The internal validity of a 24-hr. dietary recall and a seven-day dietary record was investigated among a group of non-institutionalized elderly subjects who were participating in a congregate meals program. Internal validity was assessed by comparing reported intake with unobtrusively obtained data on actual intake. Validity results suggest that the recall is prone to over-reporting low intakes and under-reporting high intakes. This pattern has been referred to as the "flat-slope syndrome." Records collected during the first few dyas were less prone to this syndrome; however, validity declined by the fifth, sixth, and seventh record days. Also, as the record progressed to the seventh day, the demographic nature of the sample became biased due to drop-outs and decreased usability of the records.