Ggression variable using a combination of the Bayesian Information Criterion (BIC

April 3, 2018

Ggression variable using a combination of the Bayesian Information Criterion (BIC) and likelihood ratio tests (LRT) (BIC; Nagin, 2005; LRT; Muth Muth , 2012). Fourth, we estimated dual trajectory order Leupeptin (hemisulfate) models where the social and physical trajectories were formed jointly rather than individually (Nagin, 2005). Last, we examined whether family factors predicted group membership in social and physical trajectory categories. Growth and prediction models were estimated using a combination of Mplus (Muth Muth , 1998?012) and Stata (StataCorp, 2011). In these analyses we considered the metric of the aggression variables; both physical and social aggression were assessed by teacher ratings that peaked at the lowest value (one) and were then skewed out to the maximum value (five). Following the recommendation of Nagin (2005), we analyzed the natural logarithm of the variables to account for the skewed nature of the data and used a censored normal (tobit) likelihood model to account for the concentration at the minimum value. In order to account for missing data in the construction of the trajectories, we used a maximum likelihood approach that allowed all observations to contribute to the estimated results (Muth Muth , 1998?012). The only constraint was that participants were required to have a minimum of two out of the nine possible teacher reports of aggression. To assess the fit of the trajectory models we used methods developed for mixture models. We assessed the reliability of the results of the models by computing the average posterior probability of assignment (AvePP) and the odds of correct classification (OCC; Nagin, 2005). These are both based on participants being assigned a probability of being in a class, j, through the estimation process, within the aggression type being estimated. The AvePP is a measure of the reliability of the model determined by averaging the actual (posterior) probability of being assigned to the class to which the student is eventually assigned. The OCC for class j is computed by:In this formula, the numerator is the odds of correct assignment based on the average posterior probability and the denominator uses the estimated population proportion of class j, j, and provides an estimate of what the odds are of a participant being classified in class jAggress Behav. Author manuscript; available in PMC 2015 September 01.Ehrenreich et al.Pageif they were 1-Deoxynojirimycin web randomly assigned. Thus, a higher OCC suggests better classification by the model compared to just randomly assigning students to a class. The guidelines developed by Nagin (1999, 2005) state that an AvePP of assignment of 0.70 or greater for each class is acceptable as well as having an OCC greater than five for each group. Descriptive Statistics Descriptive statistics and correlations are presented in Table 1. The overall pattern of correlations showed significant, positive relations between different teachers’ ratings of participants’ social and physical aggression, even across many years. There were positive relations between both authoritarian and permissive parenting with both forms of aggression at many of the time points. Growth models We began by constructing multilevel (hierarchical) linear models for social and physical aggression across grades three through twelve. Although we built both linear and quadratic versions of these models, in anticipation of the mixture models below we illustrate the linear versions of each. Let yit be either the social.Ggression variable using a combination of the Bayesian Information Criterion (BIC) and likelihood ratio tests (LRT) (BIC; Nagin, 2005; LRT; Muth Muth , 2012). Fourth, we estimated dual trajectory models where the social and physical trajectories were formed jointly rather than individually (Nagin, 2005). Last, we examined whether family factors predicted group membership in social and physical trajectory categories. Growth and prediction models were estimated using a combination of Mplus (Muth Muth , 1998?012) and Stata (StataCorp, 2011). In these analyses we considered the metric of the aggression variables; both physical and social aggression were assessed by teacher ratings that peaked at the lowest value (one) and were then skewed out to the maximum value (five). Following the recommendation of Nagin (2005), we analyzed the natural logarithm of the variables to account for the skewed nature of the data and used a censored normal (tobit) likelihood model to account for the concentration at the minimum value. In order to account for missing data in the construction of the trajectories, we used a maximum likelihood approach that allowed all observations to contribute to the estimated results (Muth Muth , 1998?012). The only constraint was that participants were required to have a minimum of two out of the nine possible teacher reports of aggression. To assess the fit of the trajectory models we used methods developed for mixture models. We assessed the reliability of the results of the models by computing the average posterior probability of assignment (AvePP) and the odds of correct classification (OCC; Nagin, 2005). These are both based on participants being assigned a probability of being in a class, j, through the estimation process, within the aggression type being estimated. The AvePP is a measure of the reliability of the model determined by averaging the actual (posterior) probability of being assigned to the class to which the student is eventually assigned. The OCC for class j is computed by:In this formula, the numerator is the odds of correct assignment based on the average posterior probability and the denominator uses the estimated population proportion of class j, j, and provides an estimate of what the odds are of a participant being classified in class jAggress Behav. Author manuscript; available in PMC 2015 September 01.Ehrenreich et al.Pageif they were randomly assigned. Thus, a higher OCC suggests better classification by the model compared to just randomly assigning students to a class. The guidelines developed by Nagin (1999, 2005) state that an AvePP of assignment of 0.70 or greater for each class is acceptable as well as having an OCC greater than five for each group. Descriptive Statistics Descriptive statistics and correlations are presented in Table 1. The overall pattern of correlations showed significant, positive relations between different teachers’ ratings of participants’ social and physical aggression, even across many years. There were positive relations between both authoritarian and permissive parenting with both forms of aggression at many of the time points. Growth models We began by constructing multilevel (hierarchical) linear models for social and physical aggression across grades three through twelve. Although we built both linear and quadratic versions of these models, in anticipation of the mixture models below we illustrate the linear versions of each. Let yit be either the social.