Bulk Samples for the Testing of Glass Containers

JSGT 1932 V16 T022-T033

The paper demonstrates the importance of sample size and sampling procedure for tests to determine the probable performance in service of glass containers. Various tests, under conditions which gave positive breakages (5–10%) with satisfactory articles, are quoted and the results set out to show the effect. on the accuracy of the result, of sample size and of (described) sampling procedures. The results indicate that (1) appreciable errors are possible with too small samples, (2) for research and similar tests, the sample should consist of at least500 systematically selected articles, and (3) for routine tests, samples need not be so large as for research tests, but with “patchy” or, “border-line” quality articles larger samples are required than with regularly “good” or “bad” articles to give trustworthy indications of quality.

E. J. C. Bowmaker & J. D. Cauwood