Abstract: The conclusion drawn from the content uniformity test for a batch of solid dosage forms is influenced by the random error of the analytical method(random methodic error for short) used. Theoretically, the test by variables based on the statistical parameters of the sample has the merit of eliminating the influence of the random methodic error acccording to the principle of superposition of errors. In this paper, random methodic errors of some analytical methods were obtained by experiment and then used to eliminate the influence of random methodic errors on the content uniformity test by variables devised by us. Thus, the accuracy of the test is further improved.