StatisticsiN Action ProbabiliTy iN a REVEI'SE CocaiNE STiNg: Was CocaiNE REally Sold? The American Statistician (May 1991) described an inter- esting application of a discrete probability distribution in a case involving illegal drugs. It all started with a "bust" in a midsized Florida city. During the bust, police seized ap- proximately 500 foil packets of a white, powdery substance, presumably cocaine. Since it is not a crime to buy or sell nonnarcotic cocaine look-alikes (e.g., inert powders), de- tectives had to prove that the packets actually contained cocaine in order to convict their suspects of drug traffick- ing. When the police laboratory randomly selected and chemically tested 4 of the packets, all 4 tested positive for cocaine. This finding led to the conviction of the traffickers. After the conviction, the police decided touse the remaining foil packets (i.e., those not tested) in reverse sting operations. Two of these packets were randomly selected and sold by undercover ofcers to a buyer. Between the sale and the arrest, however, the buyer disposed of the evidence. The key question is, Beyond a reasonable doubt, did the defendant really purchase cocaine? In court, the defendant' s attorney argued that his client should not be convicted because the police could not prove that the missing foil packets contained cocaine. The po- lice contended, however, that since 4 of the original packets tested positive for cocaine, the 2 packets sold in the reverse sting were also highly likely to contain cocaine. In this chapter, two Statistics in Action Revisited examples demonstrate how to use probability mod- els to solve the dilemma posed by the police' s reverse sting. (The case represented Florida' s first cocaine-possession con- viction without the actual physical evidence.) StatisticsmAction Revisited ' " Using the Binomial Model to Solve the Cocaine Sting Case (p. 210) ' Using the Hypergeometric Model to Solve the Cocaine Sting Case ([3. 221)