In the book Encyclopedia of Behavioral Medicine, Turner describes experimental design from a clinical perspective “as an experiment with a series of observations made under conditions in which the research scientist controls the influences of interest”, Turner, 2013. The randomized trial is a classic example of experimental design. The data is randomized to one of two or more experimental groups. Then the significant differences between the sets are analyzed. (Turner, 2013) The Completely Randomized Design, Randomized Block Design, and Factorial design are three experimental designs, and analysis of variance (ANOVA) will tell us whether or not there is a difference between the means of one or more independent categorical groups. The ANOVA is an easily calculated method we can use in R.
Thus, the importance of experimental design has significance because tested hypotheses are relatively quick to determine how to move forward with various datasets, and the design gives a good starting point for the confidence of the data in moving to other analysis functions and depth and we can:
- Maximize insight into a data set
- Uncover underlying structure
- Extract important variables
- Detect outliers and anomalies
- Test assumptions
References
Nist Sematech. (n.d.). One-Way ANOVA. Retrieved from http://www.itl.nist.gov/div898/handbook/ppc/section2/ppc231.htm
Turner, R. J. (2013). Encyclopedia of Behavioral Medicine - Experimental Designs. https://doi.org/10.1007/978-1-4419-1005-9
Yau, C. (2013). R Tutorial with Bayesian Statistics Using OpenBUGS. Retrieved from: http://www.r-tutor.com/content/r-tutorial-ebook.