Among the main research models the experimental and the non-experimental stand out. Each one has different categorizations, which makes it possible to adapt the work modality to the phenomena to be studied. There are also other classifications proposed by various scholars that offer other approaches to the problem..
For example, the Argentine philosopher and scientist Mario Augusto Bunge has proposed the distinction between basic research and applied research. The criterion of these categories is based on the function of scientific work: if it seeks to generate a merely descriptive theoretical knowledge, or if it has the practical application of knowledge.
Research models can also be classified according to the methods used to obtain the required data. The methods can be that of documentary or bibliographic research, field research or experimental research.
Likewise, the level of scope of the knowledge that is intended is fundamental to determine the types and models of research. The first level investigative scope is exploratory, after this there is the descriptive research model and finally, the explanatory one..
The first model we have is the non-experimental one. This model is also called ex − post − facto.
It consists of analyzing the phenomena as they occur in their natural context. Only events that have occurred previously, not generated in an experimental process, are taken into account.
An ex-post-facto investigation is one in which the scientist first performs a test on the object of study to measure the dependent variable, but does not manipulate the independent variable. In this method, the study subjects cannot be chosen randomly either..
It is an empirical-analytical method of a quantitative nature and is ideal for establishing cause and effect relationships. It is generally used in the social sciences because it enables the establishment of events that affect the subjects and study groups, allowing the possible causes of such events to be analyzed..
Among the ex-post-facto studies are those that respond to a transectional or cross-sectional design. In this design, the data is collected in a single moment to analyze the state and relationships of the variables at a given moment..
Among the investigations of a transectional type, there may be exploratory ones, which only focus on identifying the variables of a system..
We also have the descriptive type. These can be carried out through surveys or observational methods that allow the relevant variables to be measured..
Another type of cross-sectional study is the comparative-causal one. In this, the relationship between an independent variable and another dependent is measured, to determine if the variation of the latter is the effect of changes in the former..
Another non-experimental method of cross-sectional design is that of correlations, which consists of measuring the degree of non-causal association between two variables. These include predictive studies, factor analysis and structural equation modeling, as well as study using relational models..
The longitudinal research design seeks to make a diachronic study. Analyze the changes in variables over time trying to establish their causes and consequences.
Among these, trend studies stand out, which seek to analyze the temporary changes that a population goes through.
Another longitudinal design study is the evolutionary type, which is considered smaller than the trend study. It takes as study subjects within a population exclusively those who make up a certain generational subgroup.
There are also longitudinal panel-type studies. In these they take into account the temporal variations of individuals belonging to different generational subgroups of a population..
In these investigations there is a low degree of control. Since it is intended to study the subjects in their natural conditions, the process is at the mercy of the contextual conditions and the unforeseen events that these may cause..
These studies are not conducted in artificial, modified, or controlled environments. All data must be obtained from the observation of real situations.
For this reason, the weakness of these studies lies in the safety of the conclusions, since external agents may intervene in the specified variables and generate possible inaccuracies when establishing causality factors..
Non-experimental investigations are passive in nature. The object of study is not altered by the scientist; This only serves as a scorer, someone who measures the various variables after they have interacted in a previous event..
The researcher must select some observable effects. The effort will consist of doing a retrospective analysis to look for possible causes, establish relationships and reach conclusions. For this, ex-post-facto studies make use mainly of statistical techniques..
It is a valuable type of research when experiments are not possible for technical or ethical reasons. It allows to apprehend the nature of a real problem determining the factors that can be related in certain circumstances and times.
- A non-experimental study can be to observe the common traits of the students who obtained the highest grades in a school period.
- In the field of medicine, a non-experimental study may consist of pointing out the characteristics of cancer in its emergence and reemergence processes to determine possible relationships with its context of appearance..
- An ex-post-facto socio-political investigation could establish the relationships between the results of an election, the characteristics of the political campaign carried out and the country context in which the electoral process took place..
In this model the researcher manipulates the independent variables. This gives you a greater degree of control, making conclusions about the causality of the results more reliable..
There are two forms of pre-experimental design. The first is the one that only contemplates the study of a group by means of a single measurement; a stimulus is administered to the subjects to then measure the reaction of their variables.
The second type of pre-experimental study is the one that contemplates the use of pre-test and post-test. In this way, there is a reference point for the variable prior to the application of the stimulus: the evolution of the study subjects is monitored..
This type of experimental design has a minimum degree of control since it precisely dispenses with the control group.
There are no comparison groups; this makes the pre-experimental design prone to sources of internal invalidation. It is used especially in exploratory and descriptive studies.
This model became popular in the educational field since classroom resources do not allow certain conventional experiments to be carried out. They are typical of the field of applied sciences and usually serve to determine social variables.
In quasi-experimental models, the researcher manipulates at least one of the independent variables to see its effect on the dependent ones. In this way the causal relationships can be determined.
The study groups are not randomly selected, but are previously determined by conditions outside the experiment.
It can be a group of students in a class or a group of laborers working on a construction site. This means that the groups may not be homogeneous in terms of the variables studied, thus affecting the internal validity of the research..
In addition, they develop in natural situations, in which the environment is not controlled by the researcher. This also makes them very inexpensive and easy to apply..
A risk of this type of experiment is the placebo effect, since the subjects can change their behavior when they know they are participating in an investigation.
The true experimental research model is considered the most accurate of all scientific methods. The hypotheses are tested mathematically.
It is common in the physical sciences, but presents difficulties of application for the social sciences and psychology.
In a true experimental investigation, sample groups should be assigned randomly, in addition to having a control group.
Furthermore, any true experimental design can be statistically analyzed; for this reason, your results are always reliable and categorical, they do not leave room for ambiguity.
A fundamental element is that they must have absolute control over all variables. Since the experimental situation is totally controlled by the scientist, it is easily replicable, allowing the results to be corroborated by performing several identical tests..
An investigation with a true experimental model may be to test the effectiveness of various types of pesticides on mouse embryos..
The independent variable would be the pesticide, so a control group that is not exposed to the pesticide and other groups selected at random should be used so that each one receives a specific pesticide.
The dependent variable would be the degree to which embryonic development is affected according to the type of pesticide to which it has been exposed..
Likewise, a quasi-experimental research can take place in the field of social development: for example, a program to correct behavior and prevent crime applied to a group of young people in a community.
The study group would not be randomly selected, it is previously determined by being from the community to which the experimental program is intended..
Yet No Comments