Abstract
Extract
Sir,—Robertson (Citation1963a), in discussing the evaluation of tuberculin tests, neither distinguishes between absolute and comparative evaluation nor recognizes that the evaluation of an ancillary test is merely a special case of comparative evaluation. The consequence of these shortcomings is his quite unwarranted conclusion (p. 8) that “the use of statistics based on the classification of slaughtered reactors only is of no value whatsoever in evaluating a test”. This statement is true only in the context of absolute evaluation.