When you say empirical argument, it means facts gathered and concluded through a series of experimentation and reliable data gathered and not through theory or speculation.
Therefore, a sample size can greatly affect the validity of an empirical argument once proven inaccurate because what we want to see in an empirical argument is the clear-cut reality of what the researchers gathered through meticolous experimentation and not otherwise.
300ml of ethanol is not the same as 300L of ethanol and 2 grams of salt is not 2 kilos of salt. However, if it can be explained thoroughly that a sample size is just a fraction representation of the original, it is wise to create a control subject to compare the data and make it more reliable. Say for example, you wouild like to compare the sun and the earth, make sure to make the models realistically proportional in your miniature globe models.
X + y + z = 32 * 3
27 + 27 + x = 96
X = 54
2x = 108
Average of x and 2x = (54+108) / 2 = 164 / 2 = 82
Average of X and 2X = 82
