The study provides compelling evidence on the impact of informal peer interactions on academic and behavioral outcomes, but a few aspects could use more clarification. While the methodology accounts for changes in bus routes to establish causal effects, were any sensitivity tests performed to ensure robustness against unobserved confounders, such as family relocation or school choice policies?
Moreover, the results suggest stronger peer effects in adolescence, particularly for behavioral outcomes. Could the authors discuss whether these effects persist over time? Do peer influences on the bus translate into long-term changes in student performance or social behavior beyond school years?
Your estimates rely on the assumption that changes in bus peer groups at school transitions are conditionally exogenous. However, given that these transition points are often moments when families make strategic decisions, such as residential moves or opting into different school programs, how do you rule out the possibility that observed behavioral changes are partly driven by selection into new neighborhoods or peer environments rather than by the social interactions on the bus itself? This seems especially relevant for behavioral outcomes, which may be more sensitive to unobserved family-level motivations or constraints. Can you clarify how your model accounts for this potential source of bias?