Analytical procedures are usually based on the assumption that random errors are inevitable and must be accommodated by overdetermining the value sought. For example, if one wishes to determine the absorptivity of a compound at a particular frequency he will generally measure the absorbance for several concentrations and determine the best straight line (after demonstrating Beer's law to be valid for the system). One absorbance and concentration would be sufficient except for the experimental uncertainties.
Robert P. Bauman, "A Least Squares Method For Multicomponent Analyses," Appl. Spectrosc. 13, 156-157 (1959)