By senior year of high school, 86.2% of participants had initiated drinking and 68.4% had been intoxicated, with the percentage of those reporting AO and AI increasing over time. Participants were not asked questions about heavy drinking or problems until they reported an AO; therefore, they had “missing” data for these variables until drinking commenced. Although these “missing” data could have been handled with missing data procedures, the “missing” values may have been assigned nonzero values, implying that the participants were drinking when they were not. Rather than excluding the 160 participants who had yet to start drinking by senior year of high school, we set the values for their “missing” drinking data to 0. See Table 1 for the semiannual distributions of the percentage of participants reporting AO/AI, the mean frequencies of heavy drinking/and problems, and the percentage of participants reporting heavy drinking/problems at rates above the mean.
