ABSTRACT

Consider the data in Table 40.8. It is proposed that a new drug ‘Novom’ is effective in the treatment of postoperative nausea and vomiting. It is to be compared with cyclizine. The data satisfy the criteria for a binomial distribution, described earlier. The data is arranged into a 2×2 table (contingency table, Latin square). The null hypothesis is defined, i.e. cyclizine and Novom are equally effective in the treatment of nausea and vomiting. Statistical significance will be assumed at a probability of P<0.05. The test chosen for this type of data is ?2 (chi-squared). The mathematics behind a ?2 relationship is beyond the scope of this chapter and, as long as the conditions for its use are satisfied, detailed knowledge of the distribution is unnecessary.