Table 3.
School-level factors that shape the direction and magnitude of the heritability of daily smoking
| p.e. | beta | s.e. | t | pr < | |
|---|---|---|---|---|---|
| Social and demographic characteristics | |||||
| Proportion of college-educated mothers | −25.23 | −0.357 | 23.78 | −1.061 | 0.292 |
| Proportion non-Hispanic and white | −6.55 | −0.685 | 2.41 | −2.718 | 0.008 |
| Smoking norms | |||||
| Popular students do not smoke | −6.77 | −0.177 | 8.02 | −0.845 | 0.401 |
| Popular students are also smokers | 51.04 | 1.334 | 8.56 | 5.962 | 0.000 |
| Institutional control of smoking | |||||
| Teachers not allowed to smoke on campus | −3.31 | −0.144 | 5.05 | −0.656 | 0.514 |
| School penalties for smoking infractions | 1.72 | 0.264 | 2.34 | 0.736 | 0.464 |
| Smoking prevalence | |||||
| Proportion of students who have smoked | −19.32 | −0.213 | 174.91 | −0.110 | 0.912 |
| Smoking prevalence squared | 31.01 | 0.233 | 236.9 | 0.131 | 0.896 |
Note: Cell entries are parameter estimates the latent school-level heritability factor for daily smoking regressed on various school-level factors. These models were estimated using the GEQS command in the GLLAMM procedure available in STATA 9.2. Data obtained from the sibling and twin pair sample of the National Longitudinal Study of Adolescent Health (n = 1,198 pairs). Parameter estimates were weighted for individual and school-level weights (Chantala and Tabor 2004). The inclusion of these estimates significantly improved overall fit (Chi-square = 16.38, df = 8, P < .037)