"Increased Drug Use Among Old-for-Grade Adolescents."
From the Archives of Pediatrics & Adolescent Medicine--a medical, not educational, journal. I haven't be able to find this one online at all, but it's from Volume 150(5), May 1996, pp 470-476.
This is from the Abstract:
Thirty-six percent of adolescents surveyed were old for grade. Adjusting for multiple potential confounders, old-for-grade high school students were more likely to report being regular smokers, chewing tobacco, drinking alcoholic beverages, driving in a car with someone who had been drinking, using alcohol or other drugs before last sexual intercourse, using cocaine in the past month, ever using crack, and using injected or other illicit drugs.
Conclusions: Old-for-grade status is a potentially important marker for drug-related behaviors in adolescents. The antecedents of adolescent risk-taking behavior may begin before the teen years, and prevention of school failure or interventions targeted toward old-for-grade children could affect their propensity to experiment with or use drugs during adolescence.
And this from the conclusions:
The high proportion of students who are old for grade (36%) is striking. This rate varies from 42% in the ninth grade to 30% in the 12th grade. We believe that this observation represents a true decline in adolescents who are old for grade in the higher grades, because there is a disproportionate number of dropouts and school absence among the older old-for-grade students. The magnitude of association between being old for grade and substance use probably would have increased if we had used a population sample that included youths who were absent or who had dropped out of school. Although further determination of the magnitude and potential causes for increased substance use among subgroups of old-for-grade adolescents is warranted, for many drug-related behaviors, old-for-grade status was the most significant factor, and these older students were at twice the risk.