Feed supplement and beef from cows
Use this diagram to explain that estimates of the effect of a factor are more accurate if there is less variability between the experimental units.
The diagram models an experiment to assess the effect of a feed supplement on the weight of beef obtained from adult cows. The experiment is conducted on a herd of calves of different weights and half (at random) get the supplement.
The diagram simulates a completely randomised experiment that does not take account of the different initial weights of the calves. Click Accumulate then repeate the experiment several times.
Now use the slider to repeat the experiment with calves of similar initial weights. Again repeat the experiment several times.
The estimate of the effect of the supplement is more accurate when the calves have similar initial weight.