Prospective studies have been held out as a kind of Holy Grail in research about remembering or forgetting child sexual abuse. They seem to hold the perfect answer to the verification problems that plague retrospective self-reports in the clinical literature. Prospective studies begin with verified cases of abuse. Then they require detective work years later to find the participants in adulthood and clever questioning to assure that any disclosed abuse actually matches the "target case." These studies are extremely difficult to construct and carry out. That is undoubtedly why there were only two prospective studies of memory for child sexual abuse in the literature before the recent publication of Goodman et al. (2003). The seminal study by Williams (1994) has been criticized in various ways that challenge the meaning of the 38% nondisclosure rate (Loftus, Garry, & Feldman, 1994). A more recent prospective study by Widom and Morris (1997) has been criticized by Goodman et al. (2003) for failing to probe carefully enough to assure that any disclosed abuse matched the "target case" in the population. Goodman et al. (2003) corrected for the latter problem by constructing a careful method for ascertaining whether or not disclosures concerned the "target case." It is less clear, however, whether Goodman et al. also avoided problems assuring that the underlying abuse was verified. This commentary discusses the limitations in the Goodman et al. study. (Contains 2 notes.)