Visiting professor of civil engineering, Drexel University, Philadelphia, Pa.
Results of materials tests are influenced not only by the measured property but also by such distorting factors as random and regular errors in testing or sampling. Knowledge of the magnitude of such factors is useful, since the smaller their contribution to the test results the more reliable the results are. It is argued that the most suitable method of obtaining such knowledge is breaking down the total variation in test results into meaningful components such as testing variance, sampling variance, and material variance. Application of this method is illustrated for the ASTM Test for Slump of Portland Cement Concrete (C 143-74). This illustration also shows that, contrary to general belief, this test is no less reliable than the standard tests for compressive strength, air content, or unit weight determinations of concrete.
Paper ID: CCA10171J