Peer victimization, student engagement, and school attendance: Structural equation models

Survey data from a study of 6th grade students in Colorado (n=860) were used to estimate structural equation models in which peer victimization types were hypothesized to have significant relationships with both student engagement and attendance. Then, student engagement and attendance variables were hypothesized to have significant effects on achievement (measured as grade point average). Student engagement was viewed as a multi-faceted construct, composed of behavioral, cognitive, and emotional aspects. Four different types of peer victimization variables (verbal, physical, exclusion, and intensity) were combined to form a latent measure for peer victimization that was expected to predict absenteeism and student engagement. In addition, student engagement was expected to act as a mediating variable between the peer victimization latent variable and absenteeism. A model treating peer victimization and student engagement as latent variables fit the data well. However, the peer victimization latent variable was not statistically significantly predictive of absenteeism as was hypothesized.;Other paths between endogenous and exogenous variables, although statistically significant, had relatively weak path coefficients suggesting that victimization does not largely impact attendance for the 6th grade students. In fact, the path coefficients between student engagement and attendance were also weak. In conclusion, the relationships between peer victimization, student engagement, and attendance were simply not as strong as hypothesized. However, the structural equation models did demonstrate that school engagement mediates the effect of peer victimization on attendance and achievement. A suggestion for further study would be to examine the "school avoidance" component of the study; perhaps, an attendance variable would be more significantly impacted by peer victimization for older students who have less parental influence on their daily attendance. In addition, a longitudinal study with more measures of student behaviors across time might better capture the effect of peer victimization on the various school behavior variables.;Keywords. Peer victimization, engagement, exclusion, attendance, achievement, bullying, SEM, repeated measures.