Caveats concerning the use of computer random-number generators to select urban transportation system employees for drug tests

Abstract

Federally mandated drug-testing programs for urban transportation system employees require routine testing on a random basis. In order to assure random selection, computer random-number generators often are used in the process. Unfortunately, some of these generators cannot be relied upon to produce sequences of numbers which pass even the most fundamental statistical tests for randomness. Consequently, urban transportation system employees chosen for drug testing by using these flawed generators will not have been selected in a manner consistent with federal guidelines. Thus, those who use computers in the random selection process should ensure that the random-number generators pass two fundamental statistical tests for randomness. These tests are discussed herein, with application to a particular microcomputer-software combination, utilizing a selection of potential seeds. Also, procedures are suggested which will ensure that an otherwise well-designed testing process will not be tainted by its failure to satisfy fundamental randomness requirements.