45; Mq = ?0.01; Di = 0.06, t = 8.23; Ds = 0.03, t = 4.68; Dq = selleck kinase inhibitor 0.001, t = 5.03). The intercept and slope were significantly and negatively correlated, indicating that lower initial levels of sensation seeking were associated with faster increases across grades (r = ?.41, p < .01). The intercept and quadratic parameters were significantly but positively correlated (r = .26, p < .05), indicating that a lower initial level of sensation seeking was related to a downturn in middle school. The slope and the quadratic parameters were significantly and negatively correlated (r = ?.93, p < .000) indicating that higher increases in sensation seeking were also associated with a downturn in late middle school.

Since the linear and quadratic factors were highly correlated, the quadratic component was retained in model, but only the factor scores from the intercept and linear slope were used in further analyses (see http://statmodel.com). Relating Childhood Sensation Seeking to High School Smoking Trajectory Classes We tested three orthogonal models (a) Stable High Smokers, Experimenters, and Rapid Escalators versus the Nonsmokers; (b) Stable High Smokers versus Rapid Escalators and Experimenters; and (c) Rapid Escalators versus Experimenters. Across all three models, the interactions of the intercept and slope of sensation seeking with gender were not significant. Table 2 presents the final simplified models with nonsignificant interactions of intercept and slope of sensation seeking with gender removed. Table 2.

Regression Parameters (B) and Odds Ratios (OR) for Trait Predictors in Elementary School for Cigarette Smoking Class Comparisons in High School Girls were more likely than boys to be members of a smoker class compared with the Nonsmoker class (contrast 1), to be in the Stable High Smoker class compared with the Rapid Escalators plus Experimenters (contrast 2), and to be in the Rapid Escalator class compared with the Experimenter class (contrast 3). Having already tried cigarettes by 4th grade was only a significant predictor in contrast 2: those who had already tried cigarettes were more likely to be a Stable High Smoker as opposed to a Rapid Escalator or Experimenter. Having been on the free or reduced lunch program predicted membership in any smoker class (contrast 1), in the Stable High Smoker class (contrast 2), and in the Rapid Escalator class (contrast 3) compared with the respective contrast classes.

Higher initial level (intercept) of sensation seeking increased the likelihood of being in any smoker class compared with being a Nonsmoker (contrast 1). Higher initial level (intercept) and rate of growth AV-951 (slope) of sensation seeking increased the likelihood of being in any smoker class compared with being a Nonsmoker (contrast 1). Because the intercept and slope covaried, to illustrate the independent effects of one, the effect of the other was held constant at the mean.