Definition
The extent to which an experimental design is able to estimate main and interaction effects independently (free of aliasing) of one another. Design resolution is an issue that comes up with fractionated designs – usually the smaller the fraction, the lower (worse) the resolution. Full Factorial Designs do not have this problem because they include all possible treatment combinations, making them full resolution designs.
Low resolution designs tend to be economical in terms of the number of trials, but compensate for this economy by aliasing (combining) two or more lower-order effects in the design. As a result, it is nearly impossible to differentiate one aliased effect from the other without any extra information. In general, the focus is on resolution III, IV and V designs. The interpretation is as follows:
Resolution III Design: Main effects are aliased with two-factor interactions and two-factor Interactions are aliased with each other.
Resolution IV Design: Main effects are NOT aliased with two-factor Interactions, but two-factor Interactions are aliased with each other.
Resolution V Design: No Main Effects and two-factor Interactions are aliased with each other but two-factor Interactions are aliased with higher order Interactions.
Thus, Resolution III designs have the worst (lowest) resolution by aliasing Main Effects with two-factor Interactions. As these lower order effects are considered most important to help characterize the relationships among the Factors (under the Sparsity of Effects Principle), resolution III designs are generally not advised except in the case of a screening objective, where a large number of potential Factors needs to be pared down to the significant few.
External Links
Save Time With Fractional Factorials from iSixSigma: - http://www.isixsigma.com/library/content/c020429a.asp