Swaim Randall C
Tri-Ethnic Center for Prevention Research, Department of Psychology, Colorado State University, Fort Collins 80523-1879, USA.
Prev Sci. 2003 Jun;4(2):89-98. doi: 10.1023/a:1022922231605.
A hierarchical linear model was used to estimate the individual and school level effects for marijuana use among a national sample of 12th-grade students. School effects were small in comparison to individual level effects, accounting for 2.9% of the variance in marijuana use. At the individual level, perceived harm, perceived availability, and their interaction were significant predictors, each of which varied randomly across schools. Among two school-level predictors, the normative environment for perceived harm was not significant, but normative perceived availability predicted level of marijuana use. The effect of perceived availability on marijuana use was stronger in larger, compared to smaller communities. Results are discussed in light of the use of random regression methods for identifying school-specific patterns of risk and protection for prevention planning.
采用分层线性模型,对全国12年级学生样本中使用大麻的个体和学校层面的影响进行估计。与个体层面的影响相比,学校层面的影响较小,占大麻使用差异的2.9%。在个体层面,感知到的危害、感知到的可获得性及其相互作用是显著的预测因素,每个因素在不同学校间随机变化。在两个学校层面的预测因素中,感知到的危害的规范环境不显著,但规范的感知到的可获得性预测了大麻使用水平。与较小社区相比,在较大社区中,感知到的可获得性对大麻使用的影响更强。根据使用随机回归方法识别特定学校的风险和预防模式以进行预防规划的情况,对结果进行了讨论。