The experimental conditions under which an experiment is run should be representative of those to which conclusions of the experiments are to be applied.

However, an unfortunate consequence of increasing the scope of the experiment is an increase in the variability of the response

A random assignment of the subjects or experimental material to treatments prior to the start of the experiment ensures that observations that are favored or adversely affected by unknown sources of variation are observations "selected in the luck of draw" and not systemically selected.