Non-Experimental Research: Designs, Characteristics, Types and Examples

The non-experimental research is one in which the variables of the study are not controlled or manipulated. To develop the research, the authors observe the phenomena to be studied in their natural environment, obtaining the data directly to analyze them later.

The difference between non-experimental and experimental research is that variables are manipulated in the latter and the study is carried out in controlled environments. So, for example, you experience gravity by intentionally dropping a stone from several heights.

On the other hand, in non-experimental research, researchers go, if necessary, to the place where the phenomenon to be studied happens. For example, to know the drinking habits of young people, surveys are conducted or observed directly as they do, but no drink is offered.

This type of research is very common in fields such as psychology, the measurement of unemployment rates, consumption studies or opinion polls. In general, these are pre-existing facts, developed under their own laws or internal rules

Non-experimental research designs

As opposed to what happens with experimental research, in the non-experimental the variables studied are not deliberately manipulated. The way to proceed is to observe the phenomena to be analyzed as they are presented in their natural context.

In this way, there are no stimuli or conditions for the subjects that are being studied. These are found in their natural environment, without being transferred to any laboratory or controlled environment.

The existing variables are of two different types. The first are the independent calls, while the so-called dependents are a direct consequence of the previous ones. In this type of research, the relationships between causes and effects are investigated in order to draw valid conclusions.

Given that no exprofeso situations are created to investigate them, it can be affirmed that the non-experimental designs study the already existing situations developed under their own internal rules. In fact, another denomination that is given is that of investigations ex post facto ; that is, about facts fulfilled.

Differences with experimental designs

The main difference between both types of research is that in the experimental designs there is a manipulation of the variables by the researcher. Once the desired conditions have been created, the studies measure the effects of them.

For its part, in non-experimental investigations this manipulation does not exist, but rather the data is collected directly in the environment in which the events take place.

It can not be said that one method is better than the other. Each one is equally valid depending on what is going to be studied and / or on the perspective that the researcher wants to give to his work.

By its own characteristics, if the research is experimental it will be much easier to repeat it to ensure the results. However, the control of the environment makes some variables that may appear spontaneously more difficult to measure. It is just the opposite of what happens with non-experimental designs.

characteristics

As previously mentioned, the first characteristic of this type of research is that there is no manipulation of the variables studied.

Normally, these are phenomena that have already occurred and are analyzed a posteriori. Apart from this characteristic, other peculiarities present in these designs can be pointed out:

- Non-experimental research is widely used when, for ethical reasons (such as giving drink to young people), there is no option to conduct controlled experiments.

- No groups are formed to study them, but these are already pre-existing in their natural environments.

-The data is collected directly, and then analyzed and interpreted. There is no direct intervention on the phenomenon.

- It is very common that non-experimental designs are used in applied research, since they study the facts as they occur naturally.

- Given the characteristics presented, this type of research is not valid to establish unequivocal causal relationships.

Types

Transverse or transectional design

This type of non-experimental research design is used to observe and record the data at a specific time and, by its very nature, unique. In this way, the analysis is focused on the effects of a phenomenon that occurs at a particular time.

As an example, we can mention the study of the consequences of an earthquake on the housing in a city or the school failure rates in a given year. You can also take more than one variable, turning the study into a more complex one.

The transversal design allows to cover diverse groups of individuals, objects or phenomena. At the time of developing them, they can be divided into two different groups:

Descriptive

The objective is to investigate those incidents and their values, in which one or more variables appear. Once the data is obtained, a description of them is simply made.

Causes

In these designs, we try to establish the relationships between several variables that have occurred at a given moment. These variables are not described one by one, but rather they try to explain how they are related.

Longitudinal design

Contrary to what happens with the previous design, in the longitudinal the researchers intend to analyze the changes suffered by certain variables over time. You can also investigate how the relationships between these variables evolve during this period.

To achieve this goal it is necessary to collect data at different time points. There are three types within this design:

Trend

They study the changes that happen in some population in general.

Of group evolution

The subjects studied are smaller groups or subgroups.

Panel

Similar to the previous ones but with specific groups that are measured at all times. These investigations are useful to analyze the individual changes together with the group, allowing to know what element has produced the changes in question.

Examples

In general terms, these designs are prepared for the study of events that have already happened and, therefore, it is impossible to control the variables. They are very frequent in statistical fields of all kinds, both to measure the incidence of some factors and for opinion studies.

Effects of alcohol

A classic example of non-experimental research is studies on the effects of alcohol on the human body. Since it is not ethical to give drink to the subjects studied, these designs are used to obtain results.

The way to achieve this would be to go to the places where alcohol is habitually consumed. There is measured the degree that this substance reaches in blood (or you can take data from the police or a hospital). With this information, we will proceed to compare the different individual reactions, extracting the conclusions about it.

Opinion polls

Any survey that tries to measure the opinion of a certain group on a topic is done through non-experimental designs. For example, electoral polls are very common in most countries.

School performance

It would only be necessary to collect the statistics of the results of the school children offered by the schools themselves. If, in addition, you want to complete the study, you can search for information on the socioeconomic status of the students.

Analyzing each data and relating them to each other, a study is obtained about how the socioeconomic level of families affects the performance of school children.

References

1. APA rules. Non-experimental research - What they are and how to elaborate them. Retrieved from normasapa.net
2. EcuREd. Non-experimental research. Retrieved from ecured.cu
3. Methodology2020. Experimental and non-experimental research. Retrieved from metodologia2020.wikispaces.com
4. Rajeev H. Dehejia, Sadek Wahba. Propensity Score-Matching Methods for Nonexperimental Causal Studies. Retrieved from business.baylor.edu
5. ReadingCraze.com. Research Design: Experimental and Nonexperimental Research. Retrieved from readingcraze.com
6. Reio, Thomas G. Nonexperimental research: strengths, weaknesses and issues of precision. Retrieved from emeraldinsight.com
   
7. Wikipedia. Research design. Retrieved from en.wikipedia.org