Distinguishing the treatment effects from randomness
In a completely randomised experiment, there are two potential reasons why the response in an experiment varies between experimental units. We must distinguish between these in order to estimate the effect of the experimental factors.
Using knowledge about the amount of random variation in the experiment, we can better assess whether or not observed differences between the treatments are more than can be attributed to chance.
Crop experiment in a field