If the adolescent reports smoking within the past thirty days at a given wave, the adolescent��s data are added to the continuous portion of the model. LGCM was conducted Bicalutamide supplier using Mplus v. 6.0 software (Muth��n & Muth��n, 1998�C2010). Missing Data Because hedonic capacity was not measured until Wave 4 (N = 1,136), we used Wave 4 (age 15�C16 years) as the baseline for these analyses. Due to wave nonresponse and loss to follow-up, the number of adolescents who completed a survey in the subsequent waves were 1,110 (Wave 5), 1,092 (Wave 6), and 1,090 (Wave 7). Thirty adolescents had missing data on the covariates at Wave 4 and were not included in the analysis. To account for 14�C16 cases with missing smoking data at Waves 6 and 7, multivariate modeling used all available data.

Mplus allows modeling with missing data using maximum likelihood estimation of the mean, variance, and covariance parameters, employing the expectation maximization algorithm when data are missing at random (Muth��n & Muth��n, 1998�C2004). Final analyses were based on 1,106 adolescents. Results Descriptive Statistics Table 1 presents the means and SDs for continuous model variables, along with the proportions for the categorical model variables. Cross-sectional tabulations for cigarette smoking in the past thirty days indicated that the proportion of adolescents who did not smoke in the past month decreased from 89% at Wave 4 baseline to 87% at Wave 7. At the same time, the mean number of cigarettes smoked among those adolescents smoking at least one cigarette in the past thirty days increased from 3.

43 (SD = 3.64) at Wave 4 baseline to 3.67 (SD = 3.81) cigarettes at Wave 7. The average hedonic capacity score for the sample was 33.03 (SD = 8.32), which is very similar to mean scores in previous samples of undergraduates (M = 33.6 in Franken et al., 2007; M = 34.4 in Leventhal et al., 2006). Table 2 provides correlations between hedonic capacity and the study variables. Table 1. Characteristics of the Sample Table 2. Correlations Between Hedonic Capacity and the Study Variables Multivariate Model We began by assessing the binary and continuous conditional models separately for fit to the data (i.e., two separate models). This permitted identifying the average trajectory shape for each part. For the binary part, a linear growth curve fit the data well, ��(20,1106)2=24.91, p = .20, comparative fit index (CFI) = 1.00, root mean squared error of approximation (RMSEA) = 0.01, Weighted Root Mean Square Residual = 0.54. We next assessed the continuous model, modeling only data from participants smoking Dacomitinib at least one cigarette in the past thirty days. A visual inspection of the observed mean plot suggested a quadratic trend.