Statistical analyses: An ANOVO or Student's t-test was performed if the data were met the normality assumption. A unidentified non-parametric test was performed when normality was not met.

Other potential concerns: IQ scores were reported as a percentile rather than an actual IQ score without description of how these were determined. The study also didn't have a large gradient on exposure or outcome, which could bias the results towards the null. There were no children in this population with either grade 1 or 2 (very intelligent or intelligent, respectively) and only 1 child that was intellectually impaired. Exposure was based on a 6 point scale rather than a continuum with only 1 child in normal and 5 children in questionable and nearly half the children in Moderate.

Tamayo De la Cruz 2022

Risk of bias

Epidemiological study populations

HAWC