Classroom-Level Influences on Children's Skills in the Early School Years

Introduction. Several recent studies have examined the effects of classmates’ cognitive or behavioral skills on children’s own cognitive and behavioral skill development in the early school years. For example, children have been shown to make greater academic skill gains across the school year if their classmates have higher average levels of cognitive and language skills (Henry & Rickman, 2007; Justice, Petscher, Schatschneider, & Mashburn, 2011; Mashburn, Justice, Downer, & Pianta, 2009) or higher average self-regulation skills (Skibbe, Phillips, Day, Brophy-Herb, & Connor, 2012). In addition, lower classroom-level withdrawn and socially reticent behavior has been shown to predict greater gains in children’s social competence (Bulotsky-Shearer, Dominguez, & Bell, 2012), and greater classroom-level aggression has been shown to uniquely predict increases in children’s aggression in the early school years (Thomas, Bierman, & Powers, 2011). Classroom-level effects may operate via direct processes, such as through skill transfer among peers during small-group activities or play. Indirect processes may also be at play: For example, teachers may be able to maximize instructional time or speed up the pace of instruction in classrooms with higher average cognitive skills or more regulated, engaged behavior (Henry & Rickman, 2007; Justice, Petscher, Schatschneider, & Mashburn, 2011).

Current Study. The current study broadens the existing research base on classroom-level influences in the early school years by examining a wider range of classroom-level predictors and individual-level outcomes in the behavioral domain, with a particular emphasis on social-emotional variables. Specifically, we will test whether classroom-level prosocial behavior, classroom-level aggressive behavior, and classroom-level engaged behavior predict change in children’s individual-level prosocial, aggressive, and engaged behavior.

Method & Analysis Plan. Participants were 1124 children in a high-risk, low-income urban school district. Peer ratings were collected in kindergarten, first, and second grade during one-on-one sociometric interviews. Peers were asked to list classmates who fit various behavioral indicators (e.g., friendly, cooperative, mean, starts fights, works hard, doesn’t do their work), and these nominations were used to calculate individual scores which represented the proportion of classmates who rated that child as fitting a given behavioral indicator. These proportions were then averaged to create classroom-level behavioral aggregates for each classroom. Multi-level models (MLM) will be used to account for the nesting of time within students within classrooms. Level 2 predictors will include individual-level prosocial, aggressive, and/or engaged behavior, and Level 3 predictors will include classroom-level prosocial, aggressive, and/or engaged behavior. We hypothesize that classroom-level behavior will predict growth in children’s skill development controlling for prior individual-level skills.