This International Standard describes specific statistical methods for addressing the following questions.
a) Given a population of items from which a sample of n items has been drawn, x of the sample items are found to show a specified characteristic. What proportion of the population has that characteristic? (See A forms, subclause 8.1.)
b) Is the proportion estimated in a) different from a nominal (specified) value? (See B forms, subclause 8.2.)
c) Given two distinct populations, are the proportions with the characteristic in the two populations different? (See C forms, subclause 8.3.)
d) In b) and c) how many items must be sampled in the population(s) to be sufficiently sure that the result of the test is correct? (See 7.2.3 and 7.3.3.)
It is essential that the drawing of samples does not have any appreciable effect on the population. If the sample drawn at random is less than 10 % of the population this is usually satisfactory, but if the sample is greater than this, reliable results can be obtained only by replacing each item sampled before drawing the next item at random from the population.