Sir.—We would like to congratulate Gerber et al1 on their careful collection of data comparing BID and TID penicillin therapies for group A β-hemolytic streptococcal pharyngitis. However, the analysis of those data was incomplete and it therefore misled the authors into making improper conclusions.
Their data showed a difference in failure rates of 28% vs 18% in favor of TID therapy. However, this difference was not statistically significant by the χ2 test (P=.21), and the authors concluded that the investigation suggested "that oral penicillin V given twice daily is as effective as penicillin V given three times daily." The P value cited here only indicates the likelihood of type I or α error: the probability of accepting two forms of therapy as different when, in fact, they are the same. That is, these data do not permit the conclusion that the two treatments are different, for