Randomized Block Designs

• Published in: Design and Analysis of Experiments with SAS pp. 129 - 160
• Main Author: Lawson, John
• Format: Book Chapter
• Language: English
• Published: United Kingdom Chapman and Hall/CRC 2010; CRC Press LLC
• Subjects: BIOLOGY, LIFE SCIENCES; MATHEMATICS; Probability & statistics; Source DF Type; Tread Design; Experimental Units; Randomly Assigned; Blocking Factor; Lever Press Rate; Files Flowers; Orthogonal Main Effect Plans; Traditional ANOVA; Golf Experiment; ANOVA Table; Latin Square Design; Sub-plot Factors; Source DF Square; Latin Square; SAS Command; Proc Glm Data; SAS Proc; Source DF; Randomized Block Design; SAS Code; CRD; Treatment Factor; Block Design; RCB Design
• ISBN: 1420060600; 9781420060607
• DOI: 10.1201/9781439882740-8

In order to eliminate as much of the natural variation as possible and increase the sensitivity of experiments, it would be advisable to choose the experimental units for a study to be as homogeneous as possible. In mathematical terms this would reduce the variance, σ2, of the experimental error and increase the power for detecting treatment factor effects. On the other hand, most experimenters would like the conclusions of their work to have wide applicability. Consider the following example. An experimenter would like to compare several methods of aerobic exercise to see how they affect the stress and anxiety level of experimental subjects. Since there is wide variability in stress and anxiety levels in the general population, as measured by standardized test scores, it would be difficult to see any difference among various methods of exercise unless the subjects recruited to the study were a homogeneous group each similar in their level of stress. However, the experimenter would like to make general conclusions from his study to people of all stress levels in the general population.