Although parental help with homework has been long understood as a valuable part of parental involvement with children's education and, as such, an important mechanism shaping educational inequality, previous studies showed contradictory findings about its influence on academic achievement. Using multiple waves of the Early Childhood Longitudinal Study (ECLS-K), namely 1st-, 3rd-, and 5th-grade data from ECLS-K 1998--1999, and 1st-, 2nd-, 4th-, and 5th-grade data from ECLS-K 2011, we examined the associations between parental help with homework and children's reading and math achievement in elementary school. The initial results from cross-sectional and value-added models showed a negative association between daily parental help with homework and academic achievement. However, child fixed effects models, which address possible bias due to unobservable heterogeneity, no longer show such association. Our findings, including different sensitivity analyses, showed no statistically significant association between parental help with homework and student achievement. Further, the association between parental help with homework and achievement did not vary by parental level of education or child's achievement level. We discuss possible mechanisms behind the findings and their implications for educational policy and parenting literature.