Experimental design is the planning of an experiment so that the effects of selected inputs on measured responses can be estimated while controlling unwanted variation. An experimental factor is an input variable varied or observed in the experiment, and its selected settings are its factor levels. A design also specifies the objects or trials to which treatments are assigned and the measurements to be analyzed.
Randomization assigns treatments by chance and helps prevent systematic statistical bias. Replication applies the same treatment combination to more than one independently observed object or trial, allowing experimental variability to be estimated. Blocking groups similar objects or trials and compares treatments within each group, reducing variation due to known nuisance experimental factors.
A factorial experiment includes every selected combination of factor levels and can be used to estimate both main effects and interactions. The resulting data are often analyzed using ANOVA.
