There are two types of statistical studies: an experimental study, which will discuss in detail in the following, and an observational study, and discuss more in the next article. An experimental study is considered a randomized control trial, which means control over the exposure. It usually can demonstrate causal relationships within two variables.
There are generally three steps to perform a randomized control trial:
Choose participants satisfied with specific conditions for study and meet the requirement, which is no contact with exposure and outcome ( normally, we experiment to discover the relationship between exposure and outcome).
We randomized split participants into two groups which are treatment and control.
After some time to observe whether participants in two groups will have an outcome or not. Then record data and use data to analyze the research question.
A typical example of a randomized controlled trial is following: To determine how short wave UVA-blocking sunscreen affects the skin’s general health compared to a regular long wave UVA-blocking sunscreen, we randomly split 40 trial participants into two groups of 20 which are treatment and a control group. The experimental group wore the short wave UVA-blocking sunscreen daily, and the control group wore the long wave UVA-blocking sunscreen daily.
We initially evaluated all participants’ skin health before entering two groups. After one year, the general health of the skin was measured in both groups and statistically analyzed. In the control group, the participants’ skin health is improved by 60% by wearing long-wave UVA-blocking sunscreen. Compare to the treatment group, the participants’ skin health is improved by 75% by wearing short wave UVA-blocking sunscreen.
From the above example, we can use the above experiment to compare which group is more effective in improving general skin health. It controls external factors other than the factors in the experiment and randomly split participants into different treatments. Besides that, experiments have other advantages compared to observation studies. It can minimize selection bias through randomization. It also has statistical reliability as the p-value can be used to analyze whether the experiment’s factor is significant to the response variable. Besides, it is comparable as it allowed to compare the result of two groups.
In conclusion, if you are curious about the casual relationship between the response variable and independent variable, we can use a randomized experiment in this study. Using a randomized experiment, we can check whether the independent variable effectively predicts the response variable and uses different models to present more statistical analysis.
Reference:
Randomized controlled trial. (n.d.). Retrieved April 06, 2021, from https://himmelfarb.gwu.edu/tutorials/studydesign101/rcts.cfm
Advantages and disadvantages of randomised control study design: Deranged physiology. (n.d.). Retrieved April 06, 2021, from https://derangedphysiology.com/main/cicm-primary-exam/required-reading/research-methods-and-statistics/Chapter%202.0.2/advantages-and-disadvantages-randomised-control-study-design
