Sigmapedia - The Free Online Lean Six Sigma Encyclopedia

English |  Español |  Français |  Português |  Deutsch |  中文

Blocking

Go Back

Definition

Blocking is a method of segregating and controlling for known sources of variation, such as people, work shifts, raw materials, or equipment, that take place over the course of an experiment. These sources are not of primary interest to the study ('nuisance factors') but contribute variation to the process in the real world. If this variation is not accounted for, it could mask the effect of the true variation due to the primary variables of interest, the treatments. By blocking, the effects of this variation can be isolated and quantified. The general rule of thumb for designed experiments is to "block what you can and randomize what you cannot".

Blocking results in arranging the experimental material in such a way that it is relatively homogeneous within blocks and heterogeneous between blocks.

Examples

Blocking results in arranging the experimental material in such a way that it is relatively homogeneous within blocks and heterogeneous between blocks. One example of a blocked design is the Randomized Complete Block Design (RCBD), in which each block receives all the treatment levels. When resources are limited, however, and all treatment comparisons are equally important, the Balanced Incomplete Block Design (BIBD) is another option.

Application

The table on the right shows an RCBD testing the effect of varying amounts of Additive (treatments) on the drying time of different Batches of paint (blocks):

External Links

Randomized Block Designs: - http://www.itl.nist.gov/div898/handbook/pri/section3/pri332.htm