Blocking and Stratification
Student’s Name
Institution Affiliation
Date
Blocking and Stratification
Blocking
Blocking is concerned with arranging the experimental units into organized groups called blocks with similar traits. Typically, the blocking factor in statistical theory in experimental designs is a variability source that does not hold a fundamental interest in the experiment. Once all nuisance factors are controlled, blocking can then be used to eliminate or reduce the contribution to the experimental error caused by the annoyance elements. Blocking is typically used to eliminate the impact of a few of the most crucial annoying variables in an experiment. As a practical example of the technique, blocking is an applicable approach to moving to a safer position. In an experiment to establish a new drug’s effectiveness on male and female patients, two treatment levels: drug and placebo, could be administered to the patients in a double-blind test (Jensen, 2018). The patient’s sexual orientation becomes the blocking factor in this type of experiment and accounts for the treatment variability between either being a man or woman. This decision would significantly reduce the sources of variability, resulting in higher precision of the trial.
Stratification
Stratification is an ideal model of partitioning subjects and their subsequent outcomes using factors other than the processes involved. This technique can be used to achieve an equal allocation of smaller groups of participants to every experimental status. In general, stratification is concerned with classifying members of an experimental population into strata. Stratification has various advantages; for example, a stratified sample has the potential to provide higher precision than random samples, and the model requires a smaller sample size, thereby saving resources that could rather have been spent on large sample sizes (Montalban-Bravo, 2019). For example, if a sporting club intends to conduct a fitness study of its players where gender or age is anticipated to influence the results, participants in the study can be stratified into strata by the contradictory variable (Montalban-Bravo, 2019). However, this technique has the disadvantage of requiring knowledge of the variable that must be controlled.
References
Jensen, S. (2018). Experimental design matters for statistical analysis: how to handle
blocking. Pest Management Science, 74(3), 523–534.
Montalban-Bravo, G. (2018). Myelodysplastic syndromes: 2018 update on diagnosis, risk-
stratification and management. American Journal of Hematology, 93(1), 129–147.