Introduction to Types of Experimental Research Designs.
In this article we will go through the topic types of Experimental Research Designs.
Meaning of Experimental Research Designs
Experimental research designs entails the deliberate manipulation of one or more independent variables to observe their impact on a dependent variable, all while controlling for extraneous factors. This method allows researchers to systematically investigate cause-and-effect relationships and test hypotheses within controlled settings.
By carefully controlling variables and randomizing participant assignment, researchers can minimize bias and draw reliable conclusions about the effects of the manipulated variables. Experimental research designs is pivotal across various disciplines, enabling advancements in understanding and application in fields such as psychology, medicine, and education.
Types of Experimental Research Designs
(A) Informal Experimental Research Designs
(i) ‘After-only’ experimental design.
(ii) ‘Before- and -after’ design.
(iii) Ex-post Experiment.
(B) Formal Experimental Research Designs
(i) Completely randomized design.
(ii) Randomized block design.
(iii) Latin square design.
(iv) Factorial design.
Types of Experimental Research Designs
(A) Informal Experimental Design
(i) ‘After only’ experimental design
In this experiment two groups or two areas of homogeneous character are selected. No Change is introduced in one of these groups which is called Controlled Group. The second group is treated as Experimental Group in which certain changes are affected which are not measured before treatment. After a certain period, progress or position of both the groups is studied and the variable is measured at the same time. The impact of treatment is assessed by comparing variables of both the groups.
This may be diagrammed as follows
| Control area: Level of Phenomenon without treatment (O) Test area : Treatment-Level of Phenomenon after treatment introduced (X) Treatment effect X-O. |
(ii) ‘Before-and after’ design
Broadly there are six types of ‘Before-and- After’ experimental designs. These are as follows
(a) Before-and-after’ study of one experimental group (without control group)
Only one group is selected for this experiment. There is no controlled group in such an endeavor. The group so selected is studied with the help of some fact collection method in order to know the present status with respect of a particular issue or problem (such as health and sanitation). After this, effort is made to bring change in the status by introducing change agents such as creation of awareness among the group and providing incentives to the group people.
After a certain period of time, the same group is studied again by applying the same method of collection of information and data as was previously used. If change is found it is assumed that it is due to the impact of change measures. Such an experiment is conducted on special occasions only because this method is not treated as very apt and fully dependable.
This may be diagrammed as follows
| Level of Phenomenon before treatment Treatment effect X-O Level of Phenomenon after treatment X |
(b) ‘Before-After’ study of controlled and Experimental Groups
This is taken to be the most appropriate and easy method of experimental research. In such a design two groups namely controlled groups, and experimental groups are taken. In both the groups pre-experiment and post-experiment variables are measured and the impact of change is ascertained and studied. If the level of change is found less in the controlled group (as compared to the experimental group) it is taken as the impact of change measures introduced in the experimental group.
This may be diagrammed as
| Pre-experiment Post-experiment Control area: Level of Phenomenon Level of Phenomenon O1 O2 Treatment area: Level of Phenomenon Level of Phenomenon X1 X 2 Treatment effect = [ (X 1 – X 2) – (O1 -O2)]( |
(c) ‘Before-after’ Study by Inter-changing the groups
In the study under this design, both the groups- experimental as well as controlled will be brought under observation and study. There is no study of ‘before’ in the case of experimental group whereas in case of controlled group ‘before’ is studied and measured. New variables and factors are administered in the experimental group and their impact is measured after a certain period of time. The difference in measurement of status prevailing before’ in case of controlled group and after’ in case of experimental group is taken to be the impact of experimental factors, variables and causes.
The special feature of this design is that in case of controlled group ‘after’ is measured. The respondents of each group are interviewed only once so that there is no chance of any replication or duplication of information and the difference found in both the measures is rightly taken as the impact of experimental variables.
(d) ‘Before-After’ Study of two controlled groups
It is a mixed form of above two designs where again both-controlled as well as experimental groups are measured only once. Sometimes, ‘before measurement may impact on experimental variables so it becomes difficult to measure. the net impact of experimental variables the impact of ‘before’ measure of experimental variable is called mutual or inter-related action. Under this design both impact as well as inter-related action of experimental factors can be measured.
In this design there are two controlled groups and one experimental group. All three groups are identical in nature. The ‘before’ and ‘after’ factors of experimental group and the first controlled group are measured. Thereafter, new factors and variables are introduced in the experimental group and the first controlled group is left as it was.
The second controlled group is also impacted by variables but its ‘before’ factors are not measured. The ‘before’ measure of the second controlled group is estimated on the basis of first group and experimental group. In this process, the two controlled groups are not influenced by the interaction among themselves. However, there is probability of interaction within the experimental groups. The ‘before’ and ‘after’ measure makes the interaction process probable.
(e) ‘Before-after’ study of three controlled groups
It has been presumed in the above design (d) that the impact of contemporary events, changes and developments is almost negligible. But such an assumption is not logical. therefore, scientists have formulated this fifth design to overcome that problem and shortcoming. In this design three controlled groups are selected. The third new controlled group is neither impacted by any experimental variable nor its ‘before’ measure is taken. Only ‘after’ of this third group is measured.
Thus, there is no impact of ‘before’ measurement and experimental variables on this (third) group. However, this third group is impacted by developmental and contemporary variables and factors. All other features of this design are the same as are there in other previously discussed designs. Net impact of experimental variables can be measured by applying the method under this design.
(f) Study of Joint-Impact of two or more Experimental Variables
Sometime it becomes necessary to study the impact of two or more variables when they become operational at the same time. For conducting the study of their common impact, the researcher has to make a complex design which involves several controlled and experimental groups. These groups are formed in accordance with the impact of variables on the basis of priority by applying positive and negative rules.
Read Also : Principles of Experimental Research Designs
Types of Experimental Research Designs
(B) Formal Experimental Designs
Before we consider the formed research designs, we will define some commonly used terms:
Factors: Factor widely used to denote an independent variable. A factor may have two or more levels such as
(i) Defective and non-defective
(ii) Large, medium, small
Factors can also be classified by whether the experimenter can manipulate the levels associated with the participants.
(i) Active factor
(ii) Blocking factors.
Active Factors
Those independent variables which the researcher can manipulate by causing a participant to receive one treatment level or another.
Blocking Factors
Those factors which the researcher or experimenter can only identify and classify the participant on an existing level. Gender, age group, customer status, ethnicity are examples of blocking factors.
Types of Experimental Research Designs
(i) Completely Randomized Design (C.R. Design)
It is regarded to be the simplest possible design and is quite easy in application and analysis. It involves two principles viz., the principle of replication and the principle of randomization. Such a design can be applied in homogenous areas only. Under this design, the subjects are randomly selected for treatment and study.
There are two forms of this design which are discussed below
1.Two group simple randomized design
2. Random Replication Design
1. Two Group Simple Randomized Design
In a two-group simple randomized design, participants are randomly assigned to one of two experimental conditions or groups. This random assignment helps to ensure that each participant has an equal chance of being placed in either group, minimizing potential biases and confounding variables. By randomly assigning participants, researchers can confidently attribute any differences observed between the groups to the manipulation of the independent variable, rather than to pre-existing differences between participants.
This design is straightforward and commonly used in experimental research settings, allowing researchers to establish causal relationships between the independent variable and the dependent variable with greater confidence.
2. Random Replication Design
A random replication design involves randomly assigning participants to multiple experimental conditions or treatment groups, with each participant having an equal chance of being assigned to any condition. This design allows researchers to test the effects of different levels or variations of the independent variable simultaneously, providing more comprehensive insights into its impact on the dependent variable.
By randomly replicating the experimental conditions, researchers can enhance the generalizability and reliability of their findings, as they can assess whether the observed effects hold true across various contexts or conditions. Random replication designs are particularly useful when investigating complex phenomena that may involve multiple factors or treatment levels.
(ii) Randomized Block Design (R.B. Design)
It is an improvement over the random replication design. In this design, the subjects are divided into groups known as blocks. They are homogenous within themselves so far as certain variables are concerned, the number of subjects in a particular block will be equal to the number of treatments. From one block one subject will be randomly selected for treatment. The effect of treatment on the sample of each block will be studied and generalization will be made accordingly.
(iii) Latin Square Design
This design is used very frequently in agricultural research. Nature plays an important part in agriculture So studies concerning agriculture are different from other studies. The subject of research is treated with new variables but only once. This design helps in understanding and analyzing a situation where there are different variables of independent nature and character. For instance, we have to know the effect of three varieties of seeds. Such a study involves observation of the effects of other variables such as the type of fertilizers used, level of soil fertility etc. Such extraneous factors can be studied with the help of Latin square design.
(iv) Factorial designs
This design is useful to study economic and social situations where there are a large number of variables affecting a problem.
There are two types of factorial designs
(a) Simple Factorial Design
(b) Complex Factorial Design.
(a) Simple Factorial Design
It is also called ‘Two-factor-factorial design’. Under this design, the effects of two different factors on the dependent variables are studied and analyzed. Here the sample is divided into two parts. Each of these parts is given one single treatment. For treatment, the samples are selected randomly. In this design, the extraneous variable to be controlled is called control variable and the independent variable is called experimental variable.
There are two treatments of the experimental variable and two of the control variables. thus, there are four cells into which the sample is divided. The effects of treatment as well as levels can be studied with the help of this design. An added advantage of this design is that the interaction between treatment and levels can also be examined and studied.
(b) Complex Factorial Design
An experiment involving the use of two or more factors at the same time requires to be done through complex factorial design method. This is the design which can consider three or more independent variables simultaneously. Such designs are used mainly because they provide accurate results with minimum labour and cost. With the help of such a design, we can determine the effect of two or more factors in a single experiment. Further, such designs provide information about the effects which cannot be obtained by treating one single factor at a time.