In this work, the sampling techniques that are divided into two main categories will be explained: probabilistic and non-probabilistic sampling, which is of great help to successfully carry out or adequately select a sample to carry out an investigation, said investigation can be for a very large studio as well as for a small one.
To carry out an investigation in a population, such as examining a population to determine the prevalence or absence of a disease, we can analyze the entire population in question, that is, carry out a census or carry out the investigation at a only part of the population by sampling.
By examining the entire population, it is possible to know exactly the distribution of the variable studied in the population. Although this would take more time or the costs would be very high since a large number of elements of the population need to be examined in the census, requiring time and money. If you want to study a large number of individuals, it is necessary to have trained personnel and well-equipped facilities.
The best alternative is to use only a part of the population, that is, to carry out a sampling, since it is much faster, cheaper and the results can be more accurate if the sample is chosen correctly.
The purpose of re-carrying out this work is to make known the two types of sampling, as well as the types of techniques found in each probabilistic and non-probabilistic sampling, as well as the advantages and disadvantages of choosing each sampling.
PROBABILISTIC AND NON-PROBABILISTIC SAMPLING
To begin and understand the topic of probability and non-probability sampling, it is important to know the meaning of sampling.
WHAT IS SAMPLING?
Sampling is the process of selecting a set of individuals from a population in order to study them and to be able to characterize the total population. (Ochoa, 2015). That is, it consists of choosing a certain group of individuals who are considered representative in order to facilitate the study or to determine the characteristics of the population.
For example, suppose we want to know what percentage of the inhabitants of a city consume alcohol in their daily lives. The one way to know this data is to study the entire population but this would be very difficult to do since the number of inhabitants in a city is very high, for this reason a sample (for example 800 people) of the population is chosen in order to know the desired percentage. The way the population was selected is called sampling.
CLASSIFICATION OF SAMPLING TECHNIQUES
According to Malhotra (2008), he classifies sampling techniques as follows: In non-probabilistic and probabilistic sampling. (See in Illustration 1)
PROBABILISTIC SAMPLING
Probability sampling is a sampling technique in which individuals from the population are chosen randomly and each one has the same positive probability of being chosen and being part of the sample. Therefore, it is a type of sampling that is more recommended for research, because it is more efficient, accurate and ensures the representativeness of the sample extracted, in addition the researcher must guarantee that each individual in the population has the same opportunities. In Probability Sampling the following types can be found:
- Simple Random Sampling Systematic Random Sampling Stratified Random Sampling Cluster Random Sampling Mixed / Multistage Random Sampling.
SIMPLE RANDOM SAMPLING
In this sampling technique, all the elements of the population that will be chosen for the sample are selected by random tide, in the same way, each element has the same probability of being selected, the elements are assigned a number that will be unique for their identification.
The elements that will belong to the determined sample will be selected independently of any other element. That is, if an element is selected, it will not have any influence on selecting another element or not.
To carry out the selection procedure, the number must first be assigned to all the elements of the population, then the selection will proceed through some mechanical means (strips of paper in a crystal ball, table of random numbers, pinballs). pong with a number assigned inside a bag, random numbers generated by a computer, etc.) depending on the size of the established sample, the elements associated with the numbers obtained by mechanical means will be the study sample.
ADVANTAGES OF SIMPLE RANDOM SAMPLING
- Great ease of assembling the study sample The elements of the population are selected in an equitable way, since all elements have the same probability of selection Each element selection is independent of other selections The statistical procedures required to analyze the data and calculate the errors are easier than those required in other probability sampling procedures. tends to produce representative samples Quick calculation of means and variances.
DISADVANTAGES OF SIMPLE RANDOM SAMPLING
- The need to always have a list of all the elements of the population, in case this method is applied to a population that is very large, it would be very time consuming to make this list, as well as keeping it updated at all times. applies to a large population.
SYSTEMATIC RANDOM SAMPLING
This procedure, like simple random sampling, lists all the elements of the population, but unlike the previous method, only one element of the population will be extracted, which is done at random.
The starting point will be the element? and the elements that make up the sample are those that occupy the places?,? +?,? + 2 ?,? + 3 ?,…,? + (? - 1) ?, that is, the individuals of? where ? is the result obtained by dividing the population size by the sample size. The number ? What is used as a starting point is a random number between 1? ?
- It is recommended when the population is very large It amplifies the sample to the entire population It does not present algebraic calculation problems This type of sampling requires less time as well as it is low cost
- It presents cases in which there are periodicities in the population since when choosing the members of the sample with a constant periodicity we can introduce a homogeneity that does not occur in the population The possibility of increasing the variance if there is periodicity in the population There is no Independence of the elements in different zones, since the elements selected in each zone depend on the one selected in the first zone. There is only one random selection that is for the first element in the sample.
STRATIFIED RANDOM SAMPLING
In stratified random sampling, the elements of the population are divided into subgroups or strata, in which each element can only belong to a single stratum. That is to say, each element will be grouped in stages respecting a characteristic, for example, sex, age, profession, marital status, etc.
The purpose of this type of sampling is to ensure that each stratum of interest is appropriately represented in each sample. The strata have to be mutually exclusive and collectively exhaustive, therefore each element of the population has to belong to a single stratum without omitting any element of the population.
The distribution of the sample according to the different strata is called allocation, the types of allocation can be:
- Simple assignment: Each stratum corresponds to the same number of sample elements. Proportional allocation: The distribution is made according to the size of the population of each stratum. Optimal allocation: the predictable dispersion of the results is taken into account so that the proportion and the standard deviation are taken into account.
After defining the strata, the sample is created, to obtain said sample, the individuals are selected using any sampling technique for each of the strata separately, generally the technique that is most applied to obtain the sample. of the strata is simple random sampling.
In the strata, there can be homogeneous groups of individuals and in turn they are heterogeneous among the different groups, that is, in each stratum group the individuals present characteristics in common but are at the same time different from the other strata since each stratum has just behavior. For example, in a certain study it is expected to find a different behavior between men and women, the first stratum will be men and the second will be women, therefore men behave in a similar way to each other and women behave in a different way that is similar to each other, therefore, men and women will have different behaviors from each other. If the strata are usually internally homogeneous and heterogeneous among themselves,When using this sampling technique, you will reduce the sampling error.
ADVANTAGES OF STRATIFIED RANDOM SAMPLING
- It is advisable to use this sampling technique when you have a priori knowledge of the population Obtain more accurate estimates The objective of this type of sampling is to ensure that each stratum of interest is appropriately represented in each sample Different sampling methods can be used for each Stratum If the strata tend to be internally homogeneous and heterogeneous among themselves, when using this sampling technique it will reduce the sampling error.
DISADVANTAGES OF STRATIFIED SAMPLING
- It is necessary to have a framework that has auxiliary information It is often difficult to form each of the strata.
RANDOM SAMPLING BY CONGLOMERATES
This method is used when the population is grouped in natural conglomerates, we understand by conglomerates the grouping of elements that present characteristics similar to the entire population, unlike stratified random sampling, this procedure is grouped according to the variables to be studied and are exactly identified to the population. Clusters have to be mutually exclusive and collectively exhaustive, as well as having an equal probability of being selected.
The procedure for using this method is as follows:
- Randomly select the clusters, the elements of each cluster will be part of the sample. These elements will form the groups or conglomerates. Subsequently, the random selection of the elements that will be studied within each group or the observation of all the elements that make up the chosen groups is made.
MIXED / MULTIPLE RANDOM SAMPLING
This sampling method makes a combination of two or more of the aforementioned sampling techniques, generally when conducting an investigation it is not highly recommended that only a probabilistic sampling method be used, since different stages are carried out during the investigation. research and if this method is used in each stage, a different random sampling technique would be applied.
NON-PROBABILISTIC SAMPLING
Non-probability sampling is used when it is difficult to obtain the sample by the probability sampling method. This method is a sampling technique that does not perform random selection procedures, but rather is based on the personal judgment of the researcher to select the elements that will belong to the sample. In this technique the probability of selecting each element of the population is not known and also not all have the same probabilities of being selected for the sample.
Although this method is not very representative under the criteria of the researcher, representativeness is not guaranteed. Among the non-probabilistic methods the most used are:
- Quota sampling Intentional or convenience sampling Snowball sampling Judgment sampling.
SAMPLING BY QUOTAS
It is a non-probability sampling technique, which consists of selecting the sample after the population is divided into strata. The difference between the quota sampling method and the stratified sampling method is that the selection of the elements of the population for the sample is made at the discretion of the researcher and is not carried out randomly as in stratified sampling. The procedure for making a choice from a sample of? elements of a population ?, is next
- The population is divided into? strata or groups (by age, sex, occupation, place of residence, etc.). If the strata have? 1,? 2,…,? ? elements, such that:? =? 1 +? 2 + ⋯ +? ? Subsequently, the researcher chooses the quotas (these quotas can be chosen at her own discretion or through the criteria adapted to the sample), that is, the number of elements,? 1,? 2,, …,? ? that will be taken in each stratum, where the sum will be the total of elements? of the sample ? =? 1 +? 2 + ⋯ +? ?.
- The elements that are chosen in each stratum are carried out using the probabilistic n method.
CRITERIA FOR CHOOSING QUOTAS
As previously mentioned, the quotas can be selected at the discretion of the researcher or it can also be selected by statistical criteria, some of these criteria are:
- Simple choice: The odds would be the same in the? Individuals would be selected in each stratum. It is not recommended to use this criterion when each stratum has a different number of individuals. Choice proportional to the size of the stratum: the quota in each group is proportional to the elements of that group. In the strata will be taken? ? Which is calculated with the following formula.
Where ? is the number of elements in the population,? the sample and? ? that of the stratum?
- Choice proportional to the variety of the stratum: if the variability of the characteristic that we are taking into account in the strata is known. The quotas are proportional to it in each group. If the variance is greater in the group, the quota will also be greater.
? =? ? ? ? ?
? ∑ ? ? ?
? = 1? ?
Where n is the number of elements in the sample,? ? that of the stratum and? ? the standard deviation of the stratum?
INTENTIONAL OR CONVENIENCE SAMPLING
It is a non-probability sampling method, this method consists of selecting the elements that are convenient for the investigation for the sample, said convenience occurs since the researcher finds it easier to examine the subjects either by geographic proximity. It is characterized by the effort to obtain samples that are representative by including typical groups in the sample.
This technique is one of the cheapest and the one that takes less time, since it does not take much effort or the use of some method to select the elements of the sample, the section is carried out at the discretion of the researcher.
SNOWBALL SAMPLING
It is called snowball sampling because it consists in that each subject studied proposes others, that is, some individuals are located and they lead others and in turn these lead to others until a sufficient sample is obtained.
Producing therefore a cumulative effect similar to a snowball. This technique is a non-probabilistic sampling method and is carried out in populations in which the individuals are not known or cannot be accessed, for example in sects, the indigent, minority groups, criminals, a certain group of patients, etc.
SAMPLING BY TRIAL
The non-probability sampling method, which consists of the subjects being selected based on the knowledge and judgment of the researcher. That is, the researcher uses his judgment or experience to select the elements that will belong to the sample, since he considers that they are more representative of the population under study.
This method is recommended to be used when the person in charge of conducting the study knows similar or identical previous studies and knows exactly that the sample was useful for the study, in the same way when the population is small, therefore the researcher knows the population.
CONCLUSION
Based on this information, it is concluded that the method that provides the greatest benefit is probabilistic sampling, since its results are usually more representative and non-probability sampling strives to obtain data that is effectively representative.
In non-probability sampling they must present their original characteristics, their results are useful if they are put to good use. In probability sampling methods, their theoretical approaches turn out to be very demanding.
These sampling methods are of great importance to carry out an investigation in a certain population since if you want to carry out this investigation with greater success in terms of results, then you have to choose one of these two techniques, in the same way these techniques require less time and resources.
Sampling is therefore a research tool whose function is to determine which part of a population should be examined, in order to draw conclusions about said population.
The sample must obtain an adequate representation of the population, in which the essential features of said population that are important for the research are represented in the best way. For a sample to be representative, and therefore useful, it must reflect the similarities and differences found in the population, that is, exemplify its characteristics.
GLOSSARY
E Sampling error
They are errors caused by the act of obtaining information from a sample, rather than the entire population. They are due to the variation between one sample and another
M Sample
It is the set of individuals in the universe that was selected to study them
P Population
It is the set of all the values of a phenomenon or property that you want to observe.
REFERENCES
- ARQHYS ARCHITECTURE. (sf). Retrieved on January 02, 2017, from ARQHYS ARQUITECTURA: Need to have a framework that has auxiliary information. Difficulty creating or forming the layers. BlogsPost. (sf). Retrieved on December 30, 2016, from BlogsPost: http://metodologia02.blogspot.mx/p/tipos-de-muestreo.html Statistics. (sf). Retrieved on December 29, 2016, from Statistics: http://www.estadistica.mat.uson.mx/Material/elmuestreo.pdf Explorable. (sf). Retrieved on December 30, 2016, from Explorable: https://explorable.com/es/muestreo-probabilistico Explorable. (sf). Retrieved on December 30, 2016, from Explorable: https://explorable.com/es/muestreo-aleatorioFormulas, U. (sf). Universe Formulas. Retrieved on January 02, 2017, from http: //www.universoformulas.com / statistics / inference / systematic sampling / Malhotra, NK (2008). Sampling; design and procedures. In NK Malhotra, Market Research (pp. 332-360). Mexico: Pearson Education, Prentice Hall. Maths. (sf). Retrieved on January 02, 2017, from Mathematics: http://matematicas.unex.es/~inmatorres/teaching/muestreo/assets/cap_5.pd fOchoa, C. (2015 de Open de 08). Net Quest. Retrieved on December 30, 2016, from Net Quest: http://www.netquest.com/blog/es/blog/es/muestreo-probabilistico-muestreoaleatorio-simplepd fOchoa, C. (Open 2015 of 08). Net Quest. Retrieved on December 30, 2016, from Net Quest: http://www.netquest.com/blog/es/blog/es/muestreo-probabilistico-muestreoaleatorio-simplepd fOchoa, C. (Open 2015 of 08). Net Quest. Retrieved on December 30, 2016, from Net Quest:
- Ochoa, C. (February 19, 2015). Net Quest. Retrieved on December 30, 2016, from Net Quest:
- Ochoa, C. (April 16, 2015). Net Quest. Retrieved on January 02, 2017, from http://www.netquest.com/blog/es/blog/es/muestreo-probabilisticomuestreo-estratificado ua.es. (sf). Retrieved on December 30, 2016, from Personal.ua.es:
- Ramos, LA (September 22, 2011). Slide Share. Retrieved on January 02, 2017, from Slide Share:
- Riesco, JM (nd). Basic Concepts of Statistics. Retrieved on January 02, 2017, from Basic Concepts of Statistics: http://www.jorgegalbiati.cl/ejercicios_4/ConceptosBasicos.pdf Universe Formulas. (sf). Retrieved on January 04, 2017, from Universo Formulas: http://www.universoformulas.com/estadistica/inferencia/muestreo-noprobabilistico/ Universo Formulas. (sf). Retrieved on January 02, 2017, from http://www.universoformulas.com/estadistica/inferencia/muestreo-cuotas/ Universo Fórmula. (sf). Retrieved on January 02, 2017, from https://www.netquest.com/blog/es/blog/es/muestreo-bola-nieve Universo Fórmula. (sf). Retrieved on January 02, 2017, from https://www.netquest.com/blog/es/blog/es/muestreo-por-conveniencia Universo Fórmula. (sf). Retrieved on January 02, 2017, from http: //www.universoformulas.com / statistics / inference / discretionary sampling /
