抄録
In system suitability tests, the daily conditions of an analytical instrument are checked with the standard deviation (SD) of measurements as a criterion. This paper examines how exactly the SD values obtained from the measurements can indicate instrumental conditions. The HPLC measurement for acetaminophen is repeated six times daily to obtain an SD estimate, and these procedures are repeated for seven days. The degree of scattering of the seven SD values corresponds well to the 95% confidence intervals calculated from the chi-square distribution (n=6). Therefore it is concluded that the variability of the daily conditions of the HPLC system used in this study is too small to detect by referring to the SD estimates (n=6) and the daily change in the SD estimates only represents the error of the SD estimation itself. However, if the repetition number is 40 (n=40), the SD estimates will be a good indicator. Forty measurements are practically impossible for a slow analysis like HPLC, and this paper recommends the function of mutual information theory as a substitute for repetitive measurement.
