This paper investigates the relationship between urban green space and academic achievement (standardized test scores) for public elementary schools in New York City. In 2010 and 2017, I find evidence of a neighborhood-level positive association of tree canopy with test scores. However, there is little evidence of any association when only tree canopy close to schools is considered. Results are robust to hierarchical mixed-effects model specifications. I also conduct one of the first longitudinal analyses of the green space-educational outcomes relationship. I find initial evidence that increases in tree canopy around elementary schools between 2010 and 2017 were associated with decreased grade-average test performance for high-poverty schools and were not significantly associated with changes in test performance of non-high poverty schools, a result that somewhat contradicts the positive comparative static results. To reconcile my findings, I construct an extended econometric model detailing potential bias in the green space-educational outcomes relationship that might arise from unobserved changing dimensions of inequality, with the adoption of Common Core curricula starting in 2013 serving as a pertinent example. Standard econometric approaches do not capture time-varying unobserved characteristics associated with observational data, so omitted factors that covary both with changes in green space and changes in test scores may bias the estimated green space-educational outcomes relationship. Future researchers should intentionally consider the co-development of urban green space with salient unobserved factors when estimating and interpreting the effect of changes in green space on educational outcomes over time.