Replicates

Multiple experimental runs with the same factor settings (levels). Replicates are subject to the same sources of variability, independently of one another. You can replicate combinations of factor levels, groups of factor level combinations, or entire designs.

In experimental design, replicate measurements are taken from identical but different experimental runs. This is in contrast to repeats, which are simply repeated observations at the same settings. You can use replicates to estimate the variance (experimental error) caused by slightly different experimental conditions. The experimental error serves as a benchmark to determine whether observed differences in the data are statistically different. To make sure all the experimental variability is observed and quantified, replicates should be randomized to cover the entire range of experimental conditions. If the number of runs is too large to be completed under steady state conditions, you may block on replicates. Blocking allows you to estimate the block effects independently of the experimental error.

For example, if you have three factors with two levels each and you test all combinations of factor levels (full factorial design), one replicate of the entire design would consist of 8 runs (23). You can choose to run the design once or have multiple replicates.

Your experimental design includes the number of replicates you should run. Considerations for replicates:

·    Screening designs to reduce a large set of factors usually don't use multiple replicates.

·    If you are trying to create a prediction model, multiple replicates may increase the precision of your model.

·    If you have more data, you may be able to detect smaller effects or have greater power to detect an effect of fixed size.

·    Your resources may dictate the number of replicates you can run. For example, if your experiment is extremely costly, you may be able to run it only once.