Application of Multi- Level Approach in Detemining the Role of Inefficient Metacognitive Componenets on the Tendency of School Students to Drug Abuse

The main goal of this research was to apply the multi-level in determining the role of metacognitive components associated with drug abuse among adolescents of Kerman Province. The method used in this research was correlation type. In order to conduct this research 1000 school students of Kerman Province have been selected by using multistage cluster sampling method. For data collection in this research, ZARGAR drug addiction preparation Questionnaire and Wels’s Metacognitive Questionnaire (MCQ30) have been used. The results of multi-level approach using One-way random effects Anova showed that the four components (positive beliefs about worries, negative beliefs about worries, low cognitive trust and cognitive self-awareness) explained about 49% of the variance of addiction tendency at school level. Moreover, the random intercept model explained 12% of the student-level variance of drug abuse tendency. Here, the significance of qi-square with the value of 37.983 at the significant level (P <0.0001) reflects the significant change of the null model and model 1. In response to the question that how much the relationship between the tendency to drug abuse and student-level components is similar among schools, a comprehensive model with both student-level and school-level variables (model 2) was used. In model 2, the results showed that at the school level, the mean of positive beliefs and cognitive self-awareness about drug abuse tendency have a significant relationship, but the mean of cognitive trust and negative beliefs has not a significant relationship with the tendency to drug abuse. At the students level, there is a significant relationship between cognitive trust and negative beliefs with drug abuse tendency (P <0.01). In total, it can be said that for variables related to samples with a nesting nature, it is necessary to use all levels in data analysis.