Missing Data
At T1, participants answering less than 90% of each scale were excluded from
further analysis (n = 13). At T2, included T1 participants answering less than
90% of items on the bully, victim and disengagement scales were excluded
from subsequent analyses (n = 7). For the rest of the sample, missing data
were imputed using the expectation maximization algorithm (see Shafer &
Graham, 2002). For all variables, less than 2% of data were imputed. All
summation of continuous scores was conducted after imputation.