure of the population under study. The difference is the omission of the random sampling process. The researcher independently selects the people who will participate in the study.

There is another sampling method included in the non-random survey. It is snowball method used when analyzing hard-to-reach communities. It involves reaching out to a small group of people exhibiting characteristics of interest, and then through their acquaintances to more people with similar characteristics.

 

Market research is an integral part of the operation of most businesses and institutions. It forms the basis of many important business decisions, so it is important that the conclusions drawn coincide with the truth as accurately as possible. Correct selection of the research sample, which is the key stage of any survey, will certainly increase the reliability of the analysis and contribute to making the right decisions.

Robert Tomaszewski

See other entries

barriers to entry
Market analysis

What really are barriers to entry?

See more
3d printing
Innovations

Development and commercialization of 3D printing in medicine

See more
business consulting
Enterprise management

What is business consulting?

See more