Sample stability: a suggested definition and method of determination.

Abstract

Abstract We propose defining the stability of any chemical constituent of stored samples in terms that are quantitatively related to the precision of the measurement by which it is determined. We suggest that a constituent may be considered stable, for a stated period and under exactly defined conditions, when the average change in its measured value is less than a chosen number, K, of standard deviations of the data obtained by the measuring method over the concentration range in question. Based on this definition a technique utilizing a graphical truncated normal sequential test is presented as the appropriate experimental and statistical design for measuring stability. The statistical basis for the proposal is presented.