Abstract It is important for funding agencies to evaluate if scientists accomplish their research goals. By comparing a representative sample of National Science Foundation abstracts and project outcome reports (PORs) from 2014 to 2017, this article investigates whether scientists attain the broader impacts they propose. We find that the number of broader impacts proposed in the abstracts is significantly higher than the number of broader impacts reported in the PORs. The trend is common across directorates and type of impact, except when impacts serve advantaged groups. Only the number of broader impacts for advantaged groups increases from the abstract to the POR. Despite the difference between proposed impact and reported impact, our study does not conclude that scientists are delinquent or disingenuous when they propose their research. Rather, we question the capacity of current frameworks to capture the quality of impacts and to weigh the relative importance of impacts that serve marginalized groups versus those that sustain the status quo.