Characteristics of probability and non-probability samples

Currently the demands of the world are increasing, therefore it is up to us to adapt and survive the changes that the world demands. In the business field, it is not enough to have knowledge only about administration, but we must be able to handle an infinity of topics and develop them, since there is a lot of competition.

In the following work, the topic of probabilistic and non-probability sampling is developed, in order to provide quality information to users who need to support or learn more about the topic in a simple way, which can be used in various branches of studies and just to mention one, the business environment.

non-probabilistic-sampling-probabilistic-yesenia

It is not enough to be able to do things, many times you need the knowledge and the desire to make contributions, in order to develop the requirements that are greater each time. We are thinking and conscious beings, therefore we must act responsibly both in our actions and in those around us.

This work, as you can see, was prepared from the topics seen in probability and statistics class, from a business degree, where it is very important to have the necessary knowledge to face the world and humanity.

SAMPLING

It is the process of selecting a set of individuals from a specific population in order to study them and characterize the total population.

Here are two fundamental terms of sampling:

Universe or population: It is the total of individuals that I want to study or characterize Sample: It is the set of individuals of the universe that I select to study them, for example through a survey.

ADVANTAGES OF SAMPLING

It considerably summarizes the time and costs of the studies that are carried out in addition to reducing the resources that are used. The operation of the data is simpler.

DISADVANTAGES OF SAMPLING

We introduce (controlled) error in the result, due to the very nature of the sampling and the need to generalize results. We are at risk of introducing biases due to poor sample selection. (Jan, 2009)

PROBABILISTIC SAMPLING

Probabilistic sampling will be discussed as long as two conditions are met:

The elements of the population must have a probability greater than zero to be selected in the sample. There is precise knowledge of the probability of each element.

As we can see, we can only do probability sampling if we have a sampling frame. Subsequently, the form used to select the sample defines the different probabilistic sampling techniques, which are shown below.

SIMPLE RANDOM SAMPLING

The simple random sampling (MAS) This is the sampling technique in which all the elements forming the universe and are detailed in the sample frame are equally likely to be chosen for the sample.

Within the simple random sampling are two subdivisions, which are defined according to the possibility that the individuals of the universe can be selected more than once in the sample, it is MAS with replacement or without replacement.

If the replacement is used, when at the time of selecting an individual at random for the sample, it does not matter that he is selected again in a subsequent selection.

If replenishment is not used, an individual selected for the sample once would no longer enter the drawing again.

The following expression is used for the sample size in a MAS without replacement. The formula relates the required sample size when the universe is finite to the required size when the universe is infinite:

Where n0 is the sample size required for an infinite universe and N is the size of the finite universe. It is possible to demonstrate that the sample size when we use replacement (nr) is always equal to the size necessary for infinite universe (nr = n0).

ADVANTAGES OF SIMPLE RANDOM SAMPLING

It is fast and reliable since there are currently software that minimize the work when selecting individuals making it more reliable. The error due to chance can be calculated accurately

In this way, when using MAS we ensure that we obtain representative samples, so that the only source of error that will affect the results will be chance.

DISADVANTAGE OF SIMPLE RANDOM SAMPLING

 Difficulty putting it into practice in real investigations. Being a probabilistic technique, a sampling frame is needed with all the individuals and that all are selectable for the sample.

STRATIFIED RANDOM SAMPLING

"This technique, belonging to the family of probabilistic samplings, consists in dividing the entire population under study into different subgroups or disjoint strata, so that an individual can only belong to one stratum. Once the strata have been defined, individuals are selected to create the sample using any sampling technique for each of the strata separately. If for example we use simple random sampling in each stratum, we will speak of stratified random sampling (MAE from now on). Strata are usually homogeneous groups of individuals, which in turn are heterogeneous between different groups.

It is relatively common to define strata according to some characteristic variables of the population, such as age, sex, social class, or geographic region. These variables allow the sample to be easily divided into mutually exclusive groups and, quite frequently, allow different behaviors to be discriminated within the population… ”(Jan, 2009)

TYPES OF STRATIFIED SAMPLING

Depending on the dimension assigned to the strata, we will talk about different types of stratified sampling. It is also customary to speak of different forms of "affixation" of the sample in strata.

STRATIFIED SAMPLING PROVIDED

“When we select a characteristic of the individuals to define the strata, it usually happens that the size of the resulting subpopulations in the universe is different.

UNIFORM STRATIFIED SAMPLING

We will talk about a uniform allocation when we assign the same sample size to all the defined strata, regardless of the weight these strata have in the population.

OPTIMAL STRATIFIED SAMPLING

In this case, the size of the strata in the sample will not be proportional to the population. On the contrary, the size of the strata is defined proportionally to the standard deviation of the variables under study. In other words, larger strata are taken in strata with greater internal variability to better represent the most difficult population groups in the total sample… ”(Jan, 2009)

SYSTEMATIC SAMPLING

It is a technique which consists of selecting a certain subject randomly among the population and subsequently selecting for the sample every nth individual available in the sampling frame.

Systematic sampling is a very simple process, requiring only the random selection of an individual. The results we obtain are representative of the population, this type of sample is rarely used.

The process for systematic sampling is presented below.

SYSTEMATIC SAMPLING PROCESS

"We made an ordered list of the N individuals in the population, which would be the sampling frame. We divided the sampling frame into n fragments, where n is the sample size we want. The size of these fragments will be K = N / n

where K is called the interval or elevation coefficient.

Start number: we obtain an integer random number A, less than or equal to the interval. This number will correspond to the first subject that we will select for the sample within the first fragment in which we have divided the population. Selection of the remaining n-1 individuals: We selected the following individuals from the randomly selected individual, by means of an arithmetic succession, selecting the individuals from the rest of the fragments in which we have divided the sample that occupy the same position as the initial subject. This is equivalent to saying that we will select individuals… ”(Jan, 2009)

A, A + K, A + 2K, A + 3K,…., A + (n-1) K

ADVANTAGES OF SYSTEMATIC SAMPLING

It represents the population very well, quickly and simply since there is no need to generate many random numbers as well as subjects in our sample. It guarantees us an equitable selection of the population to study.

SAMPLING BY CONGLOMERATES

Cluster sampling is a technique that is based on groups to be analyzed, which correctly represent the total population in relation to the characteristic that we want to measure.

CONGLOMERATE SAMPLING PROCESS

Define the clusters, selecting a characteristic that allows the population to be divided into separate groups and exhaustively. Select the clusters to be studied. Investigate the subjects that are part of the cluster.

It is cheaper to select a cluster to study than to make a random or systematic sample.

That the conglomerates are not really homogeneous among them.

NON-PROBABILISTIC SAMPLING

In many occasions, other sampling techniques have to be used, which are in the non-probabilistic group. In these alternative techniques, it is commonly resorted to selecting elements for the sample based on hypotheses related to the population of interest, which is called selection criteria.

A non-probability sample informs us of what a universe is like but does not allow us to know with what precision: we cannot establish margins of error and confidence levels.

CONVENIENCE SAMPLING

As its name implies, this technique is about selecting a sample of the population due to the fact that it is accessible. In other words, the individuals who will contribute to the research are chosen because they are readily available, not because they have been selected using statistical criteria, although it does not allow a general statement about the population studied.

ADVANTAGE OF CONVENIENCE SAMPLING

 “Because the element to be analyzed is willing to cooperate with the study, it is easier to obtain truthful data apart from its disposition to the cause.

DISADVANTAGE OF CONVENIENCE SAMPLING

 The lack of representativeness, the impossibility of making statistical statements about the results and the risk of incurring biases due to the sampling criteria used. In the worst case, my convenient sample may present a systematic bias with respect to the total population, which would produce distorted results… ”(Jan, 2009)

SEQUENTIAL SAMPLING

Sequential sampling is a non-probability sampling technique in which to select the subject or group to be studied, they are chosen in a certain time interval, the study is carried out, the results are analyzed, then another group of subjects is selected, if necessary.

SAMPLING BY QUOTAS

Quota sampling consists of three steps

Segmentation division of the population under study into groups exhaustively and mutually. Setting the size of the quotas. The number of individuals to survey for each of these groups. Selection of participants and verification of fees. Participants are sought to cover each of the defined fees.

ADVANTAGE OF QUOTA SAMPLING

 The results it shows are very useful, in addition to being inexpensive and reliable.

DISADVANTAGES OF QUOTA SAMPLING

"The impossibility of limiting the mistake we are making when using this type of sampling. The risk of missing a relevant quota in a study." (Jan, 2009)

It is a non-probability sampling technique in which the individuals selected to be studied equally recruit other participants for the study. It was given the name snowball precisely because like a falling snowball, more material gathers between the subjects in question.

Snowball is very often used in populations where there is no easy access. In studies that require studying a specific population, it may be more effective to obtain a sample through the members of the same population, than by means of a purely random selection, in which a large number of individuals candidates to participate would be discarded.

SNOWBALL SAMPLING PROCESS

“The process of creating a sample using a snowball is based on using the social network of some initial individuals to access a group. We could divide this process into the following steps:

Define a participation program, which describes the process by which an individual invites or refers others to participate. Identify groups or organizations that can facilitate access to initial individuals that meet the characteristic feature of the study. Obtain initial contacts and ask for their participation. This part would be similar to a conventional sampling technique, but aimed at obtaining a reduced sample size. Request access to other contacts after the interview is completed. Ensure diversity of contacts by correctly selecting individuals. initials and promoting that the recommendation is not limited to very close contacts. ” (Jan, 2009)

TYPES OF SAMPLING SNOWBALL

Basically we can identify two types of snowball sampling:

Linear sampling: the individuals in the population must recommend another individual, in such a way that the sample grows at a linear rate. Exponential sampling: Each individual must invite more than one individual to participate. As several people will be invited simultaneously, the population will grow more in less time.

ADVANTAGES OF SNOWBALL SAMPLING

It allows sampling of hard-to-reach populations. It is an economic and simple process. It requires little planning and few human resources: the interviewed subjects themselves do labor.

DISADVANTAGES OF SNOWBALL SAMPLING

Lack of control regarding the constitution of the sample. It does not guarantee representativeness, nor does it allow us to know the degree of precision that it will offer. Uncontrolled sample size: the technique does not allow to fix a priori with precision the sample size that we are going to obtain.

DISCRETIONAL SAMPLING

Discretionary sampling is more commonly known as intentional sampling. In this type of sampling, the subjects are chosen to be part of the sample with a specific objective. With discretionary sampling, the researcher believes that some subjects are more suitable for research than others. For this reason, those are deliberately chosen as subjects… ”(Universo Fórlecciones, 2014)

The researcher selects individuals according to their professional criteria. You can make the selection based on the experience of previous studies or your knowledge of the population and its behavior against the characteristics studied.

SAMPLING FRAME

The sampling frame is the list of elements that make up the universe that you want to study and from which the sample is drawn. These elements to investigate should not always be individuals, they can also be other things that can be analyzed. Each of these elements present in the sampling frame is known as sampling units.

CONCLUSION

Unimaginable the amount of benefits that the tools that I presented in this work can provide us, depends on the situation in which we find ourselves and the objectives and type of sampling that we must use.

It is known that when carrying out studies and investigations we must always use the tools that best adapt to our needs, however it is not always possible to obtain 100% reliable results, that is why we have a wide variety of instruments at our disposal, it is our task try to meet the standards proposed for each activity.

Although it is not possible to carry out a study that does not contain a margin of error, we can consciously use sampling and make the most of it, that is up to the researchers.

I conclude that in any area in which people work, probabilistic calculations will always be used, and this is about learning more and more to develop and grow, in addition to leaving our grain of sand as a contribution.