Thank you for the opportunity to review your paper. This is a very interesting research question, and you use a rigorous approach to interrogate the relationship between mental health problems and academic achievement using multiple informants and assessing mental health at developmentally important milestones.

Missingness
While you used MICE to account for missing covariates in the sample, you do not address the larger issue of large loss-to-follow-up from the initial sample of 9,778 enrolled in the cohort to the 1,577 which were included in this study. This could be summarized in a PRISMA diagram to clearly show where the missingness occurs. This level of missingness may have created bias in the estimates, likely impacts the representativeness of the sample, and has implications for the generalizability of the results.

Results
The results section could be re-written to more clearly communicate the main findings.

When discussing the overarching scales, there was weak evidence for internalizing problems and the IQ-achievement discrepancy, but no evidence of an effect in the teacher-report.

When describing the overall scales, you indicate that in mother-report, externalizing problems are comprised of aggressive behaviour and attention problems, while in the teacher report, you only note rule breaking and aggressive behaviour. Is attention problems not included in this overarching scale?

Could you generate overarching estimates that are comparable between the informants (e.g. drop emotionally reactive behaviour from mother report so the internalizing problems is based on the same 3 subscales; and include attention problems in the estimate of externalizing problems in the teacher report?) This would increase the comparability of the results across informants.

Minor notes – Throughout the manuscript, you use “significant” often to describe the results. I am not sure if this is permitted in the journal guidelines, but you may wish to re-word some of these statements to avoid relying on the concept of statistical significance to describe the findings.

You present a lot of estimates in this paper, so I would suggest removing the subscales from the parent report and teacher report which did not overlap from Figure 1 to streamline this. Also, Table 1 includes a mention of stimulant usage which is not discussed elsewhere (e.g. as a covariate), so could be dropped from the table.

Thank you for the detailed attention to reviewer comments, which have strengthened the manuscript.