Discussion
Novel findings indicated that the BOSCC performed similarly across home
and school contexts. Specifically, scores improved in both contexts, and
SC and total scores were highly related across contexts. Children showed
improvements across all subscales and the total score over the course of
the year. Interestingly, there were larger improvements in the school
context. Larger gains in the school context may be attributed to
children receiving the intervention in school. Behavioral changes may
not generalize across contexts (Gunning et al., 2009) so for children to
show improvements in multiple settings, an intervention may need to
include home-school collaborations.
For both timepoints, the SC subscale and total scores were highly
correlated across both contexts, unlike the RRB and OAB. This makes
sense given that some RRBs likely are context-bound or elicited by
specific stimuli (e.g., familiar adults, specific routines) that may be
more or less present in one context versus another, aligning with
previous research suggesting the presence of RRBs often differ between
contexts (Stronach & Wetherby, 2014). Overall, the RRB subscale may be
more influenced by contextual features, whereas the SC subscale appeared
to be more stable.
While subscale and total scores were correlated at the individual
timepoints, we did not find a strong correlation between the changes in
child scores across contexts over time. Thus, researchers must carefully
consider the purpose of their study when selecting outcome measures and
measurement contexts. If the goal is to measure child skill acquisition,
measuring the behavior in a context similar to the intervention is
likely more important (e.g., if the intervention was implemented by a
teacher at school, then outcomes should be measured at school); however,
if the goal is to measure the generalization of skills, then measuring
across different contexts becomes more important.