SHARE THE ARTICLE ON
Experimental research design is a scientific approach that involves two sets of variables. In the research one variable is manipulated by the researcher to measure the changes it causes in the other variable.
Experimental research is best used when the researcher wants to establish a cause-effect relationship among the variables. The research can be conducted in a natural or laboratory setting depending on the nature of the experiment.
Once you understand what Experimental Research is, you will need to learn the system and process of Experimental Design to conduct good research. In this article, we will discuss designing Experimental Research.
See Voxco survey software in action with a Free demo.
The process involves determining a cause-and-effect relationship between two variables. The Experimental Design involves making decisions regarding the following components.
In a study about different types of fertilizers affecting the growth of saplings, the variables are ‘different fertilizers’ and ‘growth of saplings’.
Explanatory Variables: Different fertilizers used
This variable is under the control of the researcher. The researcher manipulates the explanatory variable to observe and measure the changes it causes in the response variable, in this case ‘growth of sapling’.
Response Variable: Growth of sapling
This variable is kept under observation. A researcher does not manipulate the response variable but simply observes the changes caused in it as a result of the explanatory variable, ‘different types of fertilizers’ applied to the samplings.
The choice of treatment in an Experimental Design depends on the Explanatory variable and its levels.
When there is one explanatory variable, the treatment depends on the level of the variables.
When there are multiple explanatory variables involved, the treatment will be a combination of the levels of variables.
The Experimental Unit is the component of the experimental design to which the treatment will be assigned.
For example, when the goal of a study is to see the effect of alcohol in the memory of Americans between the age of 19 and 22, the basic experimental units would be college students.
The Sample Size is the experimental unit in a research design. It is determined on the basis of the availability of resources, the precision of estimation, and other factors. For better statistical inference, larger sample size is the best option.
Replication involves repeating the treatment under the same condition in an experimental design. Replication prevents any error in the experiment. When treatment has been repeated any difference in the result from the previous experiment can be identified; this is a random error.
A sample size equals the number of times each treatment is repeated under the same condition.
Let’s assume you want to see the effect of temperature in an earthen pot. You prepare 3 ovens with temperatures ranging from – low, medium, and high – and randomly assign 4 earthen pots made out of the same mold to each temperature.
Objects or individuals are assigned in a random manner to an experimental group. Randomization is used in an experimental design to eliminate any potential bias by creating a homogenous treatment group.
There are two primary variations of randomization in experimental design, as mentioned in the previous point
Complete randomization: In this case, subjects or objects are randomly assigned to a group.
Randomized Block Design: The researcher divides the subjects into a homogenous group and then randomly assigns each to a treatment group.
The first step is to identify the research problem that is right for your purpose. The research problem can help you to formulate the hypothesis so that you can design and conduct your experiment to reach the right conclusion.
Identifying your research problem can make it easy to plan out the subsequent steps you need to take for the completion of the research. It can further help you to determine the factors and components that will affect the experimental design, such as – availability of resources, the significance of the research, safety measures, or ethical concerns, and other considerations.
After determining a researcher’s problem it is always a good idea to review literature related to the problem. This can help you uncover facts you may not have considered or even give you ideas for research objectives.
After identifying your research problem, Hypothesis is your next concern. A Hypothesis is not a question for the experimental design, but a theoretical or testable statement. Your hypothesis will help you determine any logical relationship among the variables.
Your hypothesis is an informed guess around which you will be conducting your experiment. The hypothesis will help you understand the reason behind why things happen.
With a research problem and a hypothesis, you now have to move forward and design your experiment. An experiment requires clear guidelines, like a blueprint. The design of the experiment explains the components it requires and also how these resources will be used to conduct the experiment.
You need to implement the experimental design you have created in the actual experiment to prove the validity of your hypothesis. Conducting the experiment involves the steps you use in designing the experiment.
The conclusion of the experiment involves the solution that you will provide at the end depending on the findings of the research. The conclusion will demonstrate whether or not your hypothesis was valid.
The findings of the research should answer your hypothesis as well as the research problem.
As mentioned above, you need to identify your research problem and translate it into your hypothesis to begin your experimental design. When you define your hypothesis you also introduce the two sets of variables.
The variables involved in the experiment are
Let’s understand the different variables by taking an example.
You are conducting a study to see the effect of listening to classical music before bedtime on sleep patterns. Your specific goal is to see how the amount of time spent listening to classical music will affect the hours of sleep.
In this case,
Independent variable: Hours spent listening to classical music
Dependent variable: Hours of sleep every night
Extraneous variable: Natural sleep pattern of every participant
Now you want to consider how to control the experiment;
Control: you measure the difference between sleep with listening to classical music and sleep without listening to classical music.
Treatment in an experimental design depends on the explanatory variable or the controlled independent variable. In experimental research how you manipulate the independent variable impacts the validity of the experiment to the extent that the result can be generalized.
Following this example, you can choose to determine the level of variables as follows:
Categorical variable: No listening to classical music, little time listening to classical music, and lot (time) listening to classical music.
Continuous variable: minutes spent listening to classical music every night
Assigning subjects to the experimental treatments is important for relevant and valid results.
You need to begin by defining your sample size. Sample size accounts for the number of participants in the experiment. The greater the sample size the more statistically accurate your experiment will be.
The next step is to randomly assign the subjects of our experiment to treatment groups. Each group is assigned to a different level of treatment. Let’s say we use a categorical variable for treatment, then each group will be subjected to – No listening to classical music, little time listening to classical music, and a lot (time) listening to classical music.
You can also choose to include a control group in the experimental design. This group receives no treatment, the researcher does not intervene, i.e., no manipulation or change in the variables. This group demonstrates what happens to the subject of the experiment without any intervention.
There are two ways you can go about assigning subjects to treatments:
Complete randomized design involves assigning subjects to treatment groups in a random manner.
Randomized block design involves grouping people based on shared characteristics and then randomly assigning them to treatments.
Between-subject design allows individuals to receive only one level of treatment in the experimental design.
Within-subject design implies that every participant or subject receives every level of treatment consecutively. The response of every treatment is observed and measured.
Explore all the survey question types possible on Voxco
Explore all the survey question types possible on Voxco
You want your results to be valid and accurate; you should be able to generalize them. The idea is to keep the result error-free or limit the error.
In the above example, you can measure the dependent variable in two ways:
A good experimental design will check out all the steps and components mentioned in the article. It will take into consideration all of the factors to generate data that is valid, relevant, and accurate. The result of a good Experimental Design will prove whether the hypothesis is valid or not.