Sampling in which observations are taken one at a time, with an appraisal, after each observation, of whether there is any need for further sampling. This approach is useful when taking an observation is very expensive, or when there are ethical considerations. After each observation has been taken, the data so far available are re-analysed and one of three possible decisions is taken: accept the null hypothesis, reject the null hypothesis, or take another observation. See also double sampling.
Subjects: Probability and Statistics.